[I work on Claude Code] I broadly agree with the author’s point: plan mode was useful, and is no longer useful.
In Claude Code, all plan mode does is add a little reminder to every user message along the lines of “you’re in plan mode, please don’t code yet”. It’s something I came up with late on a Sunday night many months ago, when I got tired of asking Claude to plan with me first before coding in each new session. Something people might not realize is plan mode has always been a prompt — it has never changed the toolset because doing so would break the prompt cache, and so would be expensive for users.
This worked well for a while, until a few months ago, using early versions of Fable, I realized that I wasn’t using plan mode anymore because the model just got it, and because for the increasingly complex work I asked the model to do, planning had become interactive and iterative. With Opus 5.5, I feel Opus has gotten to that point too.
For codebase understanding, I sometimes ask Claude to generate an artifact that explains some aspect of its changes. For complex diffs to core parts of the system, I will often ask it to make diagrams or even interactive demos so I can better understand the change and alternatives considered. I don’t do this very often, but it’s a useful way to explain code when you need it. I ask Claude to attach these artifacts to its PRs also, so others can understand and future Claudes have the context.
> This worked well for a while, until a few months ago, using early versions of Fable, I realized that I wasn’t using plan mode anymore because the model just got it, and because for the increasingly complex work I asked the model to do, planning had become interactive and iterative. With Opus 5.5, I feel Opus has gotten to that point too.
For me it’s actually the opposite, and Claude Code’s plan mode isn’t nearly sufficient. Personally I ask Claude to write down a markdown file with its plan, then review the plan using plannotator, and then go back and forth (most of the time it’s actually the comments that are the problem, not the code).
Then start a fresh session, seed it with the plan, tell Claude to find ambiguities / friction points / oversights, resolve those, and then implement it.
Review once again with plannotator, go back and forth, and then send PR.
Maybe not the “vibe coding” that was once imagined, but this does ensure I am fully aware of the code and architecture, the quality, and this also prevents long term degradation.
I've recently gotten religion on the workflow that is many (relatively) short-lived agent sessions passing planning/handoff docs between themselves. It's better for my own task tracking, better for handling "oh btw I noticed XXXX", and better as a clear review point. Overall it just feels like it takes a lot of the formerly implicit context that was whatever we happened to have talked about and turns it into a much more explicit "this is what you need to know, now go".
Currently looking for a framework for managing this in a more formal way, and I think it's probably beads, but interested to hear from others.
Spent the past 1,5 years building a tool that might be relevant, helping keep durable task state between agent sessions. It is an issue tracker persisting state as immutable event logs, allows you to inspect workflows after the fact, lets you inspect diffs inline in the tickets and it is much more lightweight than Jira/Linear. There is no central service to integrate with, as it is Git-backed and lives with your code in your repo.
This looks really nice and I'm definitely going to give it a go. I been relying on Jira so this will be a breath of fresh air. Beads was great in principle and I haven't given it a look in a while but this looks much cleaner.
I have some older projects that use beads (I still run an old version without dolt that's imho pretty good overall) but lately with Fable also have a few newer projects where I just have the agent write docs and keep a worklog with the what/why/decisions etc. (I think I read it here on HN somewhere and figured I'd give that a try.)
The latter seems to work pretty well for now (slightly better than beads) but I'm always looking for ways to improve it. This could be an interesting replacement.
I think it works well for tracking intent. I tell my agents to use a "human-input-needed" tag when there is a high stakes decision fork, and then I just filter on that tag, immediately highlighting blocked workflows. I think you can get a long way with clever tagging, filtering + time travel.
Checking this out, the skill it wants me to add to my project seems to reference issues that wouldn't make sense in the context of my project. IE: `QPPR7PE` and `MBB6WY5` ?
You are totally right in that it makes no sense. It seems like the skill.md had regressed as I had updated it via agent proxy. Thanks for pointing it out, the skill has been brushed up.
Got it! I am not up to speed on Claude cloud sessions yet. Interested to hear how that works out for you. Let me know if you figure out what's getting in the way, I'd be happy to look into it.
I have a lot of little projects and I also prefer this way of working with agents. Sometimes I would start to interrogate on a specific portion or ask questions to better understand a concept, and the session would get poisoned and the agent would fixate on that topic for all the remaining turns.
I asked fable to look at my interaction patterns and clearly stated my frustrations and the problems I wanted solved, and it designed a simple process to track things in git and built a couple simple session hook skills. It’s pretty lightweight and I’ve been very happy with it for a couple months.
The power of this mode of work is that after you deconstruct the task into smaller subtasks, it's a lot easier to use cheaper models to implement that task.
I get a long way using models like Opus to make a plan of action and a bunch of tasks, and then using Deepseek to implement that plan of action. Saves a bunch of money and is fast.
- I have a record of work done and work to be done that helps _me_ when I come back to the project after several months. It’s committed and lives with the code.
- when a task inevitably ends up more complicated than I thought, I can in that session break it up
- I initiate sessions from multiple computers, so things stay in sync (through git)
- I also have a “tooling” repo that builds out some views of the work and hosts it for me to see when I’m on my phone.
- The hooks let the agent manage all of the workflow/task management, so there’s very little management overhead for me.
I rejected beads and JIRA. I wanted something more lightweight.
Nice. I found myself doing a lot of repetitive stuff handing off between agents, time to fix that. I think I agree with your approach of keeping it lean.
I also sometimes rewind the conversation after going off on a tangent, sometimes not if I want the agent to have the sorts of things I’m thinking about in context. I probably justify keeping too much crud in context than I should
I've been on this kick since I realized the primacy of the initial part of the session context. I created a python app that reads a phased plan and kicks off a new session for each phase. There is a standard prompt and handoff mechanism to determine if we encountered any unforeseen issues that we need to address in chat, but otherwise it will just grind with a clean session with appropriate context for each phase.
Just kicking off a subagent does this automatically...?
My main conversation is usually with an orchestrator that hands off work to various (usually cheaper) subagents to plan / review / etc. It has instructions to find the correct model for each task and not to do too much itself so a multi-phase plan automatically gets a fresh subagent for each phase.
i disable subagents. it also helps that i work in microservice size repos, so a single agent can keep the full relevant context. subagents waste alot of unnecessary tokens doing code exploration that the main agent is going to end up duplicating, i think they make more sense for monorepos.
anecdotally I've tried the same plan on different branches with both approaches (single agent with subagents implement the full plan vs. my little session per phase app) and then had agents judge the code quality and my approach worked better and saves tokens. horses for courses.
I use Jobs [0] to manage this—it's an agent-first CLI to track issues and tasks. A single `job orient` command gives the agent the current task in the context of the larger plan. It's a replacement for Plan Mode and issue trackers, and it has allowed me to execute massive plans in parallel with minimal oversight. There's a web UI, but it's a work in progress.
> (most of the time it’s actually the comments that are the problem, not the code).
I wonder if it's just a consequence of a gigantic training set full of comments completely out-of-date with the code, leading to the model considering this "normal"
I don't usually see Claude leaving comments that are obviously out of date compared to the code. The problem is that the comments are just a dumping grounds for Claude's stream of consciousness, verbosely recording random bits of history and context that are true and at least somewhat relevant, but without cohesively explaining why the code is the way it is. And some sentences in the comments will be beginner-style restating what's obvious from reading the code.
Exactly. Comments are basically its scratchpad for whatever it wants. Ask it to change a number in a TOML config and it'll also add 2 new comment lines above it with some Claudeslop about it being changed and what that accomplishes, as a useless note to itself.
I now make sure to do a big decommenting pass before every PR.
This kind of stuff annoys me too, and I have instructions in agent.md to avoid it, which helps.
But I also am starting to just let go and stop caring. It’s not clear to me that it causes problems down the road, it’s easy to strip out en-masse if needed, and in my experience, agents now are really good at read git blame, the commit log, even prior agent transcripts if available to sleuth out when a change was made and why. So yeah, it’s annoying, but the code agents write for me is increasingly never read by a human, so does it matter?
This. Be honest: who reads the human-written comments? ;)
When I read, I skip most comments, especially the larger "Javadoc" style. My brain sees them colored differently in the editor and it doesn't even take mental effort. Then, when I have a question about the code, I look back up for a relevant comment. That doesn't happen very often.
If Claude writes great code and leaves a garbled Claudese-but-accurate comment ... I can read and comprehend (with like 5x the effort of a human comment) ... that's a small price to pay.
Yes, comments might be my biggest gripe with CC due to what you said. 9 out of 10 of my revisions to Claude's work is deleting or rewriting comments. Such a silly thing for it to fail on.
(I do have detailed instructions for it on how to comment (or not) but it has not fixed this.)
Nah, it’s the typical Claude-isms, and way too much info, info in the wrong places, and putting lots of planning discussions into comments that are completely irrelevant.
It’s always “you explain only what but not why” or “this is way too much prose” or “these comments don’t belong here, they should be inline comments” or “this is completely redundant as it’s already obvious from the code”.
I do find that once I beat it into submission and the codebase is “clean”, the new code it generates gets better and better, which makes sense gives its pattern-prediction nature. But it seems like there is work to do for Anthropic in terms of getting Claude to not confuse code comments with dumping its interactive discussion state into there.
Superpowers never really held up in my own testing vs /plan - not just more expensive but markedly worse code organisation, because the plan itself leads to higher cognitive burden for the models - and today it's even worse, because /plan doesn't hold up versus coming up with a few high level slices and having cc work on one per session, for much the same reason.
5.5 is much closer to Fable so i don't even need it. I am pretty sure it's got Fable's DNA in it.
I really need to find a role where I can do more DX...
Yeah I always balked at the amount bloat in that repo, and I just prefer a more lightweight setup and plannotator’s UI which I can use to interactively review stuff. It solves this one problem better, where superpowers tries to do way too much, poorly.
Going through the superpowers spec-plan-implement-review cycle took one of my test runs (local AI) from 40 minutes to 8 hours, and produced a lower-quality result. I want to use something like that skill, but I'm doubting the efficacy of this one.
I dived into one or two of the skills there and some of it feels very cargo-culty. In a similar vein to the article, the models have improved markedly now and I’m not convinced that “You are an experienced senior software engineer and an expert reviewer, here are five pages of instructions on how to do a review” style prompts are massively useful any more.
I get great review results (as good or better than colleagues using superpowers or even adversarial review skills) just by asking Claude to review a PR and spit out results in order of severity.
My workflow is very similar. But I just ask agents to write design in html instead of markdown, due to its richer layout and better interactivity. When the design is about UI or anything related to graphics, this approach is extremely efficient.
I also found that having the design reviewed by multiple agents has very little marginal value. The review agent will always find something to improve, but mostly it’s just nit and not anything super important.
I used to let Claude just upload the html design doc to Claude artifacts for me to review. Recently I switched to codex and started to use my own tool https://github.com/hyperlogue/r3 to complete this workflow.
I do the same, I don't use Claude Code or Codex planning because it is mostly pointless, even with Fable/Astra. I just have multiple agents work on a markdown file which I manually perfect, often breaking into multiple different files for large features or PRs. I also create design 'handoff' documents which I feed into Claude Design or Astra along with screenshots and wireframes. By the time an agent does something I'm well prepared.
I've tried doing the incremental, iterative approach with just Code and it's just not as effective unless you're working on something simple or experimental. Or you're shipping to something non-serious or perpetually beta.
Have you recently tried working without generating that plan? What I've been doing is first tell the model I want to plan the implementation, talk about it a few turns, then when I'm happy with the idea and method just tell the model to go ahead with the work.
Note that this is only really necessary for complex work that I don't know yet what the best way to do it is.
I've tried doing it your way as well, but there was just too much fiddling about with writing the plan somewhere, then having another session rebuild their context with whatever info is in the plan. It really didn't result in better output for me.
Currently 9 times out of 10 I just say to the model: xyz is the problem/bug/feature, fix it. Since about Fable and Opus 5, this is more than enough. Opus 5.5 (and previously Fable 5.1) got even better at this. However, this is in a codebase where there are already a few hundred thousand lines of code for the model to look at to see how we generally attack things in our codebase.
Claude Codes plan mode I never use anymore, it was useful a few months ago because the models had a tendency to just start doing work and forget I specifically told them not to. But the UX is just annoying and the models now do adhere when I tell them not to change anything.
I have, and it always does something unexpected and undesirable.
Plan mode ensures I'm spending fewer tokens on the code-test loop, and more on the arch/design, and allows me to keep appraised of what's going on, while planning for future changes better.
Maybe folks who don't need planning, don't have as much concern for the details, and are happy enough with just evaluation of if it works or not.
I have roughly the same workflow, also with plannotator - which I like a lot - and haven't used or felt the need to use _plan mode_ for at least 3-4 months.
Then telling Claude to work on a document, the instruction is kept to its core.
Now when bcherny explicitly mentioned that it merely adds a single line - it explains why I don't need it.
What may be concerning about "super plan" mode from the creators (or a skill, for that matter) - is that tuning the amount of effort, and how much deep to dig - may become too hard, as it will interfere with several embedded paragraphs explaining what to do, how to do, where to do, etc'.
What I do look for is even better plannotator ability to track changes, combining historical comments (like Google docs), and git blame of several "generations" before current reviewed doc.
While possible to commit, I'd still loose the comments themselves, and the commit will be at least per changeset, and not connected ro a specific changed section.
Roughly speaking, I'd be happy if plannotator would persist something similar to github PR reviews combined with Google docs comments & suggestions.
I do something similar. But where I change it up a bit is depending on plan complexity I divvy up parts of the implementation to different subagents with fresh, only relevant context for whatever they're responsible for doing (e.g. a part of that plan). Edit: to clarify, I will also go back and forth with the planning agent making sure edge cases are covered. Sometimes that involves invoking a new subagent without our planning context to validate it without preconceptions.
I personally still find planning a valuable mental exercise; it's not so different from pre-LLMs and whiteboarding or otherwise taking the time to consciously plan a set of work.
Yeah, sure, but you don't need a dedicated plan mode for that at all. you can just do it in auto mode, and say "let's do some planning first", and Claude will (nowadays) be smart enough to understand that it's not supposed to jump straight into the implementation.
So again: You don't need plan mode, auto mode works just fine, there is no difference in the workflows here.
I do something similar but a bit more involved, using a few informal stages. Let's say for example I'm trying to launch a new complex feature for https://coderba.se.
- strategy document
- "sprint" document with technical implementation
- actual implementation
- e2e testing scenarios updates
Every step involves iterating with Claude on it with me in the loop (setting the direction then resolving the "founder questions" as they appear), and importantly a different model for review/code-review, be it Codex (usually, it's great at it) or Antigravity/Gemini (sometimes finds novel things, its precision and recall are abysmal but on the odd occasion it has good accuracy). This iteration on the high-level plan then on the implementation plan is essential to me, and IMHO part of why people are surprised that I tend to get solid results from LLMs. At the very least, it allows me to fill gaps in my own knowledge (primarily front-end development) and be more productive than writing the code myself. I cannot stress enough how nice it is to have a partner in the high-level system design – yes, it often suggests utterly moronic ideas, but the overall experience is still net positive and getting better every quarter.
It's useful because it let's me see the decisions the model will make before it wastes a ton of time implementing them. The model is smarter now but that doesn't solve for underspecification if it guesses my intent wrong
Interesting, I don’t see this very often with the latest models. Are you using Opus 5.5/Fable 5.1?
Either way, plan mode isn’t going away. You can always /plan or ask Claude to enter plan mode. We might re-map the shift+tab keyboard shortcut to something else by default for people that don’t use plan mode.
> Interesting, I don’t see this very often with the latest models.
Strange response. I agree with the parent comment here, plan mode lets me ensure that I have specified everything correctly before it gets built which is far too late. I don't see how an improved models even matter to this workflow. Is Fable going to read my mind?
What is it with the snide responses? This threads title is "Plan mode is dead", not "You don't need to plan anymore".
Boris is saying that you don't need /plan to get the model to plan, you can just say "Let's plan this out" or similar, which at least matches my experience. Your experience may differ, of course, but it's not even clear we are talking about the same thing.
Maybe some people have not been long enough on this rodeo: This used to be an actual issue. You told the model "DONT START CODING YET" and yet, surely enough, starting to code it did. That is what /plan etc were supposed to fix.
> if I am not even quite sure what exactly I want to build myself
Sounded to me like you need a plan.
My approach is to take the statement of work or problem definition and iterate on that myself until I'm really clear on what is the goal. I therefore have a good some good ideas about what the plan should be.
If extending an existing application, which is usually the case, then make use of the plan documents that I had written before AI arrived on the scene. These sre documents in markdown form that say step-by-step how to, for example, add a new report to the system.
You tell it, "let's brainstorm, don't implement anything yet". Then you talk about what you want to build and hammer out all the corner cases. Then you tell it "now do that".
My experience with that is not very good. It gets so hung up in its initial decision. Like, it won't make changes because they're "breaking", for something never even committed. Or it will litter the code base with defensive code and comments about the path not chosen.
> My experience with that is not very good. It gets so hung up in its initial decision. Like, it won't make changes because they're "breaking", for something never even committed.
This hasn't been my experience.
> Or it will litter the code base with defensive code and comments about the path not chosen.
I love when people on HN think they're special or do more "real" software engineering.
Basically every big tech company maintains "codebases with millions of lines built upon decades." Talk to your friends at a FAANG and ask them how they're using Claude/Codex.
Merging after reading the PR description is just how it's done these days, and if you can't do it reliably, your harness, devloop, or model is simply behind the times.
The model is irrelevant, plan mode helps us ensure the model will actually do what we have in mind. The best model in the world can't work around misinterpretation because of bad specs. I'm not sure why this is even a discussion, isn't it obvious to anyone actually using LLMs?
This is an out of touch with reality thing to say. Remember: not everyone has access to unlimited usage/money to spend tokens or the most expensive models with the higher thinking settings, exceptions does not make the rules.
> This is an out of touch with reality thing to say
I mean, you're discussing this with a marketer/someone wearing a marketing hat, who works for a company which needs people to use as many tokens a possible. That's their reality
I see you didn't disagree with the main thrust of my post.
You'll also note their 'About' is empty. Regardless, engaging with HN in this manner is de facto marketing/PR. Trillion dollar companies don't just let anyone post on high-profile social media websites for any length of time without permission from marketing/legal/PR.
Hey, well first off, congrats on making the greatest product ever probably.
I absolutely see fable and opus 5.5 misunderstanding intent, but that just seems to be a feature of necessarily underspecifying in a written prompt. Just today, I gave opus 5.5 a simple task to spin up a new environment for work. It read the ticket, which was decently specified and knowing the codebase as well as "Ghasp... reading the code" I had to correct it about 5 times to do it in a way that I would have expected it to. Getting the pipelines right, environment variables, and configs. It was all relatively straight forward imo. Then I had to prompt it to clean up its corrections, because it left a workflow variable in the github action that some intermediate step required but the final solution didn't. I definitely would not have caught that if I didn't read the output. Idk, there seems to be a natural limit as to how much it can infer and I have no idea how to fix it. I did write about it [here](https://javiergonzalez.io/blog/the-assumption-problem/) though.
I’ve been having similar issues. Absolute love fable but it keeps leaving development servers running that are blocking port 3000 (rails apps run on this by default) and then when I try to launch the app and realize the port is in use I ask fable what’s up and it says sorry I left x running and then shuts it off freeing up the port.
Why would he have a field day with it? It seems he oscillates between finding models the greatest thing there is to being absolutely stupid so not sure in which direction you're implying he'd have a field day.
Theo is a YouTuber & prolific X tweeter who has a channel about AI. This is not Theo, but gives a flavour of Theo's style: https://youtu.be/h1p9zdUtUdo
I don't think me giving the model bad instructions is something a smarter model can solve. I use plan mode constantly (with opus/fable), and at least once a day I'll say something too vague or just dumb and it will sketch out the "wrong" solution in it's plan.
Which is fine because it just put together a plan and didn't spend 10 minutes rearchitecting everything.
I use fable 5.1 (tried opus but it lied to me 3 times in quick succession and ignored me in another)
This is weird to ask because I feel like of course the model isn't omniscient? Isn't the whole point of iterating on a plan to assess impact, risk, know your (the user) variables, user impact, product impact, etc for making a change? I cannot count the times even in the past few months where I start a conversation with my C suite because their desired outcome would have a potential negative impact elsewhere for other products or users.
Is this just not something that comes up at Anthropic?
It happens with Opus when I leave too much up to interpretation and the AI doesn't do what I had envisioned but didn't specify. Like sure, what it did may be a technically correct solution but it's not the correct solution that allows for further development of my idea. I'm not sure how others do their projects but I start small with proof of concepts and develop in layers until the project does what I want. I use plan mode first to layout everything I can think of that I ultimately want and describe features in the best detail I can manage. I work with Claude to figure out the best framework or find whatever existing projects can serve as a starting point. The first milestone is the proof of concept, take the framework/existing project and build something that does the bare minimum of what I need in the way I want it done then build a test suite to make sure it works. Once that's proven out, we start adding more features (both mine and the ones Claude has suggested) and adding/revising tests along the way. For small things, I won't bother with a plan since I generally already know what I want or any ambiguities can be solved in a single response. But for larger things, I try to take a waterfall approach with well defined milestones.
If I knew exactly how I was going to build something, I would have built it myself. But since there's some ambiguity in the portions of the project I'm less familiar with, I rely on the plan to not only help me understand the decisions Claude has made for me but to keep Claude constrained to the decisions I've made. It's very frustrating to waste tokens on having to refactor something because
Yeah IMO there's two entirely different traps I've seen in startup codebases in the last year or so:
1) "Maybe waterfall works now" - plan mode, take care of all the nits and issues that the bot leaves on your PR, wildly overengineered "enterprisey" solutions with a lot of bells and whistles all over the place but very poor end-to-end user story test coverage that results in user experiences with a lot of good test coverage of the edge cases of how a given step might fail but little thought towards overall user flow and throughput. Because part of the issue with waterfall was assuming you could design the right tool for your user up front.
2) "Maybe code doesn't matter anymore" - The just ask for something when you need it approach, which results in weird janky individually-sorta-working but strangely-overlapping six-variants-of-the-same-thing that makes it hard for your users to develop a single consistent mental model of the thing they're using, and that changes super frequently.
They both end up with a lot of other bad-for-velocity things that I assume are inherent to how the tools have been refined in response to last year's criticism, too. Super verbose comments. Extensive - without much eye toward runtime - low-level test coverage that might miss the forest for the trees and also slows down the next round of iteration cycles. A plethora of new proper nouns all over the place that make the documentation an ouroboros without a good entry point.
I haven’t tested with Claude specifically in a while, but I see this a lot on larger features.
It tends to be small decisions way down the stack that bubble up, or an incoherent data model that can’t handle what you’re asking for cleanly.
Eg I was messing with a state tracker the other day. The state tracker assumes a container is either currently running, or fully removed from disk.
The LLM chose to remove the state file when the container is stopped and then to remove it after, which leaks container storage.
The LLM is kind of stuck though, because every option other than “rewrite the data model” has negative outcomes and it probably violates user expectations to launch a massive rewrite there.
I've found Opus 5.5 is quite good at surfacing and clarifying these issues using Grilling[0]. Often I find that I want to expand the scope of work much larger than Claude would consider based on my original request.
We got Opus 5.5 just a few days ago so haven’t had time to learn its ways, and Fable was/is too expensive for many people including myself. Insiders have months to learn the ins and outs of new models.
It would be nice if with new model releases claude code also gave a bit of a model 101 that tells you evolving ways of prompting it that the insiders have picked up. I know there’s the prompting guide in the claude docs, but this is often very broad and most people don’t know about it.
I have primarily used Fable 5, Fable 5.1, and now Opus 5.5. I never use shift-tab to enter plan mode personally. My general workflow is chatting with the machine to work out fiddly bits, then /plan to get everything in one document, and reviewing the plan.
/plan is still useful, I still need to review what it's going to do and still make revisions. But there's two phases: hammer out key design decisions then write and amend the document.
For me, every day, although it fails far less often at this than Opus 5 did, which might as well have been a gremlin. It was a really poor technical writer too, leaving slop reminders to itself in all prose, including comments. So even if 5.5 fails less right this second, I have trouble believing that it all will not similarly bite me next week.
Summaries that don't tell me when it's changed direction in a timely fashion, but I am only told way later, when I have to undo. Really bad judgement calls regarding where to fix bugs, changes in implementation decisions, taking action when I am asking a question directly, not passive aggressively asking for action... when 5, 3 days ago, was proven to be untrustworthy, switching to very little supervision sounds like a strange thing for a customer to do.
How could a newer model be better at making up information? Do you understand basic information theory? Or maybe what you're saying is you don't actually have any ideas and are happy to do what everyone else has done before you so it makes no difference?
5.5 and 5.1 have major Rain Man (savant) syndrome. Excellent at many hyper-technical things, absofuckinglutely boneheaded at anything that a human (or an earlier model) would understand - like how to write copy, what a human would expect in a given situation, various types of norms...
it's infuriating because it's a sophies choice - dumber model but better human understanding, or better technical model that you have to explain things to over and over like a toddler.
Why is “are you using the latest model?” the automatic response to any even mildly critical of LLM coding?
OP wasn’t even actually critical of LLMs, they were just saying that plan mode was helpful to stop the model from making incorrect assumptions when you want you don’t specify everything you should.
I greatly prefer this, since it lets me iterate on the plan with Claude for a while without it repeatedly asking if I’m ready to implement the plan.
Once I’m satisfied, I usually start a fresh session and tell it to implement the plan.
For smaller plans, you don’t need the file. Just ask it to come up with a plan. I don’t recall the last time it just started implementing if I only asked for a plan.
I guess that You dont need plan mode any more is ambiguous. I took it to mean you dont need to plan because the models are so good at infering intent. You are taking it as "you dont need a dedicated modality to create a plan".
It’s not that planning is dead, but rather that planning has outgrown the simple “Plan Mode” feature as models have become capable of taking longer turns.
hi I’m the author of the post. I think that’s basically the distinction I’m trying to make.
Historically, plan mode served two different roles:
1. making the agent’s instructions precise enough to execute
2. helping the human understand what was about to happen
I think #1 is less necessary as agents get better. #2 is going the other direction, it becomes more important as the model is able to do more on its own because larger chunks of work are happening with increasing complexity.
Where I’ve changed my mind is the interface for #2. I increasingly think an interactive, iterative workflow is closer to how people actually build understanding than being handed a long generated document, especially one they didn’t author themselves.
The human-understanding problem is very real though
Awesome blog. You're a good writer. I enjoyed seeing your article on AI in 2018. Thank you for sharing your expertise.
Also I hope your delivery goes well. My wife (and co-founder) had a challenging delivery and it really put life in to perspective for both of us on a range of issues (how much women's pain is minimized in the health system requiring stronger personal advocacy than I would ever have expected).
As far as plan mode, I still find it essential in keeping agents on track. I build propelcode.app and have a variation on plan mode I still find useful, happy to trade notes on agentic coding if youre interested.
I'm not the person who responded to you but I hope they're able to help her and she's okay.
And thank you for the article - it was a good read. I still see folks in my org playing "throw spaghetti at the wall and see what works" and getting frustrated so plan mode (mostly point 2) has been their guardrails almost as much as for the AI.
I never understood fear and anxiety until I had a kid and was worried about her safety. I am sorry youre facing that and hope she recovers and thrives.
Can't help but think "Doesn't matter if a machine or a human with (even slightly) different background wrote it", maximizing information flow is maximizing common assumptions and "culture" to only have to communicate a small set of current information for the task at hand. Being a team means having built a joint context so to say. This has always been the purpose of design documents and they always were too big or too small. Because you did not write them, but the others. If you only produce code you think they are the past and useless. If you iterate and your team grows, you start seeing the value in always current docs that are containing just what is not in your everyday culture.
All the best for you and your growing family. I had a similar experience recalibrating my values...
Fascinating! I literally never use Claude without plan mode and I find it's basically useless without it, constantly wasting tokens going in circles on irrelevant things. Fable or Opus. I feel like neither has a good sense for how to architect things and if I don't use plan mode it usually wastes hours of time chasing it's tail or implementing kludges on kludges to get something working that would be a much simple fix elsewhere, especially when working on a larger codebase.
Hence why the Anthropic employee is telling you to not use it and just aimlessly throw tokens at a wall. It'll eventually get you there, sure, and consume more tokens. This sounds absurd, but trust that Anthropic (and any large company) is hyper aware of how customer behavior impacts their revenue, and they certainly will try to steer you into behavior that increases revenue.
Also this is a way less removed process that I want nothing to do with. The more removed I am from the process the more I hate my job, get burned out and genuinely wish that Anthropic never existed.
Even if it could "just know" or infer my intent. It wouldnt be desirable.
Edit: Oh yeah its Boris, hes one or the most disengenous shovel sellers on earth right now.
So you think telling a model once to "do not write code, just plan" vs having an enum that causes that same text to be added to each message will somehow waste tokens and will surely make anthropic much better off..
I mean if you want it to plan something without writing code first you literally just need to ask it to plan out the implementation first it's not exactly a difficult concept
The Anthropic employee literally told you plan mode is a "we're still planning!!!" at the end of the prompt. Plan mode is useless because you can literally type "dont write code yet" and get the same effect, not because you should never plan in general.
Sure and I should also have been more specific, I don't always use "plan mode" I often just do planning without explicitly activating it just via prompting, point being that the robot doesn't work for me without some time spent planning out what to do.
I realize there are already a ton of comments, but I think you're missing the idea of _precision_.
Most people aren't precise when initially describing their problem.
Jumping straight into implementation skips the part where we refine and better define what it is we're trying to do, and think through the implications of those changes.
I suspect "trusting the model" doesn't really work at scale with finite resources.
* Scoping discussion - do research and figure out approach (auto mode)
* Planning - take the scoping and convert it to a concrete plan for review (plan mode)
* Implementation - put the plan into action (auto mode)
I find this works really well for my workflow, and it is really easy to trigger each phase because the model has clean boundaries. (This workflow is articulated in my user CLAUDE.md) Plan mode is still useful to me as it forces the model to double-check its plans (I find even with Opus 5.5 it still discovers gaps), and it gives me an opportunity to clear context at a really good spot.
So I would still consider plan mode to be useful. It would make me sad to see it go.
PS- I have a Claude Marketplace directory submission for an MCP server that has been sitting in review hell for six(!) months. I've never received any outcome other than "In Review". I hope I'm not asking too much but would it be possible to put me in touch with someone who might be able to help here. Nobody has ever replied to messages sent to mcp-review@anthropic or the "Anthony at Platform Operations" inquiring about status, and we are getting frustrated
I tried plan mode when we first added it to github copilot and it didn’t stick for me, until very recently, when I used it the way you describe. I was just putting it too early in the process, turns out I have to do a bit of exploration and discovery on my own and most of the time I can skip plan but now I have an intuition for when to engage it so it can interview me to clarify the last few things it needs before implementation. This also made autopilot mode work a lot better for me.
> In Claude Code, all plan mode does is add a little reminder to every user message along the lines of “you’re in plan mode, please don’t code yet"
Can't help but think if plan mode isn't useful as you say because it's implementation is lacking in claude code, hypothetically speaking.
What I can say is that with other harnesses plan mode helps stabilize my workflows. Actually synthesizing code is only part of the process, lots involved in taking a work-item to production end-to-end and plan mode helps give this flow structure. More than that it's an opportunity to regroup before committing to changes. It slows down the process to a rythim that's sustainable and smooth, which ends up speeding up the process.
So if plan mode in claude was designed to speed up code churn, while oh my pi for instance designed plan mode to be strategic, that might account for the different perceptions here.
And it's not to say you should force yourself to use it, but if you are planning on cutting this mode off the loop just beware of the possible side effects.
It seems unsurprising that if your definition of "plan mode" is as tiny as appending "but don't write code yet", that it would not be that useful for that long. There have to be more sophisticated versions of what "plan mode" means out there.
Also if you are working in a heavily vibe-coded codebase, as Claude Code reportedly is, it's not that surprising if the human doesn't really understand it or have anything useful to add in a collaborative context.
GitHub Copilot for VS Code by default configures "Plan" mode to disable everything except reading, asking the user questions, and writing into the memory scratchpad.
Claude Code's tool system feels quite rudimentary in comparison, but that's what you get I guess when you're only working with models you're going to finetune on being able to handle specifically your own tools anyway.
Plan mode is critical in the beginning, because there's a lot of long tail decisions that haven't been considered, let alone resolved.
But as skills and memory are populated over time, Plan mode isn't as necessary. It becomes simpler to let Claude just build and get something general in place that works, and then refine from there. Auto mode will ask essential questions.
I still use Plan mode for big feature changes, to confirm that I've asked for what I want in the right way. I tend to prompt casually, with only a few specific details. Plan mode helps me see the whole picture before committing. In a few cases, it also helped me decide the feature I asked for was wrong.
I don't use Plan mode because I plan out work with a ticket based approach (that also works as a memory store), but I wish there were some kind of Plan mode like harness feature to try as hard as possible to stop it from beginning to code. With the newer models, I have started to have a lot of trouble with them proceeding through the QA with me where I scope out the work and refine requirements and then immediately begin to do the work rather than either making tickets or giving me the full plan to thumbs up like I tell them to do.
So it sounds like we use it somewhat similarly, just taking a glance at it before the work starts, and that's becoming harder to do in my experience. It is important because problems with the overall plan or strategy end up magnified the further you proceed with development.
I disagree with the assertion that the model gets it. Here’s a practical example I just tried with Fable 5.1. I gave it this prompt: “Write a Go function that can be used to establish secure communication to a remote system using a certificate. Keep it short, single function, and explain how to use it.” The output forced the use of a private key stored in a file even though that wasn’t specified anywhere as a requirement. The function Claude wrote takes a private key file argument and calls a Golang function that requires a private key file (tls.LoadX509KeyPair) even though Go has crypto.Signer which could support private keys in various other manifestations like HSM or KMS. I argue that a person who “gets it” (or who is reasonably experienced in security) would have opted for not requiring private key material for this to work.
For the record, this isn’t unique to Claude. ChatGPT and Gemini do the same, each with its own quirks. ChatGPT got extra credit for being the only one who allowed the function to also take a CA file for server authentication.
Don’t get me wrong: LLMs are the future (maybe even the present) of software development but I think there’s some way to go before they can be entirely hands-off in some areas. I still find myself having to course correct designs and plan mode helps me with that.
And of course, thank you for your work on Claude. :)
I'm hoping that by "gets it," he meant that if you start a discussion about the design, it doesn't misunderstand and immediately go off to do the work. Some models tend to do this.
Whether or not a distinct "plan mode" is needed, upfront planning remains essential in my experience, even with Fable (albeit not the 5.1 version). I agree that, as the models get better, you can skip planning on increasingly complicated tasks.
But there is still a ceiling above which it is necessary to "preload" the context window before starting to call tools and get into the meat of the work. You want to establish domain language (especially with Claude models which otherwise will invent their own, and it will be inscrutable) and key requirements and assumptions. You want to do a Q&A iteration cycle with the LLM. You definitely should do a sanity check that the LLM actually "understands" what you were trying to achieve, and then make sure that understanding is coherently and plainly stated in the prompt. All of that seems to be necessary still for just about any serious task, if you actually care about the quality of the results and/or don't want to burn hundreds of thousands of tokens on flailing around to get to a good quality result.
So no, you don't "need" plan mode. But you do still need to do all of the things you would do with plan mode.
But in your example you never even asked it to plan so you could check the implementation before writing it you just asked it to write it directly, so this isn't even a comparison to plan mode
Apologies, I didn’t mean to imply it’s a benchmark, I just wanted to provide a reproducible example of where I see models make decisions that seem to be fine initially but might paint the software architecture into a challenging corner. I don’t expect models to read my mind, but I do see them produce a lot of verbose output, none of which is used to say “here’s a simple response to your ask, but have you also considered...”
I think knowing which models use the standard library and which ones pull in dependencies as highly useful. As a dev I've always favored staying as close to the standard library as possible as it makes refactoring, understanding, and deletion much easier. Not everyone has the same preferences as you, it's nice to understand what matters for others too.
> In Claude Code, all plan mode does is add a little reminder to every user message
I was quite surprised when I learned this (when Claude Code edited a file despite being in plan mode). I had previously assumed "plan mode" was a harness level concern, and restricted what tools could be used. I didn't expect that it was simply an addendum to the prompt.
But I think it gave me some good insight into where the heads are at of Anthropic employees building this. Basically, leaning on pushing everything to the model. That's why alignment is so important: a "sufficiently advanced" model doesn't require any tooling infrastructure around it, and I suppose Claude Code devs are targeting that future. I had previously thought there was more to a harness, but with "auto" mode these days, it seems like there's no desire to build in that direction.
I haven't reached for Plan Mode in awhile--maybe, on some blank folder/canvas and I just want that cute "questionnaire" DX to get me going...
But on the whole, CC is smart enough to know when to "rush off and act" and when to "pushback", which is great--and it's no big deal to tell it to pause/stop by adding "what's your thoughts?" or "feel free to pushback" etc to my prompts to make it start a back-and-forth with me (the fact that you say that Plan Mode was really nothing more than a prompt anyways is reassuring).
And yeah, when you're deep in the weeds, you could (and can!) have multiple threads of thought/work going on in the same convo, that stopping and starting a plan in Plan Mode seems like a regression.
I get the idea here, but I find planning very useful as a phase when I'm doing things I am not intimately familiar with (e.g. AppKit programming).
That's when I need to learn about implied patterns, do's and don'ts; not from theory but in the context of my own project. Unfortunately, the model tends to keep implicit knowledge implicit. But I can ask during planning.
The feature did get less useful over time when the model started babbling in newspeak more and more. When it threw a thousand words at me even in concise mode.
So I don't want to let the toolmakers off the hook here. There's a lot to win that would make plan mode much much better without changing plan mode itself.
To be fair, the team also neglected the functionalities of plan mode. It was unusable for since july when they removed the option of running /clear before implementing a plan. Also, there was a period of time during the dark periods of opus 5 and nerfed fable 5, whereby if you didnt explicitly request for code, it went on an on for a few sessions planning and researching in a loop.
FYI I had Clod attempt to corrupt a prod db the other day. (Opus 5)
I was experimenting with a rather complicated backfill operation, were I had a validation script I understand and have Clod come up with the backfill script. I was running against a local prod copy, and it proposed running the actual (unfinished) backfill script against prod.
It didn't have access to the secrets and I also caught the command, but a good reminder that this stuff needs guardrails.
yeah that's crazy. its so good 99% of the time but I've seen it have some insane hallucinations before (as late as Fable... cant remember if it was Fable 5 or Fable 5.1).
Hallucination not a big deal when it's on the surface layer. But I can't imagine the damage it could do if it hallucinated while building/validating a "load-bearing" component and then continued down that path
yeah (at an early stage data-focused startup). I think when our product is a bit more mature we'd probably have a staging tier with full access and then manually promote builds / DB changes to prod. So many things to build lol
we DO daily snapshotting, so the risk is limited... but still spooks me
> plan mode has always been a prompt — it has never changed the toolset because doing so would break the prompt cache, and so would be expensive
seems like plan mode could turn off some tools, even if it doesn't change the set offered to the model, the ones that they have which would mutate your codebase could just not work with an error message, and plan mode could change permissions in the security approval prompt for "auto"
anyway, isnt the right way to know if plan mode helps or not, to run an experiment? we're all guessing unless we have data
read only agent mode sounds straightforward and useful to me
It’s bizarre to me that Claude Code doesn’t have more guardrails on plan mode. I mostly use Copilot and the plan mode there has teeth. In plan the agent does not have permission to write to the filesystem outside of a temp directory and the plan document itself, and tool calling is heavily restricted.
There’s also a “ask” mode which is read only. Both are enforced by the permissions model, not just a system prompt instruction. I’ve seen the model “forget” and try to start coding - it bounces off a hard permissions failure and that “reminds” it that it’s in plan mode.
In my experience, around the time the author describes as starting to not need plan mode is when erasing the chat history became an anti-feature. I found the agents were doing better when they had the context already, and with the history I no longer needed to micro-manage persisting various caveats and rejections to the plan artifact. The amount of prompt construction necessary went down overall.
By the time my plan’s done I’m usually between 200 and 350k context. Even if keeping that around gives a performance bump for the implementation (which I haven’t noticed to be the case) it balloons the cost. I would much rather put everything in a plan file and start fresh.
Plus, I usually plan with a more expensive model and guide implementation with a cheaper model (with smaller validation calls back to a more expensive model)
that still essentially works. ask it to write a plan doc to a file. then when you're ready to implement, start a new session with a prompt to review the plan doc and then start building.
What I do nowadays, for large changes, is have Fable create a HTML explainer for what we're gonna do with code snippets, which is not that fat from plan mode only much more convenient for me and modern models have no issue using that HTML artifact as the source of truth, and then before I get into execution - I interactively build an end to end test that also includes pieces of the plan.
When a test case fails, the relevant part of the plan is surfaced in the error. I find this helps Claude stay on track for longer - I've been able to do 12h most times and even up to 48h unattended (11h of API time) with good results.
Then whenever I do check in, I ask it to update the HTML with current state in an append only fashion (sort of like it's writing a blog), and then based on that, we iterate on the end to end test (I think of it as a "test harness") - update the test cases and error messages.
I've been able to build some truly large projects this way, both greenfield and up to spec (for example, a video game I've always wanted to play), and brownfield while staying within the conventions and design of the codebase, and with very little attention required on my part.
Plan mode was great, but I realized I progressed well beyond it. I found that I was getting these categories repeated errors and oversights from Claude (and frankly it hasn't gotten much better about this). Skills were too generic and got lost to context.
I ended up building out tool an MCP server that I use as a bit of a psuedo harness for Claude. I have a variety of multi-step workflows that are basically micro-skills stacked on top of each other. This helps me make sure that I can get Claude to think in a repeatable and reliable manner.
For coding, I've found that I have a few specific steps that Claude needs to do before I'm comfortable letting it loose:
* It must extensively explore the code base (including certain areas that it misses)
* It must think about what it doesn't know or is making assumptions about
* It MUST scaffold out it's intentions. Essentially, it can write comments, classes, and method stubs - but no actual content. Very much like a spec, but since it's in and alongside other code, it's much easier to identify problems.
* It must spike and validate key assumptions. This, plus the prior step, are the only way I've figured out how to avoid it ending up in a confusion loop. Too often it looks at poor-quality code it's written and thinks it's a long-term solution. By avoiding writing code as much as possible, it knows that it's draft content.
I use plan mode, because my current project benefits from "pair programming."
I have one session define a task, and provide a formal specification plus context in a "cover letter."
The session B, in plan mode, produces the plan back.
Session one reviews the plan and clears it, ratifying portions and often specifying specific changes.
Session one then executes.
What has been striking to me in this approach is that even with two instances of the same model (currently Opus 5.5), there are regularly corrections made. I use "project chat" for session A and Code for session B atm; it is very typical that Code finds and corrects details or oversights in the task spec; it is also typical (though less so with 5.5) that session A (chat) pushes back or clarifies things Code doesn't have the context for.
I have been afraid to open up the potential of negotiation beyond what this is costing as it is. But I am also afraid to simply skip the formalisms, because of the consistent correction that occurs in this back-and-forth.
Each component of the pattern is schematized, generated from a template, and validated, to keep things tight.
Lots of tokens! But I trust this process far more than "just typing" :)
I am curious what your take is on skill packages like Obra/Superpowers and grillme and the like. On one hand I find them extremely useful for asking me questions. I wouldn't have thought of myself or that I don't feel I would have necessarily gotten in a structured a format just by having the AI interview me. On the other hand, I am not sure how to tell when they have become obsolete because I'm not sure where to even start with generating evals for such a thing.
Yeah I think this is a natural consequence of longer task horizons. When I was chaining 4h tasks, I can mostly plan them up front.
Now that I’m frequently designing and delegating day/week scale features, the flow has to change; having the agent go off and build a spike can be a quicker way of us understanding the design space and constraints (especially in a huge codebase). I still have the agent write and update a spec doc as I go, but it’s not waterfall anymore.
At least for my kinesthetic learning mode a rough code PR stack is usually way better than a plan doc anyway, and tokens are cheap enough (vs my time) that going further than just a plan is often cost-effective overall.
The dream of course is (say it with me) loops, but that doesn’t tend to work for me on new features often.
Plan mode is what makes it easy for me to update my mental model of the codebase, and helps me decide if the mental overhead of all the changes that are needed are even worth doing, which are now the biggest limiting factors (reading the code diffs just doesn’t scale anymore).
Honestly I don't agree with this. A few reasons why I feel like I will always want a plan mode:
1. I want to know whats going to happen, at least at a high level, before changes are actually made.
2. Plan mode helps me flesh out the missing details of my plan before being mid-execution
3. In situations where I have a limited budget for AI usage I will often times use a high powered model like Opus 5.5 or Fable to make a detailed plan, then scale down to a cheaper model for implementation. I feel like this saves cost in the end.
I get plan mode is basically just a small hidden prompt. I get that I can basically just preface my prompts with "make a plan only, don't make actual changes." Maybe this is just a UX trick, but it works well for my brain.
I'm not an Anthropic model user, and the true frontier of the frontiers is beyond my budget. Maybe it's better in the rarefied atmosphere of Astra, Fable and Opus 5.5?
But with GPT 5.6 Sol, I'm still finding that the model makes conceptual mistakes, or gets edge cases wrong, or assumes incorrectly (making an ass out of both user and model). In many cases, I need to at least refine the proposed approach, or amend, correct, or flat out just stop and start over. Not planning and catching these errors, and just letting the agents code their code, would mean I'd have to rollback and redo many times. What a waste!
For a current project, which is ~33k lines of code, I'm also finding that I know the codebase better than the model, and that's vital at the planning stages too. If I wasn't in the planning loop, the model would have reinvented various wheels a few times over. How much spaghetti do you want with your code?
As always, I may simply be doing this wrong. But I'm personally not convinced that the plan is dead, or that I want the plan to be dead. Planning is also good for me -- it keeps me thinking about the code, prompting better, guiding the model better.
If I'm no longer on top of the codebase, then at some point my prompts will devolve to "Do the thing with the thing, that does thing". And I don't want that.
Agree, and for me I feel like I often have more implicit intentions than I write in a prompt. A plan helps me verify whether an agent gets these right or not. Plus, it highlights tradeoffs I might've not thought about. Removing both feels like lowering a quality bar.
On the other hand, for a low effort hobby project: just do the thing.
It is absolutely true that Opus 5.5 just ‘gets it’ far more often than gpt-5.6-sol, which is more like an idiot savant. It can nearly always do what you ask it to, but that might not be what you want.
Yer, it’s probably best to face reality and understand that for actual software engineering / complex coding work - Anthropic models are way ahead of OpenAIs…
As someone that never used the built-in plan mode, but did use a lot of spec-driven development, I’m still finding that even with Fable having “plan” docs is still quite helpful.
They’re most useful for broad changes (new features, refactors, etc.) where it’s helpful to avoid breaking changes or unnecessary scope expansion.
The new models are great, but they do more by default, which means I’m finding myself explaining what _not_ to do more often than with previous models (where they’d often end too early).
In my case, the previous plan mode was too ephemeral, and I like having one source of “truth” that sits across context windows without loss/compaction.
To be honest the new behavior is convenient when I just say "X is broken, please fix". But I find myself more and more adding "please do not do anything yet, just tell me what you would do" and please feel free to ask if anything is unclear", especially if is making decisions that are a fork in the road.
I would really like a dry run mode that just disables all external commands from the outside so Claude doesn't proactively go about changing things.
I like plan mode personally. I only use claude code for the web, and the questions claude asks me to clarify are usually pretty important - mostly because I was too vague or contradictory in my prompt, or what I was asking for conflicted with something else in the code. I don't know how claude would resolve that without plan mode.
Also for session planning, as in when-can-I-walk-away-from-computer, its nice to know the particular rhythm of initial crunch - ask questions - make plan - do it. Especially with a 5 minute cache timeout.
what worked for me is a frontend driving the agent, capturing every user message, every commit, every pull request, putting them in a graph (uses a frontend because I didn't want to load a coding agent with tools + responsibility of book keeping) and the agent get tools to search reason behind code changes and see the high quality user input underneath instead of the sloppy self written justifications
My team has been struggling to understand whether or not we should do "spec-driven" development or not. It makes a lot of sense to me to have one developer iterate with the model/harness to generate a markddown document that is a high-level of what will be implemented, and then have the team PR review it before and agent attempts to do the actual implementation work. Do you think this is a good practice?
I specifically work in tooling now, so this probably applies more to that domain than some others, but I find 'very up-front spec-driven development' unappealing for that kind of work.
I'm experimenting just like everyone else, but this is my process right now:
- Quick prototype
- Figure out the language of your app (what terms you want to use for things, what your UI design language will be, etc) and spec that, so you can use words consistently with the agent. You need to be able to describe the things you want well and consistently.
- Keep prototyping. Let the agent write unit tests along the way. Lock down behaviour you like, keep track of those things in a document.
- At some point your idea of the real architecture comes into focus, from actual use cases -- avoids the over-abstracting right away trap.
- Refactoring is cheap with tests, so start refactoring into the architecture you want.
- Your architecture won't necessarily be what would be best for a human, but it will be pretty close.
- Keep relentlessly iterating on small work.
- Things that were expensive before aren't that expensive now -- integrating a library, changing from one library to another, trying out a few architectural refactors, trying out different performance optimizations, etc. That stuff is all 'throw it there and see what sticks' now, so don't be afraid to try stuff which felt big before.
I feel like 'front loading' too much is just the wrong approach. You might feel like you're sitting there 'babysitting the agent'; but that's just what the hard part of the work (hard as in 'zjust slogging through it', not as in 'conceptually complex') looks like now. Your code is much more like clay.
Atleast that's how I'm thinking about it so far, but I'm not working on large sprawling systems that I imagine would need more pre-planning.
I believe spec driven design is a good way to go. But that means keeping all your specs either in the repo (if you want the agent to be able to edit them) or available via for example MCP (if you think only humans should be able to edit them and you got some type of external documentation system keeping specs).
But given that running an agent us cheaper than the cost of waiting for a slot to assemble the team to talk about a change (isn’t it always?), why wait with running the agent?
I propose updating the spec then do the implementation. This will most likely show that a few assumptions were wrong forcing some major or minor updates to the spec. Work through those and then let your team review the spec change together with testing the next iteration of what what’s build.
I have found it doesn't see me any direct time. I either iterate ruthlessly on the spec or on the implementation.
I find that the code is generally in a better place proportionate to the amount of SDD I actually do. But it's just a matter of where and when I want to spend my time.
SDD is good. It's a lot faster and cheaper to have an agent polish a spec, than refactor code. The implementation can then be verified against the spec artifact, and any drift can feed back into process improvements for future specs.
I would have agreed a few months ago before opus 5 but I'm back to plan mode and lots of the old tricks trying to get that insolent model to follow instructions. 5.5 so far is better but certainly doesn't feel under my control (had it go off and read other repos and make changes just yesterday while in "plan mode").
I created a skill "feasibility", which is basically "evaluate feasibility of this idea and propose solution options". So I can check that Claude's idea of how to implement the requirement is close enough to my own before it starts working. It is a lot more lightweight than plan mode, because the skill says not to build a detailed plan, just a high-level summary.
What, are you suggesting that there were people already successfully doing agentic coding in CLI harnesses with a good number of features before November 2025, the moment HN collectively deciding that it was now "good enough", if using "Opus on Claude Code"? Blasphemy!
Yeah, as a user I came to the same conclusion as I naturally used plan mode less and less over time.
I still use plan mode in Astra to come up with a plan that I then feed into Fable. I feel like OpenAI models still do better big picture investigation and planning, while Claude is the better software engineer, if that makes any sense.
Of course this could well come down to my own biases and the specific things I’m working on.
I am not seeing this and I’m wondering why. Maybe how I plan is different then you? Maybe what we each mean by plan is different? Maybe you trained yourself to write more effective prompts than old you? Maybe the problems you’re solving don’t need planning and the problems I’m solving do? Maybe I’m pickier about solutions?
I find myself endlessly ctrl+c ing claude now as it flies off doing deep first principles analysis to work out how to find a thing it isn't sure about but I know the answer. Being able to give it that answer without needing to ctrl c would be a massive improvement
I always append something along the lines of “evaluate”, “investigate” or “report only” to my prompts when I want to see what the agent is gonna do. Because especially with the new models they tend to go easily off rail and do stuff I didn’t ask. To say that they just “get it” is highly dependent on the task, scope and blast radius.
I really wonder what I’m missing because my general rule to plan first unless explicitly instructed works perfectly fine with all models I use.
“I want to ...” / “Let’s ...” -> Plan
“Do X” -> Actually Act.
But then again I also have it configured to only ever answer questions instead of inferring them to be instructions (which I’ve seen others do differently).
Thanks for the context, and makes sense a lot. That’s my primary reason to use plan mode.
However how about decisions? Do we expect the model to read our minds, just assume the best practice will be followed and that’s what the user want? Plan mode solves those, what is that am I missing?
I frequently ask claude to write and review org-mode files. I've found it useful to maintain knowledge decisions both for claude and myself. Less so for other members of my team who wont use emacs or (n)vim. But thought I'd share.
I urge you to have a look at tools like tuicr and hunk. It’s a missing component in the chat. I really want a proper chat interface, leave comments on the code and have them replied to. No clue how to implant it tho, it’s a hard UX problem IMO
Recently I started using pi as a harness and its flexibility has really changed the way I think about UI into agents. It feels like I took the red pill and have been liberated from the prescriptive nature of CC or codex. Instead of getting told here’s how you should approach ___, I pay much more attention to my workflows and when I feel friction I make the harness adapt and smooth it out.
For example, there are no shortage of web based uis for pi and they are all cool but I wanted a deep integration between artifacts and how I want to collaborate with the agent on them. So i built my own ui that mirrors what the pi tui sees. It’s chat based but has a deep integration with a GitHub style code review UI so I can leave review and comments whenever I want. Every agent message renders nicely in an annotatable markdown viewer so I don’t need an ask question tool and can more naturally get the agent on the same page as me. I want it to feel like I’m working with a colleague.
I don’t think these features are too unique but having full control over the experience is really nice. Flexibility over model provider, can tailor it to my work’s dev stack, and don’t need to worry about anyone breaking it with a million updates everyday.
Your perfect workflow can be realized in a day or so. You just need to go make it happen.
I also didn't like the chat TUI for back and forth on things, especially on long markdown files where I felt like I needed to copy a line to comment on it. I wonder if LLM companies would rather skip a step and aim for a world where most people one-shot things, which sounds much more impressive.
I've been building https://crit.md to keep that back and forth with agents - GitHub-esque interface and have agents respond to my feedback, iterating until I'm happy.
Admittedly like many others, I use it a lot less for actual plans now, models are indeed getting really good at just getting it.
I wodner what this product space will look like a year from now. Reviews are already dying.
Oh I thought it blocked tool calls like editing files...
You don't need to actually change the tool schema or break the prompt cache to do that. In the tool itself you could just check if it's in plan mode and reject the tool call...
I keep forgetting that CC has access to the "question" tool, even outside plan mode, which is mostly why I still use it sometimes. This isn't (or wasn't until recently) the case for Codex.
Reading the replies, it's interesting how idiomatic everything is, eh? We must all opine about our specific setups, panes, prompts, and how they're the true way, or the true alternative to plan mode.
I don't really use plan mode, but I do have Fable write a lot of docs, particularly when the work involved is long and multistep, because I may need to do it across multiple sessions.
As a long time user of Claude Code, I've just naturally stopped using it because the model figures it out, and i prompt it accordingly like 'come up with a plan'. Glad to see your experience matches
that's the problem though:
- LLM now output good enough results without a plan.. for coding at least. I'm not saying amazing results..just good enough. it works fine.
- Most people suck at planning anyway
- LLM still don't give you a way to verify and understand to iterate.. you have to ask and then formulate and way so people barely do it, they just trust the vibe
IMO the current successor to plan mode should be the harness knowing when to tell the use "ok here's our overall current state in a simple diagram", auromatically
I think over time more and more will be peeled back to just the model and markdown. I have a beautiful factory running with key personas all it is is a few markdown files it is building a mac app fantastically well.
I m a bit surprised by this as - during the last 2 months - I 've encountered more often models going ahead of what I have asked from them. It's gotten so annoying that now instead of just saying "Let's plan X" I will also say, "Stop once you 've written the plan." because otherwise there is a distinct (~35% anecdotally) probability that the model will go ahead and implementing the plan as well. [This applies to both Anthropic and OpenAI models btw.]
Plan mode is a useful shortcut when I want to have an agent do read-only work without having to worry about giving it appropriate stop conditions.
Hi, thanks for Claude Code. I use it, and it works well. Have you considered changing it so the text comes down from the top of the screen, in green, like The Matrix?
I ask Codex (I am not using Anthropic anymore) to always generate a plan file first, unless explicitly told otherwise.
I also tell it to write deviations and rename plans accordingly once done.
That way I keep the codebases I have to or enjoy to work on in my head and don't become too dependant on any provider or on stochastic parrots in general.
except the fact that claude code in plan mode does destructive operations, requiring a hook to tell it not to do stupid things, so its not really a "read only" or "dont code yet" tool, imho.
Plan mode still earns its keep when the blast radius is high.
For a one-file change I don’t bother. For anything that touches auth, payments, or a shared schema I still want the plan written down first — not because the model can’t figure it out, but because I need a moment where I can still say “no” before it starts editing.
The mode was never really about making the model smarter. It was about making the human stop and look.
I'm actively watching understanding slip away from developers, code review getting paired down to no comment checkmarks, and codebases go to bloated messes that nobody can read. Axioms like engineers must understand and take responsibility for the code they ship are getting torn down, and the products coming out are reflecting conway's law, becoming impenetrably obtuse and always "so complex there are no obvious deficiencies" (as opposed to "so simple there are no obvious deficiencies" which used to be the aim).
The one thing plan mode helped is for the humans to get an understanding of the strategy, and be able to poke around and look at the design and architecture. You can achieve this with some self discipline and keeping shorter leashes on agents, but it feels like a losing battle. The best devs still put out good code, but the poor devs are learning nothing while their metrics look great. I can't help but think we are racking up immense amounts of debt that will very soon become due.
I’m a big proponent of writing simple, understandable code, so I’m playing devil’s advocate here a bit, but: who cares?
A significant (majority?) portion of developers have been shipping JavaScript/node applications for the last decade that contain hundreds of MB to GB of code from god knows where doing god knows what with dependency trees the size of redwoods. It’s not like your average mediocre dev really knew what was going on behind their gluing of frameworks together - at least from what I’ve seen.
If you have remotely competent tech leadership that enforces relatively intelligent patterns (a good one I’ve found is “write everything backend in rust”) you can make AI churn out monstrous amounts of code that… isn’t all that bad? And if you enforce it writing and updating a docs/API.md on every commit/PR you’re probably doing better than 80+% of devs I’ve ever met. Up until a few years ago it wasn’t uncommon to roll up to a new job that was a “legacy” pile of garbage concocted over 20+ years with no comments or API docs and a readme that tells you to ask for help from someone who has been dead for 5 years. At least AI code is full of comments (some of which might even be accurate) and there’s a finite (relatively low!) cost to figuring out “wtf is this doing and how is it doing it”
One is abstraction, one is complexity that _you_ own. Even 100 lines of trad-coded C relies on "hundreds of MB to GB of code from god knows where" in the Linux kernel. The difference is, you can perfectly understand that C, and own it. Then delegate ownership of the rest to Linus. If AI writes 1000 lines of C instead, now there's code no human owns in the world.
The longer you work in a large engineering organization; the more clear it becomes that no amount of processes, documentation, documentation management systems or training programs can actually transfer the institutional knowledge of the origination to new people other than people-to-people interactions.
Beyond a certain size, the documentation becomes too large to ingest. Below a certain size, it can only contain a fraction of what is needed. If you take any sufficiently engineering project, and give every engineer amnesia; the project will go to shit for a undetermined amount of time, as it takes months or years go build-back the understanding that was lost.
This is clear enough then large companies fire and replace workers randomly to cut costs; a worker that has built up useful knowledge in the origination over a few years is more valuable than three cheap consultants from "low-cost countries" that are fired when the work package is over.
---
AI, looses its memory every time we press "new thread". No spec can bring that back until AIs become able to write and ingest whole books of context without getting confused.
I completely agree with everything you’ve said, in the pre-AI world. Now, I’m not so sure.
When I worked at a big tech company on a large codebase (tens of millions of LOC across dozens of repos), it was extremely common to work on something in an area of the codebase I had zero familiarity with, and due to turnover, no one else at the company did either. As you say, docs were frequently missing or out of date.
However, with some effort over a couple hours, I could make a LOT of progress in understanding the history of the code. Every commit and PR were linked to JIRA tickets, most had eng design docs with comments, slack discussions, etc. I could step through the git history and watch the code change, alongside the artifacts of the human discussion and decisions that led to the changes. It wasn’t perfect, but I could make tremendous progress. Now, this was a remote company with pretty strong culture around using JIRA, design docs with review, etc. Probably the biggest gap was meeting transcripts.
Today, an agent can chew through years of history and artifacts on a large codebase in a half hour, documenting as it goes, and have way better understanding of it than I ever will.
It’s true that an agent can’t hold all of that in its head at once without context rot (though this is improving every year), but neither can any human!
If your goal is understanding how to do something, why something was done, why something wasn’t done, etc, and the codebase is large, mature, and extremely well-“artifacted”, I’m not at all convinced that you’re better off asking Bob who has worked on this section of the codebase for a decade than just asking a really good frontier-level agent. Maybe, but it feels like that won’t be true much longer.
> Today, an agent can chew through years of history and artifacts on a large codebase in a half hour, documenting as it goes, and have way better understanding of it than I ever will.
At my last job we worked in a large, but publicly available code base. My experience was that just giving an agent the prompt to look at area X to figure out how it works would use up half or more of the context window. And that was thrown away every time we started again. We had docs, agent generated overviews but the agents still filled their context windows way too quickly to actually really be any use.
I have a hard time finding concrete examples that does not breach contract of dox me.
I feel like this issue goes beyond a single large legacy codebase. It can go beyond how the code works into what the code does. How should a thing behave may itself be institutional knowledge only Bary that was involved in the design process still remembers.
Pointing an agent at a problem in one codebase to write a patch can work quite well. But what do i do when the problem is either in the tool, the database, or the two API layers between? You may not even have the code for half of it as it's part of another department or externally purchased solution.
Bary would know, or know who to talk to, or dig up a PowerPoint slide he's been working on to explain it.
Exactly. As someone much smarter than me once argued here on HN, the value of a software company is less in the software but much more in the mental map the team has built of the problem and of how the software solves it. Throw the team / mental map away and you'll have a very hard time maintaining the software.
At the same time if the organisation want code fast, it gets code fast.
The revenue for a new feature today is something sure. While the cost associated with supporting such feature will be up to debate in the coming quarters.
As often it is the case, we are moving on a long vs short term trade-off space. And I don't think any experience will generalize
I'm not saying we can get rid of humans (yet). I'm saying that it's not critical to write and review the code yourself anymore, and that the LLM is the new layer of abstraction through which code is written.
The spec will never be able to capture all the edge cases. The code is what runs. It also doesn't capture all the edge cases (that's why code has bugs) but it does a better job at that than documentation.
Documentation can also become stale easily
I mostly agree with you, but for “Documentation can also become stale easily” … historically I’d 100% agree with you—I often forgot to update separate documentation files, forgot to update comments on related functions, etc—but AI is so much better at automatically catching and fixing this without prompting than I ever was.
That's true, but it's also prone to over documenting. As an example,
for some reason it always tries to include a diagram of the folder structure in the project. I think that's not really useful and it's something that changes so often it's not worth documenting
This, and maintaining the spec is a cat-and-mouse game precisely to cover those edge cases.
Like "if you're using this database, and the engine has these configurations, then do _x_, unless _x'_ and _y'_ are enabled, in which case, do _y_..."
Which, if you're already being THAT specific in your spec, you might as well, idk, write the code yourself?
Because at that point your human language is basically the code and AI is the compiler. A non-deterministic one.
Regardless, your business stakeholders won't understand what's going on anyway (nor should they), so we're back at square zero.
(I wrote Technical and Functional Requirements Documents as a business analyst in college. What's happening now _for most situations_ is more or less the same thing.)
I think the main problem with this statement is that different facets of an entity are abstracted at different rates.
AI abstracts effort and cognitive load away from code at a heafty rate, but it doesn’t abstract liability away from code at all.
My business is paid to produce artefacts for which it has liability in the case of error, so we need to do additional work to mitigate and eliminate the liability risk introduced with language models. So far I’ve not found a better way to do that than a plan/act/assert type approach on every feature.
What happens when the maintainers lose access to frontier models because of cost, politics or other external factors? What happens if they don't have enough hardware to spin up an open-weights model?
I've seen variants of this play out before AI, so I can tell you: they will inherit a codebase they've never seen before, take forever to ship fixes (forget new features), and they'll either scrap it, completely rewrite it, or, if they're "enterprise" enough, will pay consulting companies literal mountains of cash to make it their problem.
The whole point of writing simple code was to write code that other humans can maintain. If AI is here to stay and becomes economical enough for everyone to use it, then you're right; writing code for other humans is no longer useful. If that doesn't happen though...developers who can/want to still code by hand will be loving life
Then all the more reason to support open weights and the improvements going into them, currently state-sponsored but available to all.
If frontier AI is load-bearing, be thankful you're able to spin up comparable models (face it — 99% of us aren't writing novel software) for a fraction of the cost it would be for a third-party to come in and mop up the mess.
> hundreds of MB to GB of code from god knows where doing god knows what
If you use established libraries then actually the code IS well known to someone (and likely many), even if that's not you. Likely it was built with an actual purpose and with the foresight to not add red herrings to the design.
You can't say any of that for the equivalent amount generated lines. Literally no one knows what it does.
moreover there is not a black and white "code you can read" and "code you cannot read", there is a spectrum between. In this era, IMO if you know which module do what kind of function it should be enough, you don't need to have deep understanding of the file.
When slop code needs to do something different or new or in a different environment but still stay working on the original things, it’s hard to change, sometimes catastrophically so. When your code has concequences, quality means more people understand it which means more people can maintain it.
Arguments against it are sort of like why have devs on staff at all when you build the thing the first time, or why not outsource everything, or why should I care what my code looks like when the code seems to work?
The cost of tokens is not zero, and the bigger the thing you’re doing the more low quality will cost. When your company is the software, you take on an existential risk based on the software working or not.
Pure AI generated code without human curation is full of bad wordy comments but those comments can mislead, be stale, contain duplicates, and drastically reduce the ability to understand things. I watch teams that still care about code understanding ship good products while those vibe coding in the same org just flounder after the initial burst of features. Some problems only come up after the first 90%, and AI can help you solve them but a big ball of spaghetti is still a big ball of spaghetti.
Seems to me that if the AI writes the code, then AI can easily copy the code.
I.e. how hard is it to point an AI at a piece of software and say "AI, copy this"?
Seems like sooner or later copying just becomes a matter of spending enough on tokens.
Seems in that world, all significant software projects get copied. That turns software into a commodity loss leader for other business models or an open source project. Similar to the way Chrome works for Google and the way Firefox works.
Long term, that also just seems like a matter of spending enough tokens. I.e. we have these "red team" AIs that are finding security bugs in very mature software. As AI improves it seems you will be able to just point them at your copy, spend enough tokens and get most of the bugs.
This has been the experience of everyone making decisions in any company without being the one doing the technical work. It's not a novel concept. It's actually the opposite, compared to technical people running companies.
Well they had actual people who understood the code - because they wrote it and had a mental model pretty deep. Now the people owning the code don't understand it.
Won't you be worried if your mechanic didn't understand your car but offloaded it to a robot that made mistakes all the time?
In many places it's demanded by upper management that devs use AI.. so even if engineers wanted to avoid using it, they would have to meet their quotas.
This is what happens when executives suffer from AI psychosis. They were already impatient, now with AI all they care about is feature velocity.
The faster they can hit that refresh button to see the features, the quicker sales can close the deals for them.
I am reminded of the cartoons where the car kept going faster and faster, the driver kept pushing on the pedal, parts started to fly out, the gauges started to go in circles, smoke started to billow out of everywhere and then boom!
I do try to learn how systems work and remove "abstractions" as much as I can. I do that with coding and my personal life.
Actually baking is a good example, I used to be really bad so I spent time learning. I don't do it every day but now I understand how bread is made.
I bought a 3D printer so I could print parts to fix stuff myself.
I learned to do my own oil changes, I learned how engines work, etc.
My point is that I try to learn more, not less, which is what AI is trying to achieve
I feel the same way but I have successfully refactored some of the early experiments. Our team has settled on targeting a double output from the before times but more ambitious product vision because AI can teach us things we don't know. We actually target 2 days of coding and 3 days of learning with Ai so the increased efficiency allows upskilling rather than just pushing more code
> The best devs still put out good code, but the poor devs are learning nothing while their metrics look great. I can't help but think we are racking up immense amounts of debt that will very soon become due.
I agree with this, but the reality is that it's only the result of models empowering devs, and power in good hands amplifies positive results while power in mediocre hands amplifies technical debt.
It's a good time to choose wisely who you work with.
> It's a good time to choose wisely who you work with.
Very true, but this also makes me think what kind of ridiculous obstacle course future hiring process would look like.
In a land where anyone with a pulse can prompt AI to make an app for them - how would future hiring managers and team leads figure out who will drag codebase down with tech debt and who wouldn't?
Probably the same way they have for the last 30 years: poorly, in a dozen different ways, depending on what that particular hiring manager thinks has correlated with successful hiring in the past
I can imagine a code review where you're asked to implement and merge 10 PR in a sample codebase and the codebase is littered with the sorts of mistakes and slop that vibe coders put in when they're careless and you're asked to correct the mess and make it work.
Future hiring will select for people that shipped the most, plain and simple.
The tech debt concerns are much ado about nothing. Use the next model to clean it up, big deal. Code is cheap.
The people that sat around handwringing about tech debt and trying to read every line of LLM code will really struggle to find a job. The profession fundamentally changed, and these people did not catch up.
Of course you shouldn't evaluate people by tokens, you should evaluate them by what they ship. Did it matter to the business? Did it ship fast? Did they build what was needed?
> The people that sat around handwringing about tech debt and trying to read every line of LLM code will really struggle to find a job. The profession fundamentally changed, and these people did not catch up.
I'm inclined to agree - the latest models seem to have crossed a threshold where trying to review by hand becomes akin to a manager insisting on reviewing PRs.
Just like rushing made messes in the before times so too does rushing via LLM. The exact same outcome will happen, but at a far greater velocity and scale than anything we've seen in this industry before. Old Testament, Mr. Mayor, real wrath-of-God type stuff!
It's truly bizarre to read about all these people who just give up on any understanding about what they are working on.
I very regularly use plan mode not to even make a plan of action itself, but to better understand what possible issues might come up when implementing some feature or fixing some bug. And it is quite common for me to fix or rewrite certain findings that AI comes up because its assumptions are not quite right or don't align with overall goal.
And yet so many seem to be perfectly fine leaving all the decisions to AI - even if it's going in the wrong direction. I suppose that's all the people who got into software purely for money or status - never really caring about the actual thing they are working on.
The last couple of weeks, I’ve got rigorous about making the agent refactor old code. We’ve removed, moved, consolidated, reordered loads of cruft. It has made the code much tidier and reduced the chance that the next feature will build on top of cruft.
E.G. over time we’d gained two client-side caches of related server state. This started out as two different parts of the same model, because we couldn’t get all the data we needed from one microservice and had to merge in the client. Over time, more and more features used both caches for different aspects of related processes. At some point one of the microservices changed so as to return all the data in one call. The update to consume that kept both caches, adding code to sync them, because so many parts of the code were using one as a fallback for the other, so they both looked “necessary”. Because they were separate, and “live” sometimes they’d go out of sync after the initial load. Worse: the consumers alternated about which cache was treated as the fallback, making it very hard to see that either might be redundant. Eventually I noticed they were filled by the response to a single call. We all know paying back tech debt never gets prioritised, so I rolled the payback into two feature tasks, and just took longer about them.
My employer expects we use LLMs and provides some budget, but it’s not enough to use even Open4.7 or GLM-5.2 on every task. I do the bulk of my work with Composer 2.5. It’s quite good for “going forward” on smallish tasks and it’s written most of my code this year. It’s possible smarter models would spot these refactorinh opportunities and action them proir to building features or fixing bugs. But I wouldn’t know because I can’t afford it. I’ve never seen even a 4.8 era model spot a refactor and plan to do it prior to a “new build”.
I’m pleased I’ve spotted these trends and started to build the habit of (telling the agent to)“refactor to make the change easier”, but my percieved productivity will go down and I risk the ire of my leaders.
FWIW, our codebase is growing, and the size of each change is also growing, but it's because AI is making us fix all the bugs we'd previously check in because our code long ago surpassed what even our best developers can reason about.
The funny thing is that the AI adopters are in the middle of the bell curve. Our worst devs continue to perform worse than AI yet refuse to use it and our best devs continue to insist AI sucks despite it finding issues in their code and the reviews and designs they've approved.
Sure and sometimes that is what happens. Sometimes it's not. We should be doing a lot of things, but have to triage issues and act pragmatically.
In the case of what I'm currently working on, filing bugs for every issue I found, and then factoring out each fix, and then running each change through the 8 hour ci/cd system, and hoping an unrelated issue doesn't get misattributed to me... No, I'd rather just wrap it up into one coherent refactoring change and be done with it because when I'm done there are several more like it waiting for my attention.
I don't know your exact situation; and there are sometimes genuinely valid reasons for an 8-hour CI/CD system; but ... man, reducing that down to like 20 minutes (which is possible and often common) would pay as many dividends as all the AI stuff that's been added. Man, like ... holy crap @_@
We're uh.. we're a well known and widely used piece of software. There is an insanely large amount of testing that needs to go in to each change across many platforms and configurations and we have a lot of devs hammering that system with changes to test. Even worse now because of AI...
If your company is really interested in doing AI stuff you may be able to sell people on the idea of a "fast agentic CI" where you:
1. Track code coverage from all current test cases in your big 8hr runs.
2. Index the coverage so it goes testcase name -> methods touched.
3. On every commit run an agent which looks at the diffs, tries to predict which tests the code being edited is touching, and run just those tests (or a random subset if there are too many).
Once you have this you can also automatically kick off builds every 8hr that contain all the previously submitted patches. Once that is done, if a test begins to fail, it can automatically bisect the history to find change, and notify the author.
You could also get these benefits with a Bazel like build system which can cache test executions so others don't need to rerun them if the binary is not changed.
I see this complaint frequently about losing track of what the agents are doing, and I agree you do need to understand your system. But there seems to be this baked in assumption that if you lose track, you now need to manually wade through this massive mess to untangle it and maybe that is impossible. I don't agree.
If you don't understand the codebase, ask the agent to explain it to you. I'm not not kidding. Modern frontier models are fantastic as this - even more so than actually writing the code. It can tell you in words. It can generate architectural diagrams and sequence diagrams. It can write tests and scripts that prove it's assumptions. It can happily refactor so that the system design is aligned with your preferences.
Once you accept this, you can stop worrying so much about it and instead focusing on building the architectures and tools that lets the agents succeed better and faster - so called closed loops or agents prompting agents. Build systems that are more easily verifiable and deterministic so the agent can write very powerful property based tests. Focus more on what and why you are building, how to make sure all external properties are verifiable and leave the internals to the agents. The code is not really for us anymore.
In my experience it's really good at making you feel that you understand things.
Then when you actually dig into the code, there are many things that are not like you'd expect.
When you've experienced that a few times, you stop trusting that the agent gives you the full picture - for good reason.
When I review AI generated code I generally find so many flaws that it makes it hard for me to believe that those who are not reviewing their output are not just fooling themselves. Maybe not all the time, but quite often.
One such recent example was an SSO simulator for a local env. Instead of using a cookie to remember who was logged in, the agent remembered the last log in a variable, assuming the the next requests would come from that login.
This snowballed into our tests, where later agents had created helper tools for working around the SSO simulators statefulness.
Things improve drastically however if you spin up a second session and ask it to adversarially review everything that the first session produces (this goes for everything: not just code, but also design, planning, and explanations).
This works even better if you use models from different families to do so.
I have, on a lark, reached 20 to 30 adversarial sessions a few times. For some tasks, the sessions will just never converge on anything that yet another session won't find fault with, recommending an alternative already ruled out by another session. Even if all the reasoning in between was documented, the new sessions will endlessly claim to find flaws in past reasoning.
My experience is this is great when the model surfaces something to you. But I'm constantly caught off-guard by things the model didn't volunteer, things I would have quickly stumbled upon if I was working on code the traditional way. The model didn't think it was relevant but I sure do.
In my experience having the agent explain the code doesn’t work very well for real world apps, even the parts written by humans. For example we tried using to generate diagrams, class hierarchies, etc as part of documentation. If you don’t know the code it looks great. If you do, it’s focusing on all the wrong things, missing the mental model, and ignoring lots of important bits. And Claude tends to be extremely verbose to the point of muddling things.
"The one thing plan mode helped is for the humans to get an understanding of the strategy"
For me, this phase still happens, but a distinct "plan mode" is unnecessary: I just tell the model, "This is discussion; no code changes yet." and spend hours figuring out what will and will not be done.
To be fair, it has always been like this. People actively became dependent on the Internet, on StackOverflow, fast Internet connection, cloud services, etc.
At a company I had setup scripts to build our packages, and the CI was running those scripts. Someone more junior (only a few years, not decades) found it strictly superior to remove my scripts and replace them with GitHub actions: people could now know even less about it (as in, no need to know how to copy-paste and adapt the recipe for a new package), but now it depended on GitHub. GitHub is down, nobody can build anything anymore. And it happened once every few weeks, so people would just go have a coffee during the outage.
You know what happened next? That person got promoted for their good work. That was before AI.
Unfortunately what's happening is people just can't help themselves. Just like an addict reaching for another hit, it's really difficult to make yourself do work when you could just not. I've said it before but LLMs are our cigarettes. It's going to get a lot worse before it gets better.
These comments are so bizarre when we are what, 1-2 years away from AGI?
Like - you really think models won't be able to clean up the tech debt they created!? They are very good at this already. Ask Opus 5.5 to clean up the tech debt from some Opus 4.6 vibe coded app.
Code is cheap now. The most important thing is to ship, ship, ship. If you are handwringing over "tech debt" you have already lost - and you deeply misunderstand how good this technology is getting!
We get it, the future can't come fast enough for you. Code is a commodity and the only important thing is scale. Congratulations on leapfrogging the midwits on the path to the C-suite!
Did I say anything about tech debt? What's the point of "ship, ship, ship" if AI will be able to do it in a couple of years? Don't you see where this is going? We're rapidly losing our ability to think.
The point of "ship, ship, ship" is to cash in (in terms of your career or company) before AGI renders you jobless.
If we get to post-scarcity you never needed the money. But if we end up in some dystopian hellhole where people have no jobs but capitalism still exists, you'll be thankful.
The fundamentals of writing correct code have become dramatically cheaper by orders of magnitude. I have been involved in hiring dozens of developers. Fable/Opus 5.5 would be a superior drop-in replacement for perhaps 75% of those hires at roughly 1/100 the cost.
Perhaps for writing boilerplate, or generic websites from scratch, or early stage prototypes.
I work in embedded systems where it's harder to offload the burden of testing and maintenance. Systems knowledge is important: experienced developers can be costly to replace.
LLMs make a lot of things easier, but costs have certainly not fallen by 2 orders of magnitude.
How many people on HN work on mission critical systems? The discussion is not currently constrained to that. I agree, but the exception proves the rule.
Nothing, the people worrying about AI tech debt on HN have been worrying about it for years. 2024 HN threads had the same comments.
The reality is that models just keep getting better and are very good at cleaning up the debt they created. The "tech debt" bill never came due. It won't.
Until some catastrophic data loss event or data leak that nobody understands or has any mitigations for. Oops! Guess bankruptcy and prison time are the ultimate uncaught exceptions.
I remember this being said as far back as 2022 yet we're still here fixing things manually when the models inevitably code themselves into a hairball convoluted enough they can't find a way out.
We cannot blame just LLM models, it is brains nature to save energy. If agent did tasks consistently good, our brain try to delegate cognitive load to the model to save energy. After consistent use of LLMs anyone can have tendency commit slop just reviewing at high level, this is specially true with busy lifestyle. Also AI generated code do not give dopamine just like solving problems we did before LLMs, we tend to get lazy. Strict discipline is necessary to make good use of LLMs in order to not commit slop and not to make us dumb.
When I draft my idea for the implementation of a feature or bug fix, I don't even trust a _human_ to understand what I mean the first time. There are _always_ either errors on my part, or erroneous assumptions on theirs. Everything from "this accounts for X and Y, but not Z which breaks the whole thing" to "this part of the idea directly contradicts what with you said earlier, what do you want to do about it?"
I can't bring myself to trust that an LLM understands what I mean better than any human would, no matter how "good" people claim they are getting.
TFA seems to be advocating for regular old vibecoding. Code now and ask questions later. Which is their choice, and is perhaps even a valid choice in many cases. But at least call it what it is.
Plan mode, for me, is my opportunity to develop and understand my own plan. Sometimes, rarely, Claude demonstrates it misunderstood my intentions by writing a plan to address the wrong problem. But usually it's about me fleshing out the scope and boundary of the intended changes.
I find that faster than code first, ask questions later. But it takes more time up front.
In the before times, and sometimes now even, the way I always coded was by thinking/discussing the work, then hacking up various interesting bits, then throwing it away and doing it the "right" way. Plan mode makes it harder to go back and forth between "hacking" and "thinking" phases.
Coding with plan mode is still vibe coding. No-one outside HN calls it things like "agentic engineering" or whatever. Everyone just says they are vibe coding regardless of how hands on they are.
Whether you use plan mode or not is orthogonal to whether or not you're vibe coding. "Vibe coding" means that you don't read or edit the code; you are running the software and iterating on vibes alone. That's the meaning since the original (and very recent!) coinage and hasn't changed.
If you're reading and editing the code, you're not vibe coding. If you're not reading and editing the code, you're probably vibe coding even if you feel "hands on".
The real reason why plan mode is dead is because you can just conversationally instruct the agent to not make changes to the repository or to make changes to selected documents only, and it will listen. There was a time when we needed to enforce this via selected tool use, but we have surpassed that.
I agree. Plan mode came about because earlier models were loose cannons, doing what they pleased. Today's models follow instructions better enough to not need a separate mode. However, there is still value in using a separate, smarter model for planning than execution, and persisting it for auditing on completion.
I typically converse with the default model to point the plan in the right direction, then have it iterate with a smarter reviewer to find flaws until the plan file is converged.
Plan mode does have different permission controls though? Its difficult for me to parse the docs on this but it does seem to imply more than just a prompt although less than a full sandbox. https://code.claude.com/docs/en/permission-modes#analyze-bef...
The way it worked originally was it would trigger a different system prompt, and then the permissions system would prompt you (even on YOLO mode) before running any non-readonly actions.
Cursor would also write a plan document, which was useful when working on larger tasks (due to context size). I still find that part useful today.
Plan forced the model to write down and outline what it was going to do before writing the code. As Boris said up thread, he created it to prevent it from just rushing into writing code. So that regard, it did act like a permission system.
It helps to think more abstractly when approaching problems like this.
It also doesn't allow leaving an audit trail of plans and decisions (by default, anyway). Most of my mutating prompts look like "Propose a plan for change X and write to file Y" and "Execute steps M-N from file Y".
RE the article: I don't think it's obvious why this process is worth following until you find your time and attention wasted. Conversationally-building is the express train to waste. I'm not sure why you would even be talking to claude if you don't understand what you want to build.
Do you really trust it with a production code base or database? Telling it not to make changes feels an awful lot like "Make no mistakes". I also like how Plan mode on Codex asks clarifying questions. I'm in management now so I've used Work more than Code lately.
Anecdotally, in the Opus 4.6 days, it felt like there was something special about using plan mode to discover the approach then clearing the context to execute on it.
A mixture of defending against a disastrous mid-implementation compaction (where suddenly things would veer off the rails) and also allowing the fresh execution to double-check the assumptions and notice any subtle mistakes before context was poisoned.
I’ve found that for large enough changes I still prefer having a parent theorizing about the root cause of issues based on evidence and then dispatching targeted child sessions to fixed based on theories and concrete telemetry examples.
There’s something clean about having sandboxed context and a session you can quiz about architecture while one is heads-down working against a spec.
For me, it was the agent that decided about plan mode, when that happened. Most of the time I didn't invoke it but kept discussing and tweaking and if the context was close to full, make it dump it out to disk and start a new session. I felt like that worked more flexibly for me because I could keep steering and adjusting until it was what I wanted. At that point I made it write a detailed spec. Like my homegrown plan mode.
For me he plan mode is as useful as ever, even though I understand this is just a prompt. For large changes there are few problems exist such as context window limit and compaction, and even though models got better these are still detrimental to models performance. I found making one large high level plan and then splitting it into N plans grouped around distinct areas and then running N implementations if possible in parallel, or some consecutively, improves model "vision" far ahead. The workflow author suggests might be more useful for prototyping, but there are still types of work that are better done with massive pre-planning stages. And of course running exploratory prompt a.k.a. plan is useful even for smaller types of changes like small features or bugfixes.
I found the article interesting in that the author didn't baulk at receiving an unchecked uncheck-able bulk of code. If this is what is supposed to ship, am I just old-fashioned to think this should not be acceptable? There is no way I would ship code that a person had not read, let alone written.
A quick straw poll. Are most people here who use AI to code well-versed in their languages/software development? i.e. 10+ years experience doing it "by hand"?
I think in ten years time there will be no developers with that 10 years experience behind them.
I did everything by hand for so long and miss it, now I just read thousands of lines day in day out. But there are interesting gains such as lowered barrier to quickly run adhoc perf tests, do some formal verification, you can just add more guardrails for the thing that gets shipped. The thing I hate though is when managers sell their generated code and you are left shoehorning it to production, they take away the interesting part of the job.
Completely agree with your first paragraph and a point I've made repeatedly here is that I want agents to work more like pair programming and less like code review. I want to delegate typing, not thinking.
But what I've noticed is that lots of people don't want that. They're happy to delegate both typing and thinking.
There are folks who still want to think for themselves of course, but model providers are incentivised to maximise token spend over all else, so the "official" harnesses are not aligned with their needs. My message to those folks is: _go and build your own harnesses_.
I built one, it wasn't a huge lift and it completely changed how I work with LLMs. One consequence is that I choose smaller, cheaper models now because latency is more important with shorter feedback loops. I don't want a super-intelligent model that disappears for ages and provides a final answer, my own brain thinks and guides the process instead.
I'm deliberately not linking to my harness here because that's not the point. Think about what modes of work you want from a coding harness, then build one that works exactly like that. Immense happiness and satisfaction is just a couple of weekends away.
I'm probably just behind the curve, but I still use plan mode quite a bit in Claude Code. I iterate on the plan repeatedly until it largely agrees with what I want. Once it seems good I ask it to build the plan and open a PR when done.
My only critic of the plan mode is I wish it was easier to see the updates and changes easily in Claude Code as we iterate on the plan. It is wasteful to have to remember what parts I have reviewed and what parts are new (and need another pass). I have thought about fixing this but I also feel the review is the actual thinking (even if ineficient), and so I purposely have not removed it.
For an adhd mind that serial interface is a bless and hell at the same time. Great for the detailed planning discussion, bad for checking the results.
I am currently working on canvas-based interfaces for that reason and i would think the only way to really create value here is with a deep independent analysis and visualization of the changes afterwards to reach some ease of mind. Live would be cool (if you like that)
When it comes to planning itself, I recently tried the token-saving planning plus phases execution agents approach and had to find out that agents actually don't necessarily communicate better by prose-reduced specs than we do.
I had to go back to the planning agent to implement or fix things with our full planning context in mind. So if you want really high control for a "tight" implementation, I'd say just sharing plans is not enough. The probability of things getting filled in by the executing agents rises and you either find yourself holding those agents' hands or fixing things afterwards.
Actually phases are still to large and you would actually want the planning agent to hold that hand all the time, meaning small context is not the way to go, as you might need the full checking context much more often than current phases sizes suggested by the planning "doc" would use it.
Distributed building still is the way to go though, steady control by the overall context or one specifically thinned out for the particular job is. But don't go prose-based plans anymore. These are dead indeed.
Disagree. While it's easier to 1-shot more things now, when there's more complicated interaction of components, a good plan from plan mode can make the project a "looks good, go" and run overnight thing vs. "steering" needed.
How many thousands of dollars are you burning a month on “looks good, go” and then *letting it run overnight*?? I spend enough money on 8 human, interruptible and human-thinking hours with paused hours.
I can’t imagine how high that number would go if I got rid of the collaborative style and also the “please don’t run scary commands on random directories without permission” mode.
Write a good high level plan with functional and non-functional requirements and it works out costing less, works more efficiently, and no surprises.
It’s like this: If you know the right words to say to the LLM, you’ll likely get back the “right words” also. And the more right words up front, the less steering you need to do.
(Having 20yrs of experience writing these user stories and designs help, my “unfair” advantage)
I still rely heavily on plan modes, first to think through product design from the 3-sentence descriptions handed to me as super-complex feature requests, and second to implement the technical requirements I want to build.
Way before AI, I think the first idea, while sparked by inspiration, has always created a local minimum and suggested that something could be done around a problem, but it was rarely the best idea. I think the speed at which the models can turn a bad idea into code will lead to bad, design across systems and interfaces.
I am also very pedantic about how the data models get implemented, and I want to be sure that things are built in the way I want them to. I'm also very pedantic about how specific if statements and dicts are created, and again, the models make it easy to do things a shitty way. These are hard-learned lessons of several decades of shooting myself in the foot and locking myself into less-than-ideal design patterns.
I spent two days going through designing and planning for the LLM to build the actual feature in less than 2 hrs; once it's set, it's lost. It has made testing go very quickly and i am deploying a very complex feature and high risk feature in a 4 days when without the models it would have taken probably 4 weeks to get right.
Lately, I've been using Matt Pocock's “/grill-me” skill more and more, rather than plan mode or a complex, spec-driven set of skills. I made a personal fork of the skill to use Claude’s ask tool, which has a nicer UX.
I used plan mode for two reasons: to review the choices before execution, and to execute with another model (i.e., using the barely documented opusplan feature).
The grill-me skill is much better for reviewing and clarifying choices (and modifying it to use the ask tool makes you go faster). Instead of opusplan, you can explicitly tell Claude to start a subagent with another model to divide the tasks.
Unless I'm going to just do something, I use /wayfinder for almost everything now. It automatically goes into grill mode when appropriate, and it'll suggest just doing the work if it's light enough.
Firstly I find it’s an excellent way for me to get very good clarity about what will be built and whether it’s going to be done in a sensible way.
Very often I don’t really know what the work will need to look like until I’ve explored the problem with the LLM towards first making the plan.
Without a plan I find myself having to do the initial understanding through code review of its generated code which is much harder than reviewing a plan, and then I invariably need the LLM to fix up what it did which is much slower when it’s doing code than working on a plan, never mind the next review I need to do.
And when the plan is good enough, I clear the context before telling it do it, which I’ve found vastly improves the quality of the LLM output.
For me the most useful thing about plan mode in Claude or Codex is not the plan, I usually barely read it, it's that it forces models to gather as much context as possible before implementing. Without plan mode models tend to implement wrong solutions because they are reading not enough context before writing code.
Another useful property is the readonly nature. I can easily let multiple agents plan in parallel without having to worry about annoying worktrees or conflicts and then I can come back to the plans later.
Of course this can be done with just another prompt, but that's exactly what plan mode is. It's nothing more than a predefined prompt in the harnesses with maybe some extra guardrails (that don't always work)
Using an LLM for any large project shows how useful it is to have data and functions that aren't siloed. That's why CLI apps have had a resurgence: because the LLM can interface with them. Yet the reaction from so many who are deep into LLM development is to build their own little silo: an app.
We really need a better model. One alternative is to have an everything-app: a general purpose tool in which (almost) everything lives. The terminal is one of them. The text editor / word processor is another. (I use Emacs for everything.) In a business context the spreadsheet is probably the best choice.
Claude code’s plan mode began its slow march towards deprecation when they hid the “clear context and implement” option behind a flag. From the discussion at that time it seemed that the developers considered it mostly a legacy feature, and that newer models are smart enough in long context not to need it.
This part I find a little disappointing. The plan is the perfect boundary to cut context which still isn’t free, even if models can handle a lot of it now.
This is why open harnesses and open models will always be superior, it's only a matter of time until someone decides that something you use isn't worth maintaining anymore. The reply from the Anthropic employee up above doesn't give me much confidence that Claude users will be able to continue using plan mode forever.
Codex and Claude Code are the best harnesses out there by a long shot. They work out of the box, require no real config, and just do the thing.
The flexibility and lightweight nature of open harnesses is mainly useful for open models. These models are considerably far behind the frontier and need a lot more steering. Complex harnesses confuse them. I use open harnesses for my OSS model rig because the model needs it.
> Codex and Claude Code are the best harnesses out there by a long shot. They work out of the box, require no real config, and just do the thing.
Even if I agreed, that doesn't address my point: Claude Code might be the world's best harness right now, but that doesn't stop an Anthropic employee coming along tomorrow and taking a giant steamy dump on the thing you like most about it. You're using their workflow, not your own.
Sure, but their workflow is the best for Claude, and it's the same with Codex for OpenAI's models. These companies work and live with their own harnesses, they are ahead of the curve on what works for their models and what doesn't.
Plan mode is still very useful to me, I just need to read the plan. Even if the model is smarter than me, it’s still working on my project, and thus I need to keep a mental view of the system. My solution has been to ask the assistant for far less than it can do and leave for myself a bigger bunch of the work, so that I’m forced to keep myself up to date on the project. Yeah I’m leaving productivity on the table but… am I really?
I once tried to build something similar to your product on top of Codex, and I failed.
But my conclusion on plan mode is slightly different. I agree that plan mode itself may be a dead end, but I still believe there might be another way to achieve the same goal.
When I was building my product, I found that the biggest issue wasn't capability, but taste. The agent could build something that worked, but it often wasn't what I actually wanted. And behind that "taste" is a huge amount of implicit context — preferences, past decisions, product intuition, and trade-offs that live in my head. Distilling all of that into context takes a lot of effort, and I suspect giving it all to a single agent may eventually become overwhelming.
I've been wondering whether a better approach is to have multiple agents with different roles, prompts, and perspectives, and find a way for them to work together efficiently.
It's still just a hypothesis though. There are a lot of complicated coordination problems to figure out, and I don't have the answer yet.
For me, identifying a problem of "taste" doesn't exactly break it down. The problem is underspecified in some way. Can we be more specific?
The sense I get from this discussion is that the models/harnesses do not elicit feedback well. Where there is ambiguity, they tend to pick a solution and call it good.
A planning step aims to make these choices explicit. An iterative process is necessary to capture the detail.
It is natural to look to teams of agents to satisfy that process, but do they know where the decision points are?
So planning is tied to the spec. This much is clear.
What the author I think is hinting at is not planning alone but "the development and evolution of any program and the state of this program throughout the planning, elaboration, and eventual runtime".
I chuckle at the thought that throwing an md file or a prompt at this problem is sufficient.
So, I posit that if we want any agentic code to evolve meaningfully in the short future and over the long run, we have to have a system which holds and presents this information, the state of a program, in a coherent manner to a human operator. No other way. No other way. And this I say to both nay and yay sayers.
You can argue also that this is part of an even bigger thing. But it is not part of the current discussion on planning and speccing in agentic systems.
OpenAI and anthropic can throw all the billions they don't have at this and adjacent issues, but if this is not solved then they don't have anything.
For work, I primarily use Gemini CLI, and the one real awkward part of plan mode is that the harness can't run commands, even exploratory ones. So when I'm giving the harness a task in plan mode, it can't experiment and is either blind or relying on its training data and/or any other context (e.g. source code, documentation) I give it. And it often gets things wrong operating that way (depending on how detailed the context ends up being).
For example, the ability for the harness to call into a Python one-liner just to experiment is pretty powerful. If I'm asking it to use D2 to build an SVG graph, it can write some Python to introspect the XML to see if things appear to be placed correctly (size, x-y coords, etc.). Which is a pretty cheap way for it to experiment and verify its results before I deign to examine the rendered SVG with my own eye balls.
I use Claude's Plan mode daily and it's great. I almost always have feedback to refine the plan and I want a clean separation of planning time before writing code. I don't get what the problem is.
I also find this "kill Plan mode" push on Twitter odd, because developers have been complaining about AI supposedly killing their jobs, yet they want to take away the main feature that lets them be an active collaborator and participator in the process. Weird.
I also don’t get what the problem is. Claude’s plan mode is great, especially when you add interviewing into the equation. It means I can get to a point where I’m confident all the changes proposed make sense, and that I understand them, before anything is actually changed.
I feel like this post is hawking for commercial reasons more than it’s based in reality, especially a reality that takes into account the extremely varied experiences engineers seem to have with agents. Not to mention it ignores capabilities that, for example, Claude’s VS Code addin have had for ages (specifically select and annotate in plan mode, as I mentioned in another comment).
As such I’m not inclined to take it very seriously. More than that, the “X is dead” trope was overdone 10 years ago and I don’t think it’s been long enough to warrant a resurgence.
I see what you’re saying. That’s actually why I built Nuanced (https://www.nuanced.dev), I wanted a better way for humans and agents to work through intent before the code got written.
but I think I overfit the interface to how I used to work without agents. when I was at GitHub, I wrote a lot of ADRs, RFCs, and design docs, and I liked that because writing is how I clarify my own thinking. with agents though, I’m often not doing the writing myself. I give the model a rough intent and it fills in a bunch of gaps, and then I get back a long, polished plan containing decisions I didn’t explicitly make.
That’s the part that feels broken to me. The plan can be detailed and technically correct, but still be hard to review because the important bits are buried and feel distant from my own thinking. All the assumptions and tradeoffs and questions may or may not be legitimate, but it’s hard for me to get into flow state and carefully check them.
So I still want the collaboration step. I’m just less convinced that generated prose à la plan mode is the right interface for it.
There are many ways people solve this. I use spec directories, and the agents just keeps updating it as I make decisions. It isn't a single file, it's a directory; sometimes nested based on the size of the project. I'll have the directory open in my editor of choice. It being a directory allows me to use an arsenal of text tools, and also optionally commit it for history/tracking.
This tool is a difficult sell. Users _might_ consider an Open Source version, but switching people away from familiar tools is not easy.
I likewise don't understand this push. Surely it's a trivial feature to keep enabled. It mostly seems like advertising that "our models are so smart you don't need to plan". But I find it to be a great mental review process for my workflow.
It sounds like the Ant perspective is "you can just ask claude to plan", but at the same time that's a little more tedious than shift+tab
I see what you mean, but I’m saying something else, I think plan modes were solving two different problems:
1. giving the agent sufficiently precise instructions
2. helping the human understand what’s being built
I think #1 is increasingly going away as models get better. #2 is a separate problem, and I actually think it matters more as agents get more capable. I’m not arguing that human understanding should disappear with plan mode.
I’m also not pushing this on behalf of a model lab, I don’t represent one. These are lessons I learned from my own mistakes building Nuanced (https://www.nuanced.dev). We built around a very explicit plan-oriented workflow because that’s how I used to work. I spent years at GitHub writing ADRs, RFCs, and design docs, and I still really like writing because it’s how I clarify my own thinking.
What changed for me is that AI-generated plans don’t give me that same effect. The agent reads between the lines, generates a lot of detailed prose, and now I’m parsing decisions and assumptions I didn’t actually make.
So I still think human understanding deserves a first-class primitive. I’m just increasingly unconvinced that a big blob of generated text is the right one, especially as we move toward many agents working in parallel.
> developers have been complaining about AI supposedly killing their jobs, yet they want to take away the main feature that lets them be an active collaborator and participator in the process.
Goomba fallacy. These are two distinct groups of people.
Yeah, honestly, I'm wondering, not even why people are against plan-mode, but how they're getting stuff done without it.
Maybe it's me (it usually is), but I don't give an LLM small tasks that a human could knock out in a half-day. I give them big tasks, stuff that would take a human a few months to a year - and then I ask it to give me a plan, as a living document, and we spend a good hour or two iterating over the plan.
Then, when its finally at the point where I'm happy with the plan, and I've talked it down from wherever it first wanted to go, or pointed out that we don't need to be all-things-to-all-(wo)men and focus is important, I get it to start going through it.
Likewise, I get it to maintain TODO.md with lists of known bugs, separate lists of future-features, and again, I make a rule that this file must be updated whenever something material changes. I just asked Claude "where are we ?" and got back stuff like:
The ## Still-open detail section lists two items:
- Task #1070: the ported back end doesn't fold offsets into vector loads, so it emits an extra add on 11 files (from 892) at -O3. The output is correct, just longer.
- Task #1080: array sizes must constant-fold. For example, u8 buf[EVSZ * MAXEV] is rejected because size expressions only accept a literal.
This is for 'xc' [1] - a compiler for an Objective-C-like language (but without the excessive []). The language has ARC, blocks and bound-functions/methods named 'block' and 'callback', automatic parsing of DWARF data so you can #use a shared-object, so there's no header files - just read enums/types/functions/methods from the shared object. It's a cross-compiler, runs on mac,windows,linux and creates executables for mac,windows,linux,ios,android,WASM (amongst others). I have a binary running on my iPhone which was written on, and signed on a Linux box - no Apple software used at all. Oh, and it produces code that is very comparable to clang in speed on both arm64 and x86_64.
You can appreciate it's a reasonably large project. It's taken actual months(!) [grin] for me to get working. Months! There's no way I'd approach a problem like this without detailed plans of what I wanted the language to do, where we were going with it
FWIW, "I" wrote blewit.net [2] entirely in xc - both the server back-end (#use <psql> was very useful for binding to Postgres) and the WASM client - which share classes between client and back end, to make it very difficult to get out-of-step between them. No Apache (#use <tls>), no scripting, just a lean-and-mean daemon talking to postgres via valkey (#use <valkey>) - a reddis-alike. Oh yeah, blewit.net has a plan too. Actually it has lots of planning :)
I have never used plan mode and always worked by planning. I just say 'don't write any code yet. I want to [describe usage]. Ask me the top ten relevant questions to doing this well. We're working up a planning .md we'll handoff to a fresh agent.'
> While the ease of generating code this way triggered a greater dopamine reward
I get zero dopamine rush from this. I get a dopamine rush from building something or figuring something out myself. I don't get a dopamine rush for generating thousands of lines of code that I then have to try to understand.
I like writing spec files exclusively by hand and just asking the agent to surface questions about it, which I then clarify by editing the spec file further by hand. It keeps the spec file more manageable than having the agent generate the spec file from your conversation.
I disagree that more intelligent models make plan mode no longer useful. Even with superhuman AI, I would still like to iterate on a plan and understand all the important decisions before changing the code. AI cannot read minds, no matter how intelligent it is. Hell, often I don't even know exactly what I want before the planning phase is finished.
And yes, you can say that I can do this without an explicit plan mode.. But it's such a useful and common workflow that it deserves a special mode, IMHO.
Plan mode isn’t that useful but planning is, I never used it much, but in the agent build mode, I lay down the specs, the architecture, and everything I can think of, and I ask the agent to make the blueprint with the specs that will be used throughout the project, I review them, make the needed edits and revisions, then code and review follow. This is good because you can use the blueprint in any other model or agent, even humans can read it and understand it, sure, sometimes it gets overly verbose but it’s better than nothing.
With Claude, most tasks do not require discussion, I know exactly what to ask and what approach I want to take but for especially complex functionality with multiple possible options, I will ask it to list the tradeoffs and suggest an approach. It's still kind of 50/50 whether I take its suggestion or not; it's still a bit off in terms of weighing up importance, but it's really good at listing out relevant constraints and tradeoffs. It sometimes misses opportunities but it always sees the tradeoffs and the issues.
I still like plan mode in my harness because I restrict it's tool use so it can't go rogue and write documents outside of a specific folder. In my case it is /agile which contains a roadmap.md with epic1.md through is many epics as needed for the full vision. Then I often have one chat session per thread. If multiple agents take a sprint or a card it serves as a clear unified context for me and the agents to follow and then track what has been done.
Yes, plan mode is still a reasonable way to guide agents _before_ any code is written. They're so eager toshamelessly change the files even after a simple quesiton.
But nothing is stopping us from prompting agent to only write to plan.md this session.
Fascinating piece, puts in words what I've been feeling for a while..
Over time I found some useful patterns (indeed after not getting what I want from plan mode), but this piece convinced me to double down on them and always try to find (and let the LLM construct) clarifying abstractions that have a deterministic relation to the code.
2 examples I recently build while I'm developing a large Django system with a complex datamodel and RBAC (spending quite some time to make it look good):
A script that makes an svg of my data model with all the classes/tables laid out and their relations encoded in the line ends (1 to 1, many to 1, many to many) and their on_delete relations encoded in colors. This also helps me discuss with stakeholders. The visual is also in the README.
For the RBAC model I decided it should be declarative so a TOML in the code that seeds the DB with the roles. I quickly landed on small script that translates the TOML to a markdown table with roles as columns and perms as row. It's also in the README.
After reading this piece I'm going to actively think how I can build these visualization more often and consciously, on different levels. Great realization.
I'm using a set of skills for planning now which does something similar to plan mode but it's creating documentation that later skills reference it while building. It works so much better than anything else I've tried, because it keeps things on track throughout the feature buildout and across different agents/models/tasks. From what I read here (& in general) my take is that a lot of people haven't adopted something like this and it's just totally the wild west right now as everyone is cooking up their own flows.
I agree that as I move away from holding agents' hands through actual coding I need a different way to monitor what's going on. What step of the plan are we on, what are the tests actually doing, what agent owns what, etc. I haven't found a product that does a great job of that yet and it seems like the next frontier of the 'IDE' to me. It'd be more like an IM(management)E really. The closest thing I've seen was whiteboard [0] but I didn't have a great experience trying it out.
Maybe, I always say "let's write a plan and refine it" which ends up being a nice rubber duck session to keep claude and friends from going off the rails implementing things like privacy or encryption or some other rabbit hole when I just need it to do a few features at a time so I can review while it starts the next few things. Maybe this is slower to deliver, but I don't spend a lot of time unwinding commits.
I don't necessarily use plan mode explicitly for this, but I might turn it on if I think something destructive is coming up, like a schema change or some other devops thing where claude may decide "oh we copied it all over so I can delete it from source" "oh I'm sorry I didn't catch that error code and assumed success, your data is now gone" lol. This happens a lot less but I've been bit right on the face by it before.
Are plans the wrong abstraction? You can use a big enough plan to spam tickets into linear and have claude burn them down. I find it increases subagent accuracy.
I still find plans to be useful for big changes.
I first ask to create a plan for the specific feature i have in mind.
Then i ask for a second round of research with the current project in mind.
Then i ask if it can make improvements on the plan.
Then i read the plan and note issues down and let it correct them.
In my experience the problem with plans is that llms (at least gpt and claude) are really lazy and will stop at the first solution without even understanding your project. I had plans that wanted to add a few thousand lines of code just to support a very minimal part of a feature. After asking for a second round of research with the project specific limitations in mind it recommended a very small change to the existing codebase that did the same thing.
This is probably also the reason why most llm generated code stacks thousands of lines of code.
Plan mode - primarily for me was for Human. Models through chat could always be informed and forced to correct but real difficulty is having a good guesstimate of what model know and how its going to act. Relying on Model and its future improvement and declaring plan mode is no longer relevant - is too jumpy IMO. If you are running bunch of agents - then you need to confine them and also instruct them in BATCH fashion. Plans also useful for shaming model when it goes wayword and give it a baseline to correct. With AI, one needs a control instrument plans and spec are useful instruments.
Plan mode is "dead" because most plans are actually generated prompts to steer agents to keep doing the next thing, in hope they keep doing the right thing until the end. Real life is not so elegant and straightforward, you have a lot of "nah, screw that decision".
It is more useful to capture the motivation of the "plans" and their expected/observed end results, rather than the plan itself.
The disconnect between the prompt and actual motivation can be significantly high.
AG has dedicated plan mode now, idk why. Probably just rename into "guides" or "opinions" or "brainstorming".
A plan means it preps all work for the agents up front, tests that evals work, makes sure the dev environment is right for each agent, then finds and fixes each before the distributed tasks even begin.
What I thought would take minutes took hours as a supervisor or one agent did the prep / pre flight work.
My solution so far has been to drop all but basic setup and force the supervisor to ask before every op - if this is not the design choices, can this be run in parallel? If so, hand it off NOW.
I'm still iterating this workflow, but less setup for all the minions plus handing them work that may be incomplete/ broken is caught and fixed by the minion and its own qa gates.
This can mean a number of minions end up replicating the same fixes, but in general the time cost of that is small Vs the supervisor working in parallel instead of too sequentially.
- Inefficiencies in delegation, necessitating workflow patterns for small projects (big AIs hide this problem until you scale and they hit the same issues).
- Limits of the AI would be harder to find or notice (e.g. where time - and cost - is being spent needlessly).
Yeah, it's way better when you do design documentation, or even ticketing, to instruct it not to include any implementation specifics. You're not doing the deep dive on the zero shot that writes the ticket or the document and so it is much less informed than the agent doing the work will be.
I'm genuinely curious as to what you're working on that you find 4B models good enough. I wouldn't even let a 27B model code, never mind supervising smaller models.
At the moment I'm still doing shakedowns, so Typescript games compilation with a menu that has 4 games and retro artwork.
This seems to be a good example because things like the menu, high score boards etc are common, but the games are distinct. Then there's the artwork which requires decisions on look, and for coordinating.
The Qwen 4B model is multimodal so part of the AC is to view the output - I've a robust anti AI-look QA chain for that I've been using elsewhere, e.g. no floating parts, consistency, obvious missing fingers etc etc.
The longer term plan is to do some llama.cpp refactors specifically for some target hardware I have and implementing slightly different novel architectures I'd like to try (one I did already targeted CPU inference, which I did using 3 agents with specific roles; main planner, QA for planner, and benchmarking/environment handling)
The implementation was 85% of the speed of the original maxed out on my hardware but performance scaled with CPU core count whereas the original implementation plateaued. Unfortunately the break even mark seemed to be around 30 - non HT - threads.
I suppose I should look at that one again, since the increase in cores did not linearly drop off performance e.g. due to memory contention.
I still use plan mode. I use it for large features that I haven't fully thought out. Which I guess is like the author's "planning" phase. But I don't see the issue with using plan mode for planning. I welcome the agent to ask me for questions and clarifications. Do I read the full spec it produces at the end? Usually not. Usually the agent is smart enough to highlight the important tidbits and assumptions. I read those.
It's particularly important when you're making architecture changes or things other features will need to build on top of. I want to know what libs its going to use. The nitty gritty details I don't really care about.
At work, we use Claude Code with superpowers which organizes most work into planning followed by review followed by implementation and superpowers applies a fairly high level of scrutiny at every step. It's really quite effective, I don't generally read the code that comes out. In that workflow I have never used plan mode. I don't know what I'd use it for.
On my hobby projects I use Antigravity CLI and almost always start with /plan. This is a builtin skill - not a mode. It generates a plan artifact that I review and approve. Once approved the implementation speed is uncanny compared to Claude.
I tried using superpowers with Antigravity CLI but that slowed the agent down considerably without much benefit (plan quality was much worse).
I have been using this super simple skill[0] to make Claude ask exhaustive questions about the implementation of larger features. It has been absolutely amazing and felt like a step change in the output quality.
For complex features there can be 10 or more questions but I have a very strong sense of understanding the changes about to be made and Claude is very good at following all the decisions exactly. It's like working with an engineer who is both excellent at soliciting requirements and fast at implementation.
I think it is clear that planning before coding is generally a good idea - it's true for human developers and even more so for AI agents. The question now is more: what is the right way for an AI agent to form a plan and share it with the user? Maybe the dedicated plan mode in the agent harness is no longer needed, but I would like to make very sure that the AI agent has a plan upfront and that this plan is compliant with my requirements, my general expectations and that it is consistent and clear.
I can't speak for Claude, but in Codex, plan mode is very useful for finding serious issues with one's spec. I use it to improve my spec. I keep rerunning plan mode over my updated spec until it stops finding issues. It is only then that I execute the spec.
Personally I really like grilling before a big feature, it's a Matt Pocock skill that will iterate and ask you a lot of questions so you REALLY know what you are building. It's particularly good for complex and big changes, so you actually know what the AI is gonna build before it does. Of course, that's not always the case, sometimes you don't care about the specifics. But I think an integrated "plan mode" is really not needed, maybe a "brainstorming" skill would be more useful.
I will try to update this repositorysoon, but I had to stop because the ahem compatibility failure of Claude Code env features; observer; agent workflow message passing LACKING on installed marketplace plugins
I am still actively working on thesis : a self-directed plan mode to generate artifacts that can go through multiple evaluations of interactive interrogation is valuable
I've been using superpowers:brainstorming skill with claude for a while and have found the process to work well for me. I generally spend 10-30 minutes typing my requirements and ideas for a feature and then I feed it into brainstorming skill which asks questions to fill in the gaps before starting to build things. I don't bother reading the spec it creates. I have been consistently happy with the results.
However, since everything changes every 5-minutes now; I am curious what is now a better process than using superpowers? What works for you?
We experienced another step function in model quality over the summer. Models are now one-shotting medium sized tasks that require significant engineering judgement. We're moving into an era where the production of software will be done entirely by machine, requiring only an abstract description of the need the system is to fulfill in business language. This has been the goal since the 1970s. "Plan mode" isn't necessary when the model can suss out what you need from a few sentences of prompt and make competent engineering decisions to fulfill that need.
My own tool has a Project Starter, that takes users from idea to todo list in a number of very specific phases. I've been using it for small to medium sized projects for a year now and it still works great for me.
Just don't expect to end up with a finished project; it's more like a first draft. Once it's there, it's much easier to determine what it is you actually want, since you can directly experience what works and what should be changed.
One important caveat is that I do not work with agents; each step goes through a fairly rigid manual review phase.
I don’t agree. I think the way maybe the author used planning to maintain an understanding of the changes is largely not necessary, but there are other good use cases.
For example, it’s a great hook in the process for agentic review.
Get a second agent to look at what will be implemented and check it for inconsistencies, check it against whatever decisions were made or provided previously in the chat, or against whatever technical rules you’ve written out for your project, before going ahead. It surfaces a large number of opportunities for refinement, and generally pushes the output closer to the direction you’re looking for.
I've been working on a custom orchestration system which starts by creating an overall work plan, then splitting up development tasks among agent/model combos depending on their capabilities/expense, then coordinating their output, testing, etc. with various adversarial agents acting as gates throughout the process. It's really satisfying to watch the task board as agents go through their little routines.
It's been an absolutely fantastic way to set tokens on fire and watch them burn, and produce pretty much nothing of value.
Planning largely doesn't work because it's extrapolation. You don't know for sure how that remote system or compiler will react to the proposed change until it actually occurs.
Execution is the answer. All the complex success stories I've seen involved the LLM iteratively probing live data sources with varying filters, throwing code changes at the compiler over and over, or invoking shell commands until it succeeds.
> We still haven’t solved how to help people stay oriented as hundreds of increasingly capable agents change a system at once…
Building an issue tracker that addresses this. It can replay the workflow after the fact like a movie, and pin down the parts that require human input via tagging and inline diffs in the tickets. It is git-native, lives in your repo alongside the code doesn’t require any external service.
I am not so sure honestly. You can still wreck a codebase if you don't plan because the ai can make assumptions you don't want, or as other people say it's all a matter of taste
I also noticed this personally. Consider when an agent says
"Should I start with the spec, or go straight to building it?
Unless you're working on a super crucial piece of engineering, you can probably get away with going straight to building it. Even if something doesn't go the way you intended, I find that it's usually faster to correct the agent later, once the initial implementation is in place. It's more of an iterative approach to building and I feel that it is less cognitively demanding.
I only really used it because the harness was way too trigger happy to start making changes. Even if I just asked a question some times I would come back and it refactored the whole codebase. Now it doesn't seem to do that any more.
I still would appreciate a "read-only" mode. It's not uncommon that I start a harness ONLY to explore and understand the code and I don't really want one typo to have it off building something, or even to save a plan document.
It's quite interesting. I use blend mode pretty heavily. On both claude code and codex.
But it's no longer to plan a single PR that can be done easily without plan mode.
It is mostly because I am more geared towards having my system work on initiative level changes where it works for days and plan is a good document to maintain to keep the agent aligned with the original goal and avoid unnecessary drift.
I have rules.md read on every start of the session, in which some of the rules are:
Never start coding befire I say codenow, and add datetime.
Before patch explain in plain English what it does.
Each file has datetime added to the top of the file, and updated if exist.
Each file has a backup copy, created with name.datetime and checked after patch vs backup file, and again with new .backup file for logic vs new backup
I think Plan mode is very useful when building something that needs step by step validation, but also needs a broad strategy before starting.
For instance, in games where I work, we often need to manually test work out by playing or by using tools in ways that aren't feasible for the AI to do. In that case, getting the human in the loop between steps is an organized process when it's following a staged plan.
I find a lot of value in using plan mode to get the LLM out of dead ends. If it's in Act and starting to spiral (evidence of looking for something or trying to better understand something), I'll kick into Plan and tell it to think a bit harder—also avoids it wasting tokens retrying the same techniques—which results in a much more considered, and more often than not, correct approach.
My problem with planning and having the models ask the user questions is that the questions the models ask are still often irrelevant: you’ll find yourself answering a bunch of questions like “Should I use the existing system or create a whole new one in a different language?” - yet it will then make assumptions about other things that were actually ambiguous.
I still use plan mode, but during plan mode I will spin out sub agents to implement the current draft in a tmp dir and bring back lessons. I feel this keeps me grounded in my original starting point, rather than ending up with a implemention that meanders through the agents own discovery process. It also lets me ask concrete questions about (possible) implementation
I think with Claude Code plan mode has value even if Fable “gets it” off the bat. It’s a chance to explicitly disambiguate intent through AskUserQuestion tool as well as utilizing the option to clear content when plan is approved. The automatic disambiguation step is RLHF’d out of the model even when it’s needed.
I haven't used plan mode for a long time. I do still plan, I usually ask claude to make a plan with me first an explain how it will work, then I iterate with it until I am satisfied and then just let it go.
So far it has not started implementing on it's own before I told it to.
I also bounce around to lots of different models. Some are cheap and dumb, some not. When unfamiliar with a model, the last thing I want it to allow file edits and destructive tool calls. Having a mode explicitly for that is helpful.
I don’t think it’s dead at all. I use it as a phase transition to let the agent decide when discovery is sufficient, as an intermediate phase between that and implementation. Even as a human dev you rarely go tinkering with files all over without first forming a plan.
In Claude Code it started to feel like it just asks 2 questions and sometimes there isnt really needed 2 questions so the options there often felt like choosing to downgrade or do something actually dumb because it was forced to suggest 2 things.
It's been a while (well, 8 months) since I used Claude Code, but what I can say is that I greatly enjoy using the plan mode that comes baked into omp (oh my pi), especially for tasks around firmware development. I recognize that has a very different workflow than webapps though.
Imagining a mode where the agent is very transparent about its direction and process, and I can passively interject at any time to steer the outcome, without having to wait my turn or hard-interrupt.
I'd like a brainstorming mode, with limited side effects.
I want it to go and read stuff but also invoke some commands and generate reports and discuss on the results. Latest GPTs in codex plan mode always go.. write a plan. Shocking, i know, but thats not what i want on every turn when im in that mode.
It is?
I still found it invaluable, when having a lot of externally managed knowledge. Basically the only gate before it goes on a certain failure, when dealing with proprietary libraries, tools and services
Switched from opencode to pi and I do kinda miss planning mode. I like spending a bit more time in that phase and having multiple providers LLM look at the spec doc etc
That’s very much coming from a desire for improved quality than understanding though
I work on long horizon scientific tasks with expensive computational tools. In that case users are basically asking for the harness to stop and confirm the cost (and plan) before it goes ahead
YES, and I also find that plan mode incentivizes bad practices and results in steadily reduced thought. On bad practices, I refer to zero/one-shot big bangs over iterative features/bugfixes. On reduced thought, assuming the plan was sufficiently comprehensive because it seemed to ask reasonable questions.
It can still be used in ways that I personally consider correct, but I think the parts I personally consider incorrect are so inherently alluring that I find plan mode to be an overall net negative for software development. I celebrate its apparently impending default removal (at least Claude Code and OpenCode are openly stating that they think it's time for it to go).
I too have started using plan mode less, mainly for two reasons:
1. I plan my prompt more carefully and think about architecture up front
2. I found with more recent powerful models, the clarifying questions were generally not useful because it was pointing out issues it obviously knew the answer to and would have resolved in implementation anyway. So essentially they became time-wasting and anxiety inducing for no good reason
The mental shift I made was to not accept poor understanding of a code base on my part when writing the prompt - if I don’t understand it, I can’t predict what assumptions the model will make even on a basic level.
I used to tell myself that plan mode mitigated that, but I usually ended up mentally glossing over the generated plans anyway.
That just resulted in pure anxiety-driven engineering, where I’d often spend extra cycles verifying what was built and worrying about the design.
So invest the time understanding the system, at an appropriate level. That level will change over time as models get better.
+1 to a lot of what you said here, especially this:
> I used to tell myself that plan mode mitigated that, but I usually ended up mentally glossing over the generated plans anyway.
At a certain point I realized I far preferred delegating the first implementation pass to agents and reviewing the _code_ instead of a plan up front. If you had asked me this question ~2 years ago I'd think anyone would be insane to hand off discretion like this, but frontier models are just so _good_ nowadays.
Let's say you launch 10 working sessions in a day, would you rather:
* Review 10 plans, and _then_ review 10 PRs, and _then_ maybe adjust the approach on 1-2 and merge the other 8
* Review 10 PRs, and maybe adjust the approach on maybe 4 and merge the other 6
The first option just feels like needless attention for the sake of feeling in control; but the reality is that these models are becoming just as good if not better as us humans and our reckoning is here. Our codebase has enough linters, hooks, and guards such that generally the agent just follows our blessed patterns already, so what is plan mode actually doing beyond giving me a false sense of security? I'm still going to read the code anyways.
> how do humans maintain a coherent mental model of a software system
This is the core question and it is wrong. Because it focuses on the software system rather than the problem you are trying to solve. A better version would be:
How do humans maintain a coherent mental model of the solution they want to create?
You need to have a picture of the whole process and everything related, not just your architecture and code.
Why is a plan "mode" needed? I really hate reading plans in conversation UI. One line in AGENTS.md is always better, and you can customize it to fit your document convention.
From the communication perspective, the planning process is most about creating clarity and alignment, on things like scope, constraints and decisions, among people (and agents now), often requiring multiple rounds. The planning tool built in Codex or Claude Code is not persisted, version-controlled, nor well accessible.
In practice, My AGENTS.md contains an instruction about writing a document before starting implementation. Usually I won't read that document, because I don't want to micromanage agents. That document mostly serves a historical purpose like ADR, helping me find out what agents missed, made mistakes on or misaligned on, if needed.
He just rediscovered agile to replace waterfall.
But seriously, I am reading an early-release book Regenerative Software:
https://learning.oreilly.com/library/view/regenerative-softw...
I don't 100% agree with the book, but it contains a lot of good ideas about how to externalize knowledge about a piece of software from the code, which is what everyone tries to do with plans, specs, chats.
I totally disagree. This perspective only works if you don't care about token cost.
I'm using Astra for some homelab work and I tell it I want to do something and then it burns through a giant percentage of my 5-hour allowance (on the £20/mo subscription) doing that in a really weird way. I've had to tell it to stop zooming off and doing things, calling loads of tools and looking up websites, and just have a quick conversation first. It's much better now, but I've basically just reimplemented plan mode via AGENTS.md.
> This cut-and-paste workflow felt clunky and made it arduous to work through an idea while keeping track of the current plan.
I mean, it is, and if that were the workflow I’d be very fed up of it by now, but the plan mode in Claude’s VS Code extension has supported select and annotate directly since at the very least early 2026.
oh interesting! I haven’t used vscode for anything meaningful since shortly after leaving GitHub (2023). I generally find it overwhelming but codex added the annotations too and I love it.
hard no for sufficiently complex existing codebases. even astra misses nuances in my moderately high loc app. plan, then implement is much better and, more importantly, much safer. at least i review the plan.
Of course, it's impossible to know for sure what was LLM processed or not, but some of your posts (like this one) have been getting classified that way.
[I work on Claude Code] I broadly agree with the author’s point: plan mode was useful, and is no longer useful.
In Claude Code, all plan mode does is add a little reminder to every user message along the lines of “you’re in plan mode, please don’t code yet”. It’s something I came up with late on a Sunday night many months ago, when I got tired of asking Claude to plan with me first before coding in each new session. Something people might not realize is plan mode has always been a prompt — it has never changed the toolset because doing so would break the prompt cache, and so would be expensive for users.
This worked well for a while, until a few months ago, using early versions of Fable, I realized that I wasn’t using plan mode anymore because the model just got it, and because for the increasingly complex work I asked the model to do, planning had become interactive and iterative. With Opus 5.5, I feel Opus has gotten to that point too.
For codebase understanding, I sometimes ask Claude to generate an artifact that explains some aspect of its changes. For complex diffs to core parts of the system, I will often ask it to make diagrams or even interactive demos so I can better understand the change and alternatives considered. I don’t do this very often, but it’s a useful way to explain code when you need it. I ask Claude to attach these artifacts to its PRs also, so others can understand and future Claudes have the context.
> This worked well for a while, until a few months ago, using early versions of Fable, I realized that I wasn’t using plan mode anymore because the model just got it, and because for the increasingly complex work I asked the model to do, planning had become interactive and iterative. With Opus 5.5, I feel Opus has gotten to that point too.
For me it’s actually the opposite, and Claude Code’s plan mode isn’t nearly sufficient. Personally I ask Claude to write down a markdown file with its plan, then review the plan using plannotator, and then go back and forth (most of the time it’s actually the comments that are the problem, not the code).
Then start a fresh session, seed it with the plan, tell Claude to find ambiguities / friction points / oversights, resolve those, and then implement it.
Review once again with plannotator, go back and forth, and then send PR.
Maybe not the “vibe coding” that was once imagined, but this does ensure I am fully aware of the code and architecture, the quality, and this also prevents long term degradation.
I've recently gotten religion on the workflow that is many (relatively) short-lived agent sessions passing planning/handoff docs between themselves. It's better for my own task tracking, better for handling "oh btw I noticed XXXX", and better as a clear review point. Overall it just feels like it takes a lot of the formerly implicit context that was whatever we happened to have talked about and turns it into a much more explicit "this is what you need to know, now go".
Currently looking for a framework for managing this in a more formal way, and I think it's probably beads, but interested to hear from others.
Spent the past 1,5 years building a tool that might be relevant, helping keep durable task state between agent sessions. It is an issue tracker persisting state as immutable event logs, allows you to inspect workflows after the fact, lets you inspect diffs inline in the tickets and it is much more lightweight than Jira/Linear. There is no central service to integrate with, as it is Git-backed and lives with your code in your repo.
https://ljtn.github.io/epiq/
Might be worth a look if you’re evaluating alternatives to Beads.
This looks really nice and I'm definitely going to give it a go. I been relying on Jira so this will be a breath of fresh air. Beads was great in principle and I haven't given it a look in a while but this looks much cleaner.
Curious to hear how you find it after trying it!
Not OP but will definitely have a look.
I have some older projects that use beads (I still run an old version without dolt that's imho pretty good overall) but lately with Fable also have a few newer projects where I just have the agent write docs and keep a worklog with the what/why/decisions etc. (I think I read it here on HN somewhere and figured I'd give that a try.)
The latter seems to work pretty well for now (slightly better than beads) but I'm always looking for ways to improve it. This could be an interesting replacement.
I think it works well for tracking intent. I tell my agents to use a "human-input-needed" tag when there is a high stakes decision fork, and then I just filter on that tag, immediately highlighting blocked workflows. I think you can get a long way with clever tagging, filtering + time travel.
Wow. It's like beads++
Great work man.
Checking this out, the skill it wants me to add to my project seems to reference issues that wouldn't make sense in the context of my project. IE: `QPPR7PE` and `MBB6WY5` ?
You are totally right in that it makes no sense. It seems like the skill.md had regressed as I had updated it via agent proxy. Thanks for pointing it out, the skill has been brushed up.
I’ve been trying it out on a personal project. It’s been going well. The only difficulty I’ve had is getting it to work with cloud sessions on Claude.
Got it! I am not up to speed on Claude cloud sessions yet. Interested to hear how that works out for you. Let me know if you figure out what's getting in the way, I'd be happy to look into it.
Oh shit, this looks incredible. Thanks!
> 13:37 - 13:49
“ayy lmao”
I have a lot of little projects and I also prefer this way of working with agents. Sometimes I would start to interrogate on a specific portion or ask questions to better understand a concept, and the session would get poisoned and the agent would fixate on that topic for all the remaining turns.
I asked fable to look at my interaction patterns and clearly stated my frustrations and the problems I wanted solved, and it designed a simple process to track things in git and built a couple simple session hook skills. It’s pretty lightweight and I’ve been very happy with it for a couple months.
The power of this mode of work is that after you deconstruct the task into smaller subtasks, it's a lot easier to use cheaper models to implement that task.
I get a long way using models like Opus to make a plan of action and a bunch of tasks, and then using Deepseek to implement that plan of action. Saves a bunch of money and is fast.
Yep! The other big advantages for me:
- I have a record of work done and work to be done that helps _me_ when I come back to the project after several months. It’s committed and lives with the code.
- when a task inevitably ends up more complicated than I thought, I can in that session break it up
- I initiate sessions from multiple computers, so things stay in sync (through git)
- I also have a “tooling” repo that builds out some views of the work and hosts it for me to see when I’m on my phone.
- The hooks let the agent manage all of the workflow/task management, so there’s very little management overhead for me.
I rejected beads and JIRA. I wanted something more lightweight.
What do the session hooks do?
- start: invokes `work brief` skill to put the current tasks into context
- pre-compact: does the same so work state survives compaction
- stop: lints the work-tracking yaml files to make sure that updates mid session don't break
Nice. I found myself doing a lot of repetitive stuff handing off between agents, time to fix that. I think I agree with your approach of keeping it lean.
I also sometimes rewind the conversation after going off on a tangent, sometimes not if I want the agent to have the sorts of things I’m thinking about in context. I probably justify keeping too much crud in context than I should
I've been on this kick since I realized the primacy of the initial part of the session context. I created a python app that reads a phased plan and kicks off a new session for each phase. There is a standard prompt and handoff mechanism to determine if we encountered any unforeseen issues that we need to address in chat, but otherwise it will just grind with a clean session with appropriate context for each phase.
Just kicking off a subagent does this automatically...?
My main conversation is usually with an orchestrator that hands off work to various (usually cheaper) subagents to plan / review / etc. It has instructions to find the correct model for each task and not to do too much itself so a multi-phase plan automatically gets a fresh subagent for each phase.
i disable subagents. it also helps that i work in microservice size repos, so a single agent can keep the full relevant context. subagents waste alot of unnecessary tokens doing code exploration that the main agent is going to end up duplicating, i think they make more sense for monorepos.
anecdotally I've tried the same plan on different branches with both approaches (single agent with subagents implement the full plan vs. my little session per phase app) and then had agents judge the code quality and my approach worked better and saves tokens. horses for courses.
[dead]
That sounds very similar to just manually compacting after every message. Is there a difference I'm not seeing?
I used taskwarrior for myself and agents, but felt it was insufficient for agentic era in many ways, so I started building my own a while back:
https://aventasks.dev/
I use Jobs [0] to manage this—it's an agent-first CLI to track issues and tasks. A single `job orient` command gives the agent the current task in the context of the larger plan. It's a replacement for Plan Mode and issue trackers, and it has allowed me to execute massive plans in parallel with minimal oversight. There's a web UI, but it's a work in progress.
[0]: https://github.com/bensyverson/jobs
Look into beads/dolt then - it does this pretty much with a cli - Jira for agents :)
https://recursive-mode.dev
httpd://recursive-mode.dev
[dead]
[flagged]
> (most of the time it’s actually the comments that are the problem, not the code).
I wonder if it's just a consequence of a gigantic training set full of comments completely out-of-date with the code, leading to the model considering this "normal"
I don't usually see Claude leaving comments that are obviously out of date compared to the code. The problem is that the comments are just a dumping grounds for Claude's stream of consciousness, verbosely recording random bits of history and context that are true and at least somewhat relevant, but without cohesively explaining why the code is the way it is. And some sentences in the comments will be beginner-style restating what's obvious from reading the code.
Exactly. Comments are basically its scratchpad for whatever it wants. Ask it to change a number in a TOML config and it'll also add 2 new comment lines above it with some Claudeslop about it being changed and what that accomplishes, as a useless note to itself.
I now make sure to do a big decommenting pass before every PR.
This kind of stuff annoys me too, and I have instructions in agent.md to avoid it, which helps.
But I also am starting to just let go and stop caring. It’s not clear to me that it causes problems down the road, it’s easy to strip out en-masse if needed, and in my experience, agents now are really good at read git blame, the commit log, even prior agent transcripts if available to sleuth out when a change was made and why. So yeah, it’s annoying, but the code agents write for me is increasingly never read by a human, so does it matter?
This. Be honest: who reads the human-written comments? ;)
When I read, I skip most comments, especially the larger "Javadoc" style. My brain sees them colored differently in the editor and it doesn't even take mental effort. Then, when I have a question about the code, I look back up for a relevant comment. That doesn't happen very often.
If Claude writes great code and leaves a garbled Claudese-but-accurate comment ... I can read and comprehend (with like 5x the effort of a human comment) ... that's a small price to pay.
Yes, comments might be my biggest gripe with CC due to what you said. 9 out of 10 of my revisions to Claude's work is deleting or rewriting comments. Such a silly thing for it to fail on.
(I do have detailed instructions for it on how to comment (or not) but it has not fixed this.)
You can only fit so many bits of watermarks in code... but comments on the other hand, the possibilities are endless!
[dead]
Nah, it’s the typical Claude-isms, and way too much info, info in the wrong places, and putting lots of planning discussions into comments that are completely irrelevant.
It’s always “you explain only what but not why” or “this is way too much prose” or “these comments don’t belong here, they should be inline comments” or “this is completely redundant as it’s already obvious from the code”.
I do find that once I beat it into submission and the codebase is “clean”, the new code it generates gets better and better, which makes sense gives its pattern-prediction nature. But it seems like there is work to do for Anthropic in terms of getting Claude to not confuse code comments with dumping its interactive discussion state into there.
There is a popular skill for this kind of workflows: https://github.com/obra/superpowers
Superpowers never really held up in my own testing vs /plan - not just more expensive but markedly worse code organisation, because the plan itself leads to higher cognitive burden for the models - and today it's even worse, because /plan doesn't hold up versus coming up with a few high level slices and having cc work on one per session, for much the same reason.
5.5 is much closer to Fable so i don't even need it. I am pretty sure it's got Fable's DNA in it.
I really need to find a role where I can do more DX...
Yeah I always balked at the amount bloat in that repo, and I just prefer a more lightweight setup and plannotator’s UI which I can use to interactively review stuff. It solves this one problem better, where superpowers tries to do way too much, poorly.
https://plannotator.ai/
Going through the superpowers spec-plan-implement-review cycle took one of my test runs (local AI) from 40 minutes to 8 hours, and produced a lower-quality result. I want to use something like that skill, but I'm doubting the efficacy of this one.
I dived into one or two of the skills there and some of it feels very cargo-culty. In a similar vein to the article, the models have improved markedly now and I’m not convinced that “You are an experienced senior software engineer and an expert reviewer, here are five pages of instructions on how to do a review” style prompts are massively useful any more.
I get great review results (as good or better than colleagues using superpowers or even adversarial review skills) just by asking Claude to review a PR and spit out results in order of severity.
My workflow is very similar. But I just ask agents to write design in html instead of markdown, due to its richer layout and better interactivity. When the design is about UI or anything related to graphics, this approach is extremely efficient.
I also found that having the design reviewed by multiple agents has very little marginal value. The review agent will always find something to improve, but mostly it’s just nit and not anything super important.
I used to let Claude just upload the html design doc to Claude artifacts for me to review. Recently I switched to codex and started to use my own tool https://github.com/hyperlogue/r3 to complete this workflow.
I do the same, I don't use Claude Code or Codex planning because it is mostly pointless, even with Fable/Astra. I just have multiple agents work on a markdown file which I manually perfect, often breaking into multiple different files for large features or PRs. I also create design 'handoff' documents which I feed into Claude Design or Astra along with screenshots and wireframes. By the time an agent does something I'm well prepared.
I've tried doing the incremental, iterative approach with just Code and it's just not as effective unless you're working on something simple or experimental. Or you're shipping to something non-serious or perpetually beta.
Have you recently tried working without generating that plan? What I've been doing is first tell the model I want to plan the implementation, talk about it a few turns, then when I'm happy with the idea and method just tell the model to go ahead with the work.
Note that this is only really necessary for complex work that I don't know yet what the best way to do it is.
I've tried doing it your way as well, but there was just too much fiddling about with writing the plan somewhere, then having another session rebuild their context with whatever info is in the plan. It really didn't result in better output for me.
Currently 9 times out of 10 I just say to the model: xyz is the problem/bug/feature, fix it. Since about Fable and Opus 5, this is more than enough. Opus 5.5 (and previously Fable 5.1) got even better at this. However, this is in a codebase where there are already a few hundred thousand lines of code for the model to look at to see how we generally attack things in our codebase.
Claude Codes plan mode I never use anymore, it was useful a few months ago because the models had a tendency to just start doing work and forget I specifically told them not to. But the UX is just annoying and the models now do adhere when I tell them not to change anything.
I have, and it always does something unexpected and undesirable.
Plan mode ensures I'm spending fewer tokens on the code-test loop, and more on the arch/design, and allows me to keep appraised of what's going on, while planning for future changes better.
Maybe folks who don't need planning, don't have as much concern for the details, and are happy enough with just evaluation of if it works or not.
I have roughly the same workflow, also with plannotator - which I like a lot - and haven't used or felt the need to use _plan mode_ for at least 3-4 months.
Then telling Claude to work on a document, the instruction is kept to its core.
Now when bcherny explicitly mentioned that it merely adds a single line - it explains why I don't need it.
What may be concerning about "super plan" mode from the creators (or a skill, for that matter) - is that tuning the amount of effort, and how much deep to dig - may become too hard, as it will interfere with several embedded paragraphs explaining what to do, how to do, where to do, etc'.
What I do look for is even better plannotator ability to track changes, combining historical comments (like Google docs), and git blame of several "generations" before current reviewed doc.
Couldn't you commit the plan markdown to git to get the tracking you're referring to?
While possible to commit, I'd still loose the comments themselves, and the commit will be at least per changeset, and not connected ro a specific changed section.
Roughly speaking, I'd be happy if plannotator would persist something similar to github PR reviews combined with Google docs comments & suggestions.
I do something similar. But where I change it up a bit is depending on plan complexity I divvy up parts of the implementation to different subagents with fresh, only relevant context for whatever they're responsible for doing (e.g. a part of that plan). Edit: to clarify, I will also go back and forth with the planning agent making sure edge cases are covered. Sometimes that involves invoking a new subagent without our planning context to validate it without preconceptions.
I personally still find planning a valuable mental exercise; it's not so different from pre-LLMs and whiteboarding or otherwise taking the time to consciously plan a set of work.
Yeah, sure, but you don't need a dedicated plan mode for that at all. you can just do it in auto mode, and say "let's do some planning first", and Claude will (nowadays) be smart enough to understand that it's not supposed to jump straight into the implementation.
So again: You don't need plan mode, auto mode works just fine, there is no difference in the workflows here.
> write down a markdown file with its plan, then review the plan using plannotator
How do I use plannotator to review an arbitrary markdown file? It always opens the Claude Code plan file for me.
On the quick start page it looks like you just pass it as a CLI argument: https://docs.plannotator.ai/open-source/start/quickstart
I use similar approach, but use gpt to review plan written by opus.
This is pretty much exactly the flow we built Nuanced.dev around.
That used to be my flow but I stripped it all out with opus 5.5 and it's working great. You don't need all the process anymore
I do something similar but a bit more involved, using a few informal stages. Let's say for example I'm trying to launch a new complex feature for https://coderba.se.
- strategy document
- "sprint" document with technical implementation
- actual implementation
- e2e testing scenarios updates
Every step involves iterating with Claude on it with me in the loop (setting the direction then resolving the "founder questions" as they appear), and importantly a different model for review/code-review, be it Codex (usually, it's great at it) or Antigravity/Gemini (sometimes finds novel things, its precision and recall are abysmal but on the odd occasion it has good accuracy). This iteration on the high-level plan then on the implementation plan is essential to me, and IMHO part of why people are surprised that I tend to get solid results from LLMs. At the very least, it allows me to fill gaps in my own knowledge (primarily front-end development) and be more productive than writing the code myself. I cannot stress enough how nice it is to have a partner in the high-level system design – yes, it often suggests utterly moronic ideas, but the overall experience is still net positive and getting better every quarter.
both points to the same outcome: plan mode within harness is dead.
Only because the harness’ current plan mode is too limited.
It's useful because it let's me see the decisions the model will make before it wastes a ton of time implementing them. The model is smarter now but that doesn't solve for underspecification if it guesses my intent wrong
Interesting, I don’t see this very often with the latest models. Are you using Opus 5.5/Fable 5.1?
Either way, plan mode isn’t going away. You can always /plan or ask Claude to enter plan mode. We might re-map the shift+tab keyboard shortcut to something else by default for people that don’t use plan mode.
> Interesting, I don’t see this very often with the latest models.
Strange response. I agree with the parent comment here, plan mode lets me ensure that I have specified everything correctly before it gets built which is far too late. I don't see how an improved models even matter to this workflow. Is Fable going to read my mind?
What is it with the snide responses? This threads title is "Plan mode is dead", not "You don't need to plan anymore".
Boris is saying that you don't need /plan to get the model to plan, you can just say "Let's plan this out" or similar, which at least matches my experience. Your experience may differ, of course, but it's not even clear we are talking about the same thing.
Maybe some people have not been long enough on this rodeo: This used to be an actual issue. You told the model "DONT START CODING YET" and yet, surely enough, starting to code it did. That is what /plan etc were supposed to fix.
I also use Plan Mode. How does it otherwise work if I am not even quite sure what exactly I want to build myself?
> if I am not even quite sure what exactly I want to build myself
Sounded to me like you need a plan.
My approach is to take the statement of work or problem definition and iterate on that myself until I'm really clear on what is the goal. I therefore have a good some good ideas about what the plan should be.
If extending an existing application, which is usually the case, then make use of the plan documents that I had written before AI arrived on the scene. These sre documents in markdown form that say step-by-step how to, for example, add a new report to the system.
You tell it, "let's brainstorm, don't implement anything yet". Then you talk about what you want to build and hammer out all the corner cases. Then you tell it "now do that".
I have something like this in my default rules which Claude has consistently always loaded before doing any work. Works really well!
> How does it otherwise work if I am not even quite sure what exactly I want to build myself?
It makes some plausible choices and you can retroactively ask it to make different ones later, if you want.
My experience with that is not very good. It gets so hung up in its initial decision. Like, it won't make changes because they're "breaking", for something never even committed. Or it will litter the code base with defensive code and comments about the path not chosen.
This is my experience as well. Changing the initial decision is a hassle so I just start over most of the time.
> My experience with that is not very good. It gets so hung up in its initial decision. Like, it won't make changes because they're "breaking", for something never even committed.
This hasn't been my experience.
> Or it will litter the code base with defensive code and comments about the path not chosen.
I've definitely seen this, though.
Edit: Deleted.
I love when people on HN think they're special or do more "real" software engineering.
Basically every big tech company maintains "codebases with millions of lines built upon decades." Talk to your friends at a FAANG and ask them how they're using Claude/Codex.
Merging after reading the PR description is just how it's done these days, and if you can't do it reliably, your harness, devloop, or model is simply behind the times.
>when people on HN think they're special or do more "real" software engineering.
But you're doing the same with your "how it's done these days".
These days things are done in many contradicting ways, and it probably will take at least a few years to settle on common normal.
The model is irrelevant, plan mode helps us ensure the model will actually do what we have in mind. The best model in the world can't work around misinterpretation because of bad specs. I'm not sure why this is even a discussion, isn't it obvious to anyone actually using LLMs?
This is an out of touch with reality thing to say. Remember: not everyone has access to unlimited usage/money to spend tokens or the most expensive models with the higher thinking settings, exceptions does not make the rules.
> This is an out of touch with reality thing to say
I mean, you're discussing this with a marketer/someone wearing a marketing hat, who works for a company which needs people to use as many tokens a possible. That's their reality
[flagged]
"Marketing hat" is extremely civil. Back off.
I see you didn't disagree with the main thrust of my post.
You'll also note their 'About' is empty. Regardless, engaging with HN in this manner is de facto marketing/PR. Trillion dollar companies don't just let anyone post on high-profile social media websites for any length of time without permission from marketing/legal/PR.
Engineers can nevertheless also be marketers for their employer's products.
Hey, well first off, congrats on making the greatest product ever probably.
I absolutely see fable and opus 5.5 misunderstanding intent, but that just seems to be a feature of necessarily underspecifying in a written prompt. Just today, I gave opus 5.5 a simple task to spin up a new environment for work. It read the ticket, which was decently specified and knowing the codebase as well as "Ghasp... reading the code" I had to correct it about 5 times to do it in a way that I would have expected it to. Getting the pipelines right, environment variables, and configs. It was all relatively straight forward imo. Then I had to prompt it to clean up its corrections, because it left a workflow variable in the github action that some intermediate step required but the final solution didn't. I definitely would not have caught that if I didn't read the output. Idk, there seems to be a natural limit as to how much it can infer and I have no idea how to fix it. I did write about it [here](https://javiergonzalez.io/blog/the-assumption-problem/) though.
I’ve been having similar issues. Absolute love fable but it keeps leaving development servers running that are blocking port 3000 (rails apps run on this by default) and then when I try to launch the app and realize the port is in use I ask fable what’s up and it says sorry I left x running and then shuts it off freeing up the port.
Theo would have a field day with this, should surely read your article too imo.
Why would he have a field day with it? It seems he oscillates between finding models the greatest thing there is to being absolutely stupid so not sure in which direction you're implying he'd have a field day.
I'm curious. Who is Theo?
Theo is a YouTuber & prolific X tweeter who has a channel about AI. This is not Theo, but gives a flavour of Theo's style: https://youtu.be/h1p9zdUtUdo
thanks for the info. I'll politely pass on the video though. priorities
I don't think me giving the model bad instructions is something a smarter model can solve. I use plan mode constantly (with opus/fable), and at least once a day I'll say something too vague or just dumb and it will sketch out the "wrong" solution in it's plan.
Which is fine because it just put together a plan and didn't spend 10 minutes rearchitecting everything.
I use fable 5.1 (tried opus but it lied to me 3 times in quick succession and ignored me in another)
This is weird to ask because I feel like of course the model isn't omniscient? Isn't the whole point of iterating on a plan to assess impact, risk, know your (the user) variables, user impact, product impact, etc for making a change? I cannot count the times even in the past few months where I start a conversation with my C suite because their desired outcome would have a potential negative impact elsewhere for other products or users.
Is this just not something that comes up at Anthropic?
It happens with Opus when I leave too much up to interpretation and the AI doesn't do what I had envisioned but didn't specify. Like sure, what it did may be a technically correct solution but it's not the correct solution that allows for further development of my idea. I'm not sure how others do their projects but I start small with proof of concepts and develop in layers until the project does what I want. I use plan mode first to layout everything I can think of that I ultimately want and describe features in the best detail I can manage. I work with Claude to figure out the best framework or find whatever existing projects can serve as a starting point. The first milestone is the proof of concept, take the framework/existing project and build something that does the bare minimum of what I need in the way I want it done then build a test suite to make sure it works. Once that's proven out, we start adding more features (both mine and the ones Claude has suggested) and adding/revising tests along the way. For small things, I won't bother with a plan since I generally already know what I want or any ambiguities can be solved in a single response. But for larger things, I try to take a waterfall approach with well defined milestones.
If I knew exactly how I was going to build something, I would have built it myself. But since there's some ambiguity in the portions of the project I'm less familiar with, I rely on the plan to not only help me understand the decisions Claude has made for me but to keep Claude constrained to the decisions I've made. It's very frustrating to waste tokens on having to refactor something because
I can relate to this. But this part doesn't make sense to me.
> But for larger things, I try to take a waterfall approach with well defined milestones.
> If I knew exactly how I was going to build something, I would have built it myself
Aren't these contradictory? If you don't know exactly what/how to build, how can you do waterfall?
Yeah IMO there's two entirely different traps I've seen in startup codebases in the last year or so:
1) "Maybe waterfall works now" - plan mode, take care of all the nits and issues that the bot leaves on your PR, wildly overengineered "enterprisey" solutions with a lot of bells and whistles all over the place but very poor end-to-end user story test coverage that results in user experiences with a lot of good test coverage of the edge cases of how a given step might fail but little thought towards overall user flow and throughput. Because part of the issue with waterfall was assuming you could design the right tool for your user up front.
2) "Maybe code doesn't matter anymore" - The just ask for something when you need it approach, which results in weird janky individually-sorta-working but strangely-overlapping six-variants-of-the-same-thing that makes it hard for your users to develop a single consistent mental model of the thing they're using, and that changes super frequently.
They both end up with a lot of other bad-for-velocity things that I assume are inherent to how the tools have been refined in response to last year's criticism, too. Super verbose comments. Extensive - without much eye toward runtime - low-level test coverage that might miss the forest for the trees and also slows down the next round of iteration cycles. A plethora of new proper nouns all over the place that make the documentation an ouroboros without a good entry point.
I haven’t tested with Claude specifically in a while, but I see this a lot on larger features.
It tends to be small decisions way down the stack that bubble up, or an incoherent data model that can’t handle what you’re asking for cleanly.
Eg I was messing with a state tracker the other day. The state tracker assumes a container is either currently running, or fully removed from disk.
The LLM chose to remove the state file when the container is stopped and then to remove it after, which leaks container storage.
The LLM is kind of stuck though, because every option other than “rewrite the data model” has negative outcomes and it probably violates user expectations to launch a massive rewrite there.
I've found Opus 5.5 is quite good at surfacing and clarifying these issues using Grilling[0]. Often I find that I want to expand the scope of work much larger than Claude would consider based on my original request.
0. https://github.com/mattpocock/skills/blob/main/skills/produc...
We got Opus 5.5 just a few days ago so haven’t had time to learn its ways, and Fable was/is too expensive for many people including myself. Insiders have months to learn the ins and outs of new models.
It would be nice if with new model releases claude code also gave a bit of a model 101 that tells you evolving ways of prompting it that the insiders have picked up. I know there’s the prompting guide in the claude docs, but this is often very broad and most people don’t know about it.
I have primarily used Fable 5, Fable 5.1, and now Opus 5.5. I never use shift-tab to enter plan mode personally. My general workflow is chatting with the machine to work out fiddly bits, then /plan to get everything in one document, and reviewing the plan.
/plan is still useful, I still need to review what it's going to do and still make revisions. But there's two phases: hammer out key design decisions then write and amend the document.
For me, every day, although it fails far less often at this than Opus 5 did, which might as well have been a gremlin. It was a really poor technical writer too, leaving slop reminders to itself in all prose, including comments. So even if 5.5 fails less right this second, I have trouble believing that it all will not similarly bite me next week.
Summaries that don't tell me when it's changed direction in a timely fashion, but I am only told way later, when I have to undo. Really bad judgement calls regarding where to fix bugs, changes in implementation decisions, taking action when I am asking a question directly, not passive aggressively asking for action... when 5, 3 days ago, was proven to be untrustworthy, switching to very little supervision sounds like a strange thing for a customer to do.
How could a newer model be better at making up information? Do you understand basic information theory? Or maybe what you're saying is you don't actually have any ideas and are happy to do what everyone else has done before you so it makes no difference?
It’s just surprising how few people understand this. It’s not like it’s new either. Polyani’s theory of tacit knowledge captured this back in 1958
The older models were less good at inferring intent. The new ones just seem to do a better job.
They're just more subtly wrong when they are
That sounds like a contorted way of saying "better."
Hi! Taking this moment to gripe; forgive me...
5.5 and 5.1 have major Rain Man (savant) syndrome. Excellent at many hyper-technical things, absofuckinglutely boneheaded at anything that a human (or an earlier model) would understand - like how to write copy, what a human would expect in a given situation, various types of norms...
it's infuriating because it's a sophies choice - dumber model but better human understanding, or better technical model that you have to explain things to over and over like a toddler.
Why is “are you using the latest model?” the automatic response to any even mildly critical of LLM coding?
OP wasn’t even actually critical of LLMs, they were just saying that plan mode was helpful to stop the model from making incorrect assumptions when you want you don’t specify everything you should.
You can just tell it to write out a plan.md file.
I greatly prefer this, since it lets me iterate on the plan with Claude for a while without it repeatedly asking if I’m ready to implement the plan.
Once I’m satisfied, I usually start a fresh session and tell it to implement the plan.
For smaller plans, you don’t need the file. Just ask it to come up with a plan. I don’t recall the last time it just started implementing if I only asked for a plan.
This is the process I generally use too. Small plans you can just ask for, and big plans you work through building a plan.md file before you build it.
That also makes it easier to adversarially review the plan (I have Fable write the plan, then review it with Astra and another Fable instance).
But that’s just the value of planning, not having it be a special mode.
I guess that You dont need plan mode any more is ambiguous. I took it to mean you dont need to plan because the models are so good at infering intent. You are taking it as "you dont need a dedicated modality to create a plan".
> wastes a ton of time
It wastes a ton of tokens as well and those are not cheap.
It’s not that planning is dead, but rather that planning has outgrown the simple “Plan Mode” feature as models have become capable of taking longer turns.
the thing is it fails on CSS most of time, I experienced it and it takes a lot of time to fix again and again, and ruin the code sometimes
[flagged]
hi I’m the author of the post. I think that’s basically the distinction I’m trying to make.
Historically, plan mode served two different roles:
1. making the agent’s instructions precise enough to execute 2. helping the human understand what was about to happen
I think #1 is less necessary as agents get better. #2 is going the other direction, it becomes more important as the model is able to do more on its own because larger chunks of work are happening with increasing complexity.
Where I’ve changed my mind is the interface for #2. I increasingly think an interactive, iterative workflow is closer to how people actually build understanding than being handed a long generated document, especially one they didn’t author themselves.
The human-understanding problem is very real though
Awesome blog. You're a good writer. I enjoyed seeing your article on AI in 2018. Thank you for sharing your expertise.
Also I hope your delivery goes well. My wife (and co-founder) had a challenging delivery and it really put life in to perspective for both of us on a range of issues (how much women's pain is minimized in the health system requiring stronger personal advocacy than I would ever have expected).
As far as plan mode, I still find it essential in keeping agents on track. I build propelcode.app and have a variation on plan mode I still find useful, happy to trade notes on agentic coding if youre interested.
Thank you! Baby was born in July, but we are actually back in the ER right now because she’s struggling :(
Propelcode looks awesome! So cool to see different people and perspectives shaping this space.
I'm not the person who responded to you but I hope they're able to help her and she's okay.
And thank you for the article - it was a good read. I still see folks in my org playing "throw spaghetti at the wall and see what works" and getting frustrated so plan mode (mostly point 2) has been their guardrails almost as much as for the AI.
I never understood fear and anxiety until I had a kid and was worried about her safety. I am sorry youre facing that and hope she recovers and thrives.
Did you just invent "agile" maybe. ;)
Can't help but think "Doesn't matter if a machine or a human with (even slightly) different background wrote it", maximizing information flow is maximizing common assumptions and "culture" to only have to communicate a small set of current information for the task at hand. Being a team means having built a joint context so to say. This has always been the purpose of design documents and they always were too big or too small. Because you did not write them, but the others. If you only produce code you think they are the past and useless. If you iterate and your team grows, you start seeing the value in always current docs that are containing just what is not in your everyday culture.
All the best for you and your growing family. I had a similar experience recalibrating my values...
Fascinating! I literally never use Claude without plan mode and I find it's basically useless without it, constantly wasting tokens going in circles on irrelevant things. Fable or Opus. I feel like neither has a good sense for how to architect things and if I don't use plan mode it usually wastes hours of time chasing it's tail or implementing kludges on kludges to get something working that would be a much simple fix elsewhere, especially when working on a larger codebase.
Hence why the Anthropic employee is telling you to not use it and just aimlessly throw tokens at a wall. It'll eventually get you there, sure, and consume more tokens. This sounds absurd, but trust that Anthropic (and any large company) is hyper aware of how customer behavior impacts their revenue, and they certainly will try to steer you into behavior that increases revenue.
Also this is a way less removed process that I want nothing to do with. The more removed I am from the process the more I hate my job, get burned out and genuinely wish that Anthropic never existed.
Even if it could "just know" or infer my intent. It wouldnt be desirable.
Edit: Oh yeah its Boris, hes one or the most disengenous shovel sellers on earth right now.
So you think telling a model once to "do not write code, just plan" vs having an enum that causes that same text to be added to each message will somehow waste tokens and will surely make anthropic much better off..
I mean if you want it to plan something without writing code first you literally just need to ask it to plan out the implementation first it's not exactly a difficult concept
It must be fun to see everything as a conspiracy.
The Anthropic employee literally told you plan mode is a "we're still planning!!!" at the end of the prompt. Plan mode is useless because you can literally type "dont write code yet" and get the same effect, not because you should never plan in general.
Sure and I should also have been more specific, I don't always use "plan mode" I often just do planning without explicitly activating it just via prompting, point being that the robot doesn't work for me without some time spent planning out what to do.
I realize there are already a ton of comments, but I think you're missing the idea of _precision_.
Most people aren't precise when initially describing their problem.
Jumping straight into implementation skips the part where we refine and better define what it is we're trying to do, and think through the implications of those changes.
I suspect "trusting the model" doesn't really work at scale with finite resources.
So my Claude Code workflow for a ticket is:
* Scoping discussion - do research and figure out approach (auto mode)
* Planning - take the scoping and convert it to a concrete plan for review (plan mode)
* Implementation - put the plan into action (auto mode)
I find this works really well for my workflow, and it is really easy to trigger each phase because the model has clean boundaries. (This workflow is articulated in my user CLAUDE.md) Plan mode is still useful to me as it forces the model to double-check its plans (I find even with Opus 5.5 it still discovers gaps), and it gives me an opportunity to clear context at a really good spot.
So I would still consider plan mode to be useful. It would make me sad to see it go.
PS- I have a Claude Marketplace directory submission for an MCP server that has been sitting in review hell for six(!) months. I've never received any outcome other than "In Review". I hope I'm not asking too much but would it be possible to put me in touch with someone who might be able to help here. Nobody has ever replied to messages sent to mcp-review@anthropic or the "Anthony at Platform Operations" inquiring about status, and we are getting frustrated
(Disclaimer, I work at github)
I tried plan mode when we first added it to github copilot and it didn’t stick for me, until very recently, when I used it the way you describe. I was just putting it too early in the process, turns out I have to do a bit of exploration and discovery on my own and most of the time I can skip plan but now I have an intuition for when to engage it so it can interview me to clarify the last few things it needs before implementation. This also made autopilot mode work a lot better for me.
> plan mode was useful, and is no longer useful.
> In Claude Code, all plan mode does is add a little reminder to every user message along the lines of “you’re in plan mode, please don’t code yet"
Can't help but think if plan mode isn't useful as you say because it's implementation is lacking in claude code, hypothetically speaking.
What I can say is that with other harnesses plan mode helps stabilize my workflows. Actually synthesizing code is only part of the process, lots involved in taking a work-item to production end-to-end and plan mode helps give this flow structure. More than that it's an opportunity to regroup before committing to changes. It slows down the process to a rythim that's sustainable and smooth, which ends up speeding up the process.
So if plan mode in claude was designed to speed up code churn, while oh my pi for instance designed plan mode to be strategic, that might account for the different perceptions here.
And it's not to say you should force yourself to use it, but if you are planning on cutting this mode off the loop just beware of the possible side effects.
It seems unsurprising that if your definition of "plan mode" is as tiny as appending "but don't write code yet", that it would not be that useful for that long. There have to be more sophisticated versions of what "plan mode" means out there.
Also if you are working in a heavily vibe-coded codebase, as Claude Code reportedly is, it's not that surprising if the human doesn't really understand it or have anything useful to add in a collaborative context.
GitHub Copilot for VS Code by default configures "Plan" mode to disable everything except reading, asking the user questions, and writing into the memory scratchpad. Claude Code's tool system feels quite rudimentary in comparison, but that's what you get I guess when you're only working with models you're going to finetune on being able to handle specifically your own tools anyway.
Plan mode is critical in the beginning, because there's a lot of long tail decisions that haven't been considered, let alone resolved.
But as skills and memory are populated over time, Plan mode isn't as necessary. It becomes simpler to let Claude just build and get something general in place that works, and then refine from there. Auto mode will ask essential questions.
I still use Plan mode for big feature changes, to confirm that I've asked for what I want in the right way. I tend to prompt casually, with only a few specific details. Plan mode helps me see the whole picture before committing. In a few cases, it also helped me decide the feature I asked for was wrong.
I don't use Plan mode because I plan out work with a ticket based approach (that also works as a memory store), but I wish there were some kind of Plan mode like harness feature to try as hard as possible to stop it from beginning to code. With the newer models, I have started to have a lot of trouble with them proceeding through the QA with me where I scope out the work and refine requirements and then immediately begin to do the work rather than either making tickets or giving me the full plan to thumbs up like I tell them to do.
So it sounds like we use it somewhat similarly, just taking a glance at it before the work starts, and that's becoming harder to do in my experience. It is important because problems with the overall plan or strategy end up magnified the further you proceed with development.
I disagree with the assertion that the model gets it. Here’s a practical example I just tried with Fable 5.1. I gave it this prompt: “Write a Go function that can be used to establish secure communication to a remote system using a certificate. Keep it short, single function, and explain how to use it.” The output forced the use of a private key stored in a file even though that wasn’t specified anywhere as a requirement. The function Claude wrote takes a private key file argument and calls a Golang function that requires a private key file (tls.LoadX509KeyPair) even though Go has crypto.Signer which could support private keys in various other manifestations like HSM or KMS. I argue that a person who “gets it” (or who is reasonably experienced in security) would have opted for not requiring private key material for this to work.
For the record, this isn’t unique to Claude. ChatGPT and Gemini do the same, each with its own quirks. ChatGPT got extra credit for being the only one who allowed the function to also take a CA file for server authentication.
Don’t get me wrong: LLMs are the future (maybe even the present) of software development but I think there’s some way to go before they can be entirely hands-off in some areas. I still find myself having to course correct designs and plan mode helps me with that.
And of course, thank you for your work on Claude. :)
I'm hoping that by "gets it," he meant that if you start a discussion about the design, it doesn't misunderstand and immediately go off to do the work. Some models tend to do this.
Whether or not a distinct "plan mode" is needed, upfront planning remains essential in my experience, even with Fable (albeit not the 5.1 version). I agree that, as the models get better, you can skip planning on increasingly complicated tasks.
But there is still a ceiling above which it is necessary to "preload" the context window before starting to call tools and get into the meat of the work. You want to establish domain language (especially with Claude models which otherwise will invent their own, and it will be inscrutable) and key requirements and assumptions. You want to do a Q&A iteration cycle with the LLM. You definitely should do a sanity check that the LLM actually "understands" what you were trying to achieve, and then make sure that understanding is coherently and plainly stated in the prompt. All of that seems to be necessary still for just about any serious task, if you actually care about the quality of the results and/or don't want to burn hundreds of thousands of tokens on flailing around to get to a good quality result.
So no, you don't "need" plan mode. But you do still need to do all of the things you would do with plan mode.
But in your example you never even asked it to plan so you could check the implementation before writing it you just asked it to write it directly, so this isn't even a comparison to plan mode
This is a pretty obscure and in-the-weeds benchmark, but to me the models’ interpretation feels quite reasonable.
Apologies, I didn’t mean to imply it’s a benchmark, I just wanted to provide a reproducible example of where I see models make decisions that seem to be fine initially but might paint the software architecture into a challenging corner. I don’t expect models to read my mind, but I do see them produce a lot of verbose output, none of which is used to say “here’s a simple response to your ask, but have you also considered...”
It's obscure to use common functions from the standard library?
It’s just an odd way to judge whether an entire model “gets it” in my opinion. There are so many variables.
I think knowing which models use the standard library and which ones pull in dependencies as highly useful. As a dev I've always favored staying as close to the standard library as possible as it makes refactoring, understanding, and deletion much easier. Not everyone has the same preferences as you, it's nice to understand what matters for others too.
It can't read your mind that you have an HSM, can it?
[dead]
> In Claude Code, all plan mode does is add a little reminder to every user message
I was quite surprised when I learned this (when Claude Code edited a file despite being in plan mode). I had previously assumed "plan mode" was a harness level concern, and restricted what tools could be used. I didn't expect that it was simply an addendum to the prompt.
But I think it gave me some good insight into where the heads are at of Anthropic employees building this. Basically, leaning on pushing everything to the model. That's why alignment is so important: a "sufficiently advanced" model doesn't require any tooling infrastructure around it, and I suppose Claude Code devs are targeting that future. I had previously thought there was more to a harness, but with "auto" mode these days, it seems like there's no desire to build in that direction.
If you want to see an even more heavy/"extreme" example of this, there was this HN post & blog some months ago on Claude Design.
Basically, Claude Design is primarily a huge prompt.
That's all.
There's no "magic sauce", and with the right tools to call, even your local LLM can implement "Qwen Design".
Good to hear, this is my experience also:
I haven't reached for Plan Mode in awhile--maybe, on some blank folder/canvas and I just want that cute "questionnaire" DX to get me going...
But on the whole, CC is smart enough to know when to "rush off and act" and when to "pushback", which is great--and it's no big deal to tell it to pause/stop by adding "what's your thoughts?" or "feel free to pushback" etc to my prompts to make it start a back-and-forth with me (the fact that you say that Plan Mode was really nothing more than a prompt anyways is reassuring).
And yeah, when you're deep in the weeds, you could (and can!) have multiple threads of thought/work going on in the same convo, that stopping and starting a plan in Plan Mode seems like a regression.
(happily using Claude Code Opus 5.5 on High rn)
I get the idea here, but I find planning very useful as a phase when I'm doing things I am not intimately familiar with (e.g. AppKit programming).
That's when I need to learn about implied patterns, do's and don'ts; not from theory but in the context of my own project. Unfortunately, the model tends to keep implicit knowledge implicit. But I can ask during planning.
The feature did get less useful over time when the model started babbling in newspeak more and more. When it threw a thousand words at me even in concise mode.
So I don't want to let the toolmakers off the hook here. There's a lot to win that would make plan mode much much better without changing plan mode itself.
You can still do planning as a phase without using an official plan mode you literally just ask it to plan first...
To be fair, the team also neglected the functionalities of plan mode. It was unusable for since july when they removed the option of running /clear before implementing a plan. Also, there was a period of time during the dark periods of opus 5 and nerfed fable 5, whereby if you didnt explicitly request for code, it went on an on for a few sessions planning and researching in a loop.
for me it basically all boils down to:
1. I dont want to have to accept every time Claude touches our DB
2. I'm scared out of my mind it might do something bad to the DB
Plan mode gives me enough confidence that it wont do (2) --> allowing me to give it enough permissions to do (1)
Oh gods, I don't give it write access to my actual DB.
For my small-scale sqlite dB, it gets read access, and I encourage it to test modifications by copying it somewhere and writing into that.
Scale-dependent, but I hope to not have to work at a scale where it gets write access to the production DB. That just seems like asking for trouble.
FYI I had Clod attempt to corrupt a prod db the other day. (Opus 5)
I was experimenting with a rather complicated backfill operation, were I had a validation script I understand and have Clod come up with the backfill script. I was running against a local prod copy, and it proposed running the actual (unfinished) backfill script against prod.
It didn't have access to the secrets and I also caught the command, but a good reminder that this stuff needs guardrails.
yeah that's crazy. its so good 99% of the time but I've seen it have some insane hallucinations before (as late as Fable... cant remember if it was Fable 5 or Fable 5.1).
Hallucination not a big deal when it's on the surface layer. But I can't imagine the damage it could do if it hallucinated while building/validating a "load-bearing" component and then continued down that path
Make a db replica or just a db user account with readonly permissions, and have only those in your env, or docs accessible to agent. It's liberating.
Do you mean when you're making changes to a production DB?
yeah (at an early stage data-focused startup). I think when our product is a bit more mature we'd probably have a staging tier with full access and then manually promote builds / DB changes to prod. So many things to build lol
we DO daily snapshotting, so the risk is limited... but still spooks me
> data-focused startup
What could possibly go wrong with building a data-focused company on a foundation of violating the most fundamental precepts of data management
Write schema/data migration files [with agents] in version control and deploy those deterministically (after review).
> plan mode has always been a prompt — it has never changed the toolset because doing so would break the prompt cache, and so would be expensive
seems like plan mode could turn off some tools, even if it doesn't change the set offered to the model, the ones that they have which would mutate your codebase could just not work with an error message, and plan mode could change permissions in the security approval prompt for "auto"
anyway, isnt the right way to know if plan mode helps or not, to run an experiment? we're all guessing unless we have data
read only agent mode sounds straightforward and useful to me
It’s bizarre to me that Claude Code doesn’t have more guardrails on plan mode. I mostly use Copilot and the plan mode there has teeth. In plan the agent does not have permission to write to the filesystem outside of a temp directory and the plan document itself, and tool calling is heavily restricted.
There’s also a “ask” mode which is read only. Both are enforced by the permissions model, not just a system prompt instruction. I’ve seen the model “forget” and try to start coding - it bounces off a hard permissions failure and that “reminds” it that it’s in plan mode.
I've had Claude generate me videos of before & after to better convey new features or bugfixes, that's been quite helpful.
In case anyone else is interested, the skill is public: https://github.com/Mudlet/Mudlet/blob/development/docs/demo-...
you must have tokens to burn.
Early versions of plan mode would erase the chat history and present the plan as the start of a new session when you approved it.
I miss that. It worked really well, and it kept the context clean.
In my experience, around the time the author describes as starting to not need plan mode is when erasing the chat history became an anti-feature. I found the agents were doing better when they had the context already, and with the history I no longer needed to micro-manage persisting various caveats and rejections to the plan artifact. The amount of prompt construction necessary went down overall.
By the time my plan’s done I’m usually between 200 and 350k context. Even if keeping that around gives a performance bump for the implementation (which I haven’t noticed to be the case) it balloons the cost. I would much rather put everything in a plan file and start fresh.
Plus, I usually plan with a more expensive model and guide implementation with a cheaper model (with smaller validation calls back to a more expensive model)
that still essentially works. ask it to write a plan doc to a file. then when you're ready to implement, start a new session with a prompt to review the plan doc and then start building.
thats a feature flag you can gurn on still
What I do nowadays, for large changes, is have Fable create a HTML explainer for what we're gonna do with code snippets, which is not that fat from plan mode only much more convenient for me and modern models have no issue using that HTML artifact as the source of truth, and then before I get into execution - I interactively build an end to end test that also includes pieces of the plan.
When a test case fails, the relevant part of the plan is surfaced in the error. I find this helps Claude stay on track for longer - I've been able to do 12h most times and even up to 48h unattended (11h of API time) with good results.
Then whenever I do check in, I ask it to update the HTML with current state in an append only fashion (sort of like it's writing a blog), and then based on that, we iterate on the end to end test (I think of it as a "test harness") - update the test cases and error messages.
I've been able to build some truly large projects this way, both greenfield and up to spec (for example, a video game I've always wanted to play), and brownfield while staying within the conventions and design of the codebase, and with very little attention required on my part.
Plan mode was great, but I realized I progressed well beyond it. I found that I was getting these categories repeated errors and oversights from Claude (and frankly it hasn't gotten much better about this). Skills were too generic and got lost to context.
I ended up building out tool an MCP server that I use as a bit of a psuedo harness for Claude. I have a variety of multi-step workflows that are basically micro-skills stacked on top of each other. This helps me make sure that I can get Claude to think in a repeatable and reliable manner.
For coding, I've found that I have a few specific steps that Claude needs to do before I'm comfortable letting it loose:
* It must extensively explore the code base (including certain areas that it misses)
* It must think about what it doesn't know or is making assumptions about
* It MUST scaffold out it's intentions. Essentially, it can write comments, classes, and method stubs - but no actual content. Very much like a spec, but since it's in and alongside other code, it's much easier to identify problems.
* It must spike and validate key assumptions. This, plus the prior step, are the only way I've figured out how to avoid it ending up in a confusion loop. Too often it looks at poor-quality code it's written and thinks it's a long-term solution. By avoiding writing code as much as possible, it knows that it's draft content.
* Only, then can I review it and send it it.
Said MCP server (missing the actual ops): https://github.com/clops-mcp/clops-mcp
I use plan mode, because my current project benefits from "pair programming."
I have one session define a task, and provide a formal specification plus context in a "cover letter."
The session B, in plan mode, produces the plan back.
Session one reviews the plan and clears it, ratifying portions and often specifying specific changes.
Session one then executes.
What has been striking to me in this approach is that even with two instances of the same model (currently Opus 5.5), there are regularly corrections made. I use "project chat" for session A and Code for session B atm; it is very typical that Code finds and corrects details or oversights in the task spec; it is also typical (though less so with 5.5) that session A (chat) pushes back or clarifies things Code doesn't have the context for.
I have been afraid to open up the potential of negotiation beyond what this is costing as it is. But I am also afraid to simply skip the formalisms, because of the consistent correction that occurs in this back-and-forth.
Each component of the pattern is schematized, generated from a template, and validated, to keep things tight.
Lots of tokens! But I trust this process far more than "just typing" :)
I am curious what your take is on skill packages like Obra/Superpowers and grillme and the like. On one hand I find them extremely useful for asking me questions. I wouldn't have thought of myself or that I don't feel I would have necessarily gotten in a structured a format just by having the AI interview me. On the other hand, I am not sure how to tell when they have become obsolete because I'm not sure where to even start with generating evals for such a thing.
Yeah I think this is a natural consequence of longer task horizons. When I was chaining 4h tasks, I can mostly plan them up front.
Now that I’m frequently designing and delegating day/week scale features, the flow has to change; having the agent go off and build a spike can be a quicker way of us understanding the design space and constraints (especially in a huge codebase). I still have the agent write and update a spec doc as I go, but it’s not waterfall anymore.
At least for my kinesthetic learning mode a rough code PR stack is usually way better than a plan doc anyway, and tokens are cheap enough (vs my time) that going further than just a plan is often cost-effective overall.
The dream of course is (say it with me) loops, but that doesn’t tend to work for me on new features often.
Plan mode is what makes it easy for me to update my mental model of the codebase, and helps me decide if the mental overhead of all the changes that are needed are even worth doing, which are now the biggest limiting factors (reading the code diffs just doesn’t scale anymore).
Honestly I don't agree with this. A few reasons why I feel like I will always want a plan mode:
1. I want to know whats going to happen, at least at a high level, before changes are actually made. 2. Plan mode helps me flesh out the missing details of my plan before being mid-execution 3. In situations where I have a limited budget for AI usage I will often times use a high powered model like Opus 5.5 or Fable to make a detailed plan, then scale down to a cheaper model for implementation. I feel like this saves cost in the end.
I get plan mode is basically just a small hidden prompt. I get that I can basically just preface my prompts with "make a plan only, don't make actual changes." Maybe this is just a UX trick, but it works well for my brain.
I'm not an Anthropic model user, and the true frontier of the frontiers is beyond my budget. Maybe it's better in the rarefied atmosphere of Astra, Fable and Opus 5.5?
But with GPT 5.6 Sol, I'm still finding that the model makes conceptual mistakes, or gets edge cases wrong, or assumes incorrectly (making an ass out of both user and model). In many cases, I need to at least refine the proposed approach, or amend, correct, or flat out just stop and start over. Not planning and catching these errors, and just letting the agents code their code, would mean I'd have to rollback and redo many times. What a waste!
For a current project, which is ~33k lines of code, I'm also finding that I know the codebase better than the model, and that's vital at the planning stages too. If I wasn't in the planning loop, the model would have reinvented various wheels a few times over. How much spaghetti do you want with your code?
As always, I may simply be doing this wrong. But I'm personally not convinced that the plan is dead, or that I want the plan to be dead. Planning is also good for me -- it keeps me thinking about the code, prompting better, guiding the model better.
If I'm no longer on top of the codebase, then at some point my prompts will devolve to "Do the thing with the thing, that does thing". And I don't want that.
Agree, and for me I feel like I often have more implicit intentions than I write in a prompt. A plan helps me verify whether an agent gets these right or not. Plus, it highlights tradeoffs I might've not thought about. Removing both feels like lowering a quality bar.
On the other hand, for a low effort hobby project: just do the thing.
It is absolutely true that Opus 5.5 just ‘gets it’ far more often than gpt-5.6-sol, which is more like an idiot savant. It can nearly always do what you ask it to, but that might not be what you want.
Yer, it’s probably best to face reality and understand that for actual software engineering / complex coding work - Anthropic models are way ahead of OpenAIs…
As someone that never used the built-in plan mode, but did use a lot of spec-driven development, I’m still finding that even with Fable having “plan” docs is still quite helpful.
They’re most useful for broad changes (new features, refactors, etc.) where it’s helpful to avoid breaking changes or unnecessary scope expansion.
The new models are great, but they do more by default, which means I’m finding myself explaining what _not_ to do more often than with previous models (where they’d often end too early).
In my case, the previous plan mode was too ephemeral, and I like having one source of “truth” that sits across context windows without loss/compaction.
To be honest the new behavior is convenient when I just say "X is broken, please fix". But I find myself more and more adding "please do not do anything yet, just tell me what you would do" and please feel free to ask if anything is unclear", especially if is making decisions that are a fork in the road.
I would really like a dry run mode that just disables all external commands from the outside so Claude doesn't proactively go about changing things.
I like plan mode personally. I only use claude code for the web, and the questions claude asks me to clarify are usually pretty important - mostly because I was too vague or contradictory in my prompt, or what I was asking for conflicted with something else in the code. I don't know how claude would resolve that without plan mode.
Also for session planning, as in when-can-I-walk-away-from-computer, its nice to know the particular rhythm of initial crunch - ask questions - make plan - do it. Especially with a 5 minute cache timeout.
what worked for me is a frontend driving the agent, capturing every user message, every commit, every pull request, putting them in a graph (uses a frontend because I didn't want to load a coding agent with tools + responsibility of book keeping) and the agent get tools to search reason behind code changes and see the high quality user input underneath instead of the sloppy self written justifications
My team has been struggling to understand whether or not we should do "spec-driven" development or not. It makes a lot of sense to me to have one developer iterate with the model/harness to generate a markddown document that is a high-level of what will be implemented, and then have the team PR review it before and agent attempts to do the actual implementation work. Do you think this is a good practice?
I specifically work in tooling now, so this probably applies more to that domain than some others, but I find 'very up-front spec-driven development' unappealing for that kind of work.
I'm experimenting just like everyone else, but this is my process right now:
- Quick prototype
- Figure out the language of your app (what terms you want to use for things, what your UI design language will be, etc) and spec that, so you can use words consistently with the agent. You need to be able to describe the things you want well and consistently.
- Keep prototyping. Let the agent write unit tests along the way. Lock down behaviour you like, keep track of those things in a document.
- At some point your idea of the real architecture comes into focus, from actual use cases -- avoids the over-abstracting right away trap.
- Refactoring is cheap with tests, so start refactoring into the architecture you want.
- Your architecture won't necessarily be what would be best for a human, but it will be pretty close.
- Keep relentlessly iterating on small work.
- Things that were expensive before aren't that expensive now -- integrating a library, changing from one library to another, trying out a few architectural refactors, trying out different performance optimizations, etc. That stuff is all 'throw it there and see what sticks' now, so don't be afraid to try stuff which felt big before.
I feel like 'front loading' too much is just the wrong approach. You might feel like you're sitting there 'babysitting the agent'; but that's just what the hard part of the work (hard as in 'zjust slogging through it', not as in 'conceptually complex') looks like now. Your code is much more like clay.
Atleast that's how I'm thinking about it so far, but I'm not working on large sprawling systems that I imagine would need more pre-planning.
I believe spec driven design is a good way to go. But that means keeping all your specs either in the repo (if you want the agent to be able to edit them) or available via for example MCP (if you think only humans should be able to edit them and you got some type of external documentation system keeping specs).
But given that running an agent us cheaper than the cost of waiting for a slot to assemble the team to talk about a change (isn’t it always?), why wait with running the agent?
I propose updating the spec then do the implementation. This will most likely show that a few assumptions were wrong forcing some major or minor updates to the spec. Work through those and then let your team review the spec change together with testing the next iteration of what what’s build.
I have found it doesn't see me any direct time. I either iterate ruthlessly on the spec or on the implementation.
I find that the code is generally in a better place proportionate to the amount of SDD I actually do. But it's just a matter of where and when I want to spend my time.
SDD is good. It's a lot faster and cheaper to have an agent polish a spec, than refactor code. The implementation can then be verified against the spec artifact, and any drift can feed back into process improvements for future specs.
> Do you think this is a good practice?
For who? The more control you hand over to the AI and let it think for you with no supervision, the better it is for Anthropic
I would have agreed a few months ago before opus 5 but I'm back to plan mode and lots of the old tricks trying to get that insolent model to follow instructions. 5.5 so far is better but certainly doesn't feel under my control (had it go off and read other repos and make changes just yesterday while in "plan mode").
I created a skill "feasibility", which is basically "evaluate feasibility of this idea and propose solution options". So I can check that Claude's idea of how to implement the requirement is close enough to my own before it starts working. It is a lot more lightweight than plan mode, because the skill says not to build a detailed plan, just a high-level summary.
> It’s something I came up with late on a Sunday night many months ago
Aider, Cline and many other agents had plan mode before Claude Code existed.
What, are you suggesting that there were people already successfully doing agentic coding in CLI harnesses with a good number of features before November 2025, the moment HN collectively deciding that it was now "good enough", if using "Opus on Claude Code"? Blasphemy!
Yeah, as a user I came to the same conclusion as I naturally used plan mode less and less over time.
I still use plan mode in Astra to come up with a plan that I then feed into Fable. I feel like OpenAI models still do better big picture investigation and planning, while Claude is the better software engineer, if that makes any sense.
Of course this could well come down to my own biases and the specific things I’m working on.
I am not seeing this and I’m wondering why. Maybe how I plan is different then you? Maybe what we each mean by plan is different? Maybe you trained yourself to write more effective prompts than old you? Maybe the problems you’re solving don’t need planning and the problems I’m solving do? Maybe I’m pickier about solutions?
> planning had become interactive and iterative
I find myself endlessly ctrl+c ing claude now as it flies off doing deep first principles analysis to work out how to find a thing it isn't sure about but I know the answer. Being able to give it that answer without needing to ctrl c would be a massive improvement
I always append something along the lines of “evaluate”, “investigate” or “report only” to my prompts when I want to see what the agent is gonna do. Because especially with the new models they tend to go easily off rail and do stuff I didn’t ask. To say that they just “get it” is highly dependent on the task, scope and blast radius.
I really wonder what I’m missing because my general rule to plan first unless explicitly instructed works perfectly fine with all models I use.
“I want to ...” / “Let’s ...” -> Plan
“Do X” -> Actually Act.
But then again I also have it configured to only ever answer questions instead of inferring them to be instructions (which I’ve seen others do differently).
Thanks for the context, and makes sense a lot. That’s my primary reason to use plan mode.
However how about decisions? Do we expect the model to read our minds, just assume the best practice will be followed and that’s what the user want? Plan mode solves those, what is that am I missing?
I frequently ask claude to write and review org-mode files. I've found it useful to maintain knowledge decisions both for claude and myself. Less so for other members of my team who wont use emacs or (n)vim. But thought I'd share.
I urge you to have a look at tools like tuicr and hunk. It’s a missing component in the chat. I really want a proper chat interface, leave comments on the code and have them replied to. No clue how to implant it tho, it’s a hard UX problem IMO
Recently I started using pi as a harness and its flexibility has really changed the way I think about UI into agents. It feels like I took the red pill and have been liberated from the prescriptive nature of CC or codex. Instead of getting told here’s how you should approach ___, I pay much more attention to my workflows and when I feel friction I make the harness adapt and smooth it out.
For example, there are no shortage of web based uis for pi and they are all cool but I wanted a deep integration between artifacts and how I want to collaborate with the agent on them. So i built my own ui that mirrors what the pi tui sees. It’s chat based but has a deep integration with a GitHub style code review UI so I can leave review and comments whenever I want. Every agent message renders nicely in an annotatable markdown viewer so I don’t need an ask question tool and can more naturally get the agent on the same page as me. I want it to feel like I’m working with a colleague.
I don’t think these features are too unique but having full control over the experience is really nice. Flexibility over model provider, can tailor it to my work’s dev stack, and don’t need to worry about anyone breaking it with a million updates everyday.
Your perfect workflow can be realized in a day or so. You just need to go make it happen.
I also didn't like the chat TUI for back and forth on things, especially on long markdown files where I felt like I needed to copy a line to comment on it. I wonder if LLM companies would rather skip a step and aim for a world where most people one-shot things, which sounds much more impressive.
I've been building https://crit.md to keep that back and forth with agents - GitHub-esque interface and have agents respond to my feedback, iterating until I'm happy.
Admittedly like many others, I use it a lot less for actual plans now, models are indeed getting really good at just getting it.
I wodner what this product space will look like a year from now. Reviews are already dying.
Oh I thought it blocked tool calls like editing files...
You don't need to actually change the tool schema or break the prompt cache to do that. In the tool itself you could just check if it's in plan mode and reject the tool call...
> it has never changed the toolset because doing so would break the prompt cache, and so would be expensive for users.
It could still make the tools into no-ops or disabled if it actually tries to use them, without changing the context history at all.
I keep forgetting that CC has access to the "question" tool, even outside plan mode, which is mostly why I still use it sometimes. This isn't (or wasn't until recently) the case for Codex.
Reading the replies, it's interesting how idiomatic everything is, eh? We must all opine about our specific setups, panes, prompts, and how they're the true way, or the true alternative to plan mode.
I don't really use plan mode, but I do have Fable write a lot of docs, particularly when the work involved is long and multistep, because I may need to do it across multiple sessions.
As a long time user of Claude Code, I've just naturally stopped using it because the model figures it out, and i prompt it accordingly like 'come up with a plan'. Glad to see your experience matches
that's the problem though: - LLM now output good enough results without a plan.. for coding at least. I'm not saying amazing results..just good enough. it works fine.
- Most people suck at planning anyway
- LLM still don't give you a way to verify and understand to iterate.. you have to ask and then formulate and way so people barely do it, they just trust the vibe
IMO the current successor to plan mode should be the harness knowing when to tell the use "ok here's our overall current state in a simple diagram", auromatically
Well that explains why I was able to coax it to partially execute on some things without actually exiting plan mode.
I was thinking of making (locally) an “artifact mode” along those same lines. It’s a much better way to see proposed changes
I’m always telling the model “don’t write code yet” so I maybe ill try plan mode if it’s just a prompt change
> so I can better understand the change and alternatives considered. I don’t do this very often,
# YOLO mode
I just tell Claude to write an md file and do research first anyway… makes sense
I think over time more and more will be peeled back to just the model and markdown. I have a beautiful factory running with key personas all it is is a few markdown files it is building a mac app fantastically well.
https://innerloop.test/breadcrumb (for reference)
Good old "Check out my new awesome app: http://localhost:3000"
Is this what happens when you vibe code long enough?
Hahaha
https://innerloop.works/breadcrumb
What a rookie mistake!
Perhaps not that fantastically given the url doesn't load. :)
So planning mode is dead, long live planning?
I m a bit surprised by this as - during the last 2 months - I 've encountered more often models going ahead of what I have asked from them. It's gotten so annoying that now instead of just saying "Let's plan X" I will also say, "Stop once you 've written the plan." because otherwise there is a distinct (~35% anecdotally) probability that the model will go ahead and implementing the plan as well. [This applies to both Anthropic and OpenAI models btw.]
Plan mode is a useful shortcut when I want to have an agent do read-only work without having to worry about giving it appropriate stop conditions.
Hi, thanks for Claude Code. I use it, and it works well. Have you considered changing it so the text comes down from the top of the screen, in green, like The Matrix?
It’s been like that with me in Grok 4.6 for the last couple of months as well.
I just start my day writing about 20 queues /goal prompts and then check the work at the end of the day. It’s almost always right!
For me its like a dry-run.
I ask Codex (I am not using Anthropic anymore) to always generate a plan file first, unless explicitly told otherwise.
I also tell it to write deviations and rename plans accordingly once done.
That way I keep the codebases I have to or enjoy to work on in my head and don't become too dependant on any provider or on stochastic parrots in general.
Btw random question about Claude Code. Is it just me, or has it become a lot slower in the last 2-3 months? I am on Fable but also latest Opus.
except the fact that claude code in plan mode does destructive operations, requiring a hook to tell it not to do stupid things, so its not really a "read only" or "dont code yet" tool, imho.
Plan mode still earns its keep when the blast radius is high.
For a one-file change I don’t bother. For anything that touches auth, payments, or a shared schema I still want the plan written down first — not because the model can’t figure it out, but because I need a moment where I can still say “no” before it starts editing.
The mode was never really about making the model smarter. It was about making the human stop and look.
Now if only the human would write its own comment.
…did Opus 5 write this?
I’ve heard “earns its keep” in only two contexts in my life - the intro to the song “Regulate” and terrible Claude docs
Perhaps it depends on where you live? It is a common enough term in parts of England
It is 100% Claude
> I work on Claude Code
So basically you don't know what the fuck you're talking about. Not everyone is employed to burn money.
I'm actively watching understanding slip away from developers, code review getting paired down to no comment checkmarks, and codebases go to bloated messes that nobody can read. Axioms like engineers must understand and take responsibility for the code they ship are getting torn down, and the products coming out are reflecting conway's law, becoming impenetrably obtuse and always "so complex there are no obvious deficiencies" (as opposed to "so simple there are no obvious deficiencies" which used to be the aim).
The one thing plan mode helped is for the humans to get an understanding of the strategy, and be able to poke around and look at the design and architecture. You can achieve this with some self discipline and keeping shorter leashes on agents, but it feels like a losing battle. The best devs still put out good code, but the poor devs are learning nothing while their metrics look great. I can't help but think we are racking up immense amounts of debt that will very soon become due.
100% this. Some people are basically adding AI as a dependency for their projects. They no longer understand the code
I’m a big proponent of writing simple, understandable code, so I’m playing devil’s advocate here a bit, but: who cares?
A significant (majority?) portion of developers have been shipping JavaScript/node applications for the last decade that contain hundreds of MB to GB of code from god knows where doing god knows what with dependency trees the size of redwoods. It’s not like your average mediocre dev really knew what was going on behind their gluing of frameworks together - at least from what I’ve seen.
If you have remotely competent tech leadership that enforces relatively intelligent patterns (a good one I’ve found is “write everything backend in rust”) you can make AI churn out monstrous amounts of code that… isn’t all that bad? And if you enforce it writing and updating a docs/API.md on every commit/PR you’re probably doing better than 80+% of devs I’ve ever met. Up until a few years ago it wasn’t uncommon to roll up to a new job that was a “legacy” pile of garbage concocted over 20+ years with no comments or API docs and a readme that tells you to ask for help from someone who has been dead for 5 years. At least AI code is full of comments (some of which might even be accurate) and there’s a finite (relatively low!) cost to figuring out “wtf is this doing and how is it doing it”
One is abstraction, one is complexity that _you_ own. Even 100 lines of trad-coded C relies on "hundreds of MB to GB of code from god knows where" in the Linux kernel. The difference is, you can perfectly understand that C, and own it. Then delegate ownership of the rest to Linus. If AI writes 1000 lines of C instead, now there's code no human owns in the world.
> The difference is, you can perfectly understand that C, and own it.
and there is a community of developers who care about the quality of those libraries and take the weight of their responsibility seriously
The point is that saying "I own the code" is just drawing an arbitrary line in the sand.
You can give the AI the spec, and own the spec instead of the code.
Eventually, the spec too will be something the model owns, and you'll work at a higher level of abstraction.
The longer you work in a large engineering organization; the more clear it becomes that no amount of processes, documentation, documentation management systems or training programs can actually transfer the institutional knowledge of the origination to new people other than people-to-people interactions.
Beyond a certain size, the documentation becomes too large to ingest. Below a certain size, it can only contain a fraction of what is needed. If you take any sufficiently engineering project, and give every engineer amnesia; the project will go to shit for a undetermined amount of time, as it takes months or years go build-back the understanding that was lost.
This is clear enough then large companies fire and replace workers randomly to cut costs; a worker that has built up useful knowledge in the origination over a few years is more valuable than three cheap consultants from "low-cost countries" that are fired when the work package is over.
---
AI, looses its memory every time we press "new thread". No spec can bring that back until AIs become able to write and ingest whole books of context without getting confused.
I completely agree with everything you’ve said, in the pre-AI world. Now, I’m not so sure.
When I worked at a big tech company on a large codebase (tens of millions of LOC across dozens of repos), it was extremely common to work on something in an area of the codebase I had zero familiarity with, and due to turnover, no one else at the company did either. As you say, docs were frequently missing or out of date.
However, with some effort over a couple hours, I could make a LOT of progress in understanding the history of the code. Every commit and PR were linked to JIRA tickets, most had eng design docs with comments, slack discussions, etc. I could step through the git history and watch the code change, alongside the artifacts of the human discussion and decisions that led to the changes. It wasn’t perfect, but I could make tremendous progress. Now, this was a remote company with pretty strong culture around using JIRA, design docs with review, etc. Probably the biggest gap was meeting transcripts.
Today, an agent can chew through years of history and artifacts on a large codebase in a half hour, documenting as it goes, and have way better understanding of it than I ever will.
It’s true that an agent can’t hold all of that in its head at once without context rot (though this is improving every year), but neither can any human!
If your goal is understanding how to do something, why something was done, why something wasn’t done, etc, and the codebase is large, mature, and extremely well-“artifacted”, I’m not at all convinced that you’re better off asking Bob who has worked on this section of the codebase for a decade than just asking a really good frontier-level agent. Maybe, but it feels like that won’t be true much longer.
> Today, an agent can chew through years of history and artifacts on a large codebase in a half hour, documenting as it goes, and have way better understanding of it than I ever will.
At my last job we worked in a large, but publicly available code base. My experience was that just giving an agent the prompt to look at area X to figure out how it works would use up half or more of the context window. And that was thrown away every time we started again. We had docs, agent generated overviews but the agents still filled their context windows way too quickly to actually really be any use.
I have a hard time finding concrete examples that does not breach contract of dox me.
I feel like this issue goes beyond a single large legacy codebase. It can go beyond how the code works into what the code does. How should a thing behave may itself be institutional knowledge only Bary that was involved in the design process still remembers.
Pointing an agent at a problem in one codebase to write a patch can work quite well. But what do i do when the problem is either in the tool, the database, or the two API layers between? You may not even have the code for half of it as it's part of another department or externally purchased solution.
Bary would know, or know who to talk to, or dig up a PowerPoint slide he's been working on to explain it.
Exactly. As someone much smarter than me once argued here on HN, the value of a software company is less in the software but much more in the mental map the team has built of the problem and of how the software solves it. Throw the team / mental map away and you'll have a very hard time maintaining the software.
At the same time if the organisation want code fast, it gets code fast.
The revenue for a new feature today is something sure. While the cost associated with supporting such feature will be up to debate in the coming quarters.
As often it is the case, we are moving on a long vs short term trade-off space. And I don't think any experience will generalize
I'm not saying we can get rid of humans (yet). I'm saying that it's not critical to write and review the code yourself anymore, and that the LLM is the new layer of abstraction through which code is written.
The spec will never be able to capture all the edge cases. The code is what runs. It also doesn't capture all the edge cases (that's why code has bugs) but it does a better job at that than documentation. Documentation can also become stale easily
I mostly agree with you, but for “Documentation can also become stale easily” … historically I’d 100% agree with you—I often forgot to update separate documentation files, forgot to update comments on related functions, etc—but AI is so much better at automatically catching and fixing this without prompting than I ever was.
That's true, but it's also prone to over documenting. As an example, for some reason it always tries to include a diagram of the folder structure in the project. I think that's not really useful and it's something that changes so often it's not worth documenting
This, and maintaining the spec is a cat-and-mouse game precisely to cover those edge cases.
Like "if you're using this database, and the engine has these configurations, then do _x_, unless _x'_ and _y'_ are enabled, in which case, do _y_..."
Which, if you're already being THAT specific in your spec, you might as well, idk, write the code yourself?
Because at that point your human language is basically the code and AI is the compiler. A non-deterministic one.
Regardless, your business stakeholders won't understand what's going on anyway (nor should they), so we're back at square zero.
(I wrote Technical and Functional Requirements Documents as a business analyst in college. What's happening now _for most situations_ is more or less the same thing.)
AI is not an abstraction.
I beg to differ! Everything is an abstraction.
I think the main problem with this statement is that different facets of an entity are abstracted at different rates.
AI abstracts effort and cognitive load away from code at a heafty rate, but it doesn’t abstract liability away from code at all.
My business is paid to produce artefacts for which it has liability in the case of error, so we need to do additional work to mitigate and eliminate the liability risk introduced with language models. So far I’ve not found a better way to do that than a plan/act/assert type approach on every feature.
What a very useful definition.
You’re being too abstract ;)
I care!
What happens when the maintainers lose access to frontier models because of cost, politics or other external factors? What happens if they don't have enough hardware to spin up an open-weights model?
I've seen variants of this play out before AI, so I can tell you: they will inherit a codebase they've never seen before, take forever to ship fixes (forget new features), and they'll either scrap it, completely rewrite it, or, if they're "enterprise" enough, will pay consulting companies literal mountains of cash to make it their problem.
The whole point of writing simple code was to write code that other humans can maintain. If AI is here to stay and becomes economical enough for everyone to use it, then you're right; writing code for other humans is no longer useful. If that doesn't happen though...developers who can/want to still code by hand will be loving life
Then all the more reason to support open weights and the improvements going into them, currently state-sponsored but available to all.
If frontier AI is load-bearing, be thankful you're able to spin up comparable models (face it — 99% of us aren't writing novel software) for a fraction of the cost it would be for a third-party to come in and mop up the mess.
> hundreds of MB to GB of code from god knows where doing god knows what
If you use established libraries then actually the code IS well known to someone (and likely many), even if that's not you. Likely it was built with an actual purpose and with the foresight to not add red herrings to the design.
You can't say any of that for the equivalent amount generated lines. Literally no one knows what it does.
moreover there is not a black and white "code you can read" and "code you cannot read", there is a spectrum between. In this era, IMO if you know which module do what kind of function it should be enough, you don't need to have deep understanding of the file.
When slop code needs to do something different or new or in a different environment but still stay working on the original things, it’s hard to change, sometimes catastrophically so. When your code has concequences, quality means more people understand it which means more people can maintain it.
Arguments against it are sort of like why have devs on staff at all when you build the thing the first time, or why not outsource everything, or why should I care what my code looks like when the code seems to work?
The cost of tokens is not zero, and the bigger the thing you’re doing the more low quality will cost. When your company is the software, you take on an existential risk based on the software working or not.
Pure AI generated code without human curation is full of bad wordy comments but those comments can mislead, be stale, contain duplicates, and drastically reduce the ability to understand things. I watch teams that still care about code understanding ship good products while those vibe coding in the same org just flounder after the initial burst of features. Some problems only come up after the first 90%, and AI can help you solve them but a big ball of spaghetti is still a big ball of spaghetti.
Magecart has been very $ucce$$ful at hiding in JS dependencies.
Seems to me that if the AI writes the code, then AI can easily copy the code.
I.e. how hard is it to point an AI at a piece of software and say "AI, copy this"?
Seems like sooner or later copying just becomes a matter of spending enough on tokens.
Seems in that world, all significant software projects get copied. That turns software into a commodity loss leader for other business models or an open source project. Similar to the way Chrome works for Google and the way Firefox works.
It's harder than you might think. Mature software is battle-tested. All of the bugs and edge cases are handled. You can't just copy that from the UI.
Long term, that also just seems like a matter of spending enough tokens. I.e. we have these "red team" AIs that are finding security bugs in very mature software. As AI improves it seems you will be able to just point them at your copy, spend enough tokens and get most of the bugs.
This has been the experience of everyone making decisions in any company without being the one doing the technical work. It's not a novel concept. It's actually the opposite, compared to technical people running companies.
Well they had actual people who understood the code - because they wrote it and had a mental model pretty deep. Now the people owning the code don't understand it.
Won't you be worried if your mechanic didn't understand your car but offloaded it to a robot that made mistakes all the time?
In many places it's demanded by upper management that devs use AI.. so even if engineers wanted to avoid using it, they would have to meet their quotas.
This is what happens when executives suffer from AI psychosis. They were already impatient, now with AI all they care about is feature velocity.
The faster they can hit that refresh button to see the features, the quicker sales can close the deals for them.
AI has basically sold them to wet dream.
I am reminded of the cartoons where the car kept going faster and faster, the driver kept pushing on the pedal, parts started to fly out, the gauges started to go in circles, smoke started to billow out of everywhere and then boom!
I guess we just wait for the boom.
Do you understand every part of your dependencies now, pre-AI?
I see it kind of like baking/cooking. Do you bake your bread from scratch? Do you grow your own wheat and grist your own flour?
I think over reliance on it or not even trying to understand what is happening is a big problem to be sure, but it's certainly not a new problem.
I do try to learn how systems work and remove "abstractions" as much as I can. I do that with coding and my personal life.
Actually baking is a good example, I used to be really bad so I spent time learning. I don't do it every day but now I understand how bread is made. I bought a 3D printer so I could print parts to fix stuff myself. I learned to do my own oil changes, I learned how engines work, etc.
My point is that I try to learn more, not less, which is what AI is trying to achieve
The person who grows the wheat understands how to grow the wheat. The person who grist the flour understands how to grist the flour.
And I would understand how to bake bread from scratch.
Now in the entire chain, we will get to a point where no one knows anything.
I feel the same way but I have successfully refactored some of the early experiments. Our team has settled on targeting a double output from the before times but more ambitious product vision because AI can teach us things we don't know. We actually target 2 days of coding and 3 days of learning with Ai so the increased efficiency allows upskilling rather than just pushing more code
> The best devs still put out good code, but the poor devs are learning nothing while their metrics look great. I can't help but think we are racking up immense amounts of debt that will very soon become due.
I agree with this, but the reality is that it's only the result of models empowering devs, and power in good hands amplifies positive results while power in mediocre hands amplifies technical debt.
It's a good time to choose wisely who you work with.
> It's a good time to choose wisely who you work with.
Very true, but this also makes me think what kind of ridiculous obstacle course future hiring process would look like.
In a land where anyone with a pulse can prompt AI to make an app for them - how would future hiring managers and team leads figure out who will drag codebase down with tech debt and who wouldn't?
Probably the same way they have for the last 30 years: poorly, in a dozen different ways, depending on what that particular hiring manager thinks has correlated with successful hiring in the past
I can imagine a code review where you're asked to implement and merge 10 PR in a sample codebase and the codebase is littered with the sorts of mistakes and slop that vibe coders put in when they're careless and you're asked to correct the mess and make it work.
Future hiring will select for people that shipped the most, plain and simple.
The tech debt concerns are much ado about nothing. Use the next model to clean it up, big deal. Code is cheap.
The people that sat around handwringing about tech debt and trying to read every line of LLM code will really struggle to find a job. The profession fundamentally changed, and these people did not catch up.
Sounds like a great way to end up with people that run 73 agents to create bunch of useless AI harnesses that nobody (including the repo owner) uses.
I think there has to be a better rubric by which person would be evaluated, not just amount of tokens they managed to waste.
Of course you shouldn't evaluate people by tokens, you should evaluate them by what they ship. Did it matter to the business? Did it ship fast? Did they build what was needed?
> The people that sat around handwringing about tech debt and trying to read every line of LLM code will really struggle to find a job. The profession fundamentally changed, and these people did not catch up.
I'm inclined to agree - the latest models seem to have crossed a threshold where trying to review by hand becomes akin to a manager insisting on reviewing PRs.
Just like rushing made messes in the before times so too does rushing via LLM. The exact same outcome will happen, but at a far greater velocity and scale than anything we've seen in this industry before. Old Testament, Mr. Mayor, real wrath-of-God type stuff!
It's truly bizarre to read about all these people who just give up on any understanding about what they are working on.
I very regularly use plan mode not to even make a plan of action itself, but to better understand what possible issues might come up when implementing some feature or fixing some bug. And it is quite common for me to fix or rewrite certain findings that AI comes up because its assumptions are not quite right or don't align with overall goal.
And yet so many seem to be perfectly fine leaving all the decisions to AI - even if it's going in the wrong direction. I suppose that's all the people who got into software purely for money or status - never really caring about the actual thing they are working on.
They believe they know, because they have LLMs to tell them, but they don’t often get to the point of being able to have a conversation about it
The last couple of weeks, I’ve got rigorous about making the agent refactor old code. We’ve removed, moved, consolidated, reordered loads of cruft. It has made the code much tidier and reduced the chance that the next feature will build on top of cruft.
E.G. over time we’d gained two client-side caches of related server state. This started out as two different parts of the same model, because we couldn’t get all the data we needed from one microservice and had to merge in the client. Over time, more and more features used both caches for different aspects of related processes. At some point one of the microservices changed so as to return all the data in one call. The update to consume that kept both caches, adding code to sync them, because so many parts of the code were using one as a fallback for the other, so they both looked “necessary”. Because they were separate, and “live” sometimes they’d go out of sync after the initial load. Worse: the consumers alternated about which cache was treated as the fallback, making it very hard to see that either might be redundant. Eventually I noticed they were filled by the response to a single call. We all know paying back tech debt never gets prioritised, so I rolled the payback into two feature tasks, and just took longer about them.
My employer expects we use LLMs and provides some budget, but it’s not enough to use even Open4.7 or GLM-5.2 on every task. I do the bulk of my work with Composer 2.5. It’s quite good for “going forward” on smallish tasks and it’s written most of my code this year. It’s possible smarter models would spot these refactorinh opportunities and action them proir to building features or fixing bugs. But I wouldn’t know because I can’t afford it. I’ve never seen even a 4.8 era model spot a refactor and plan to do it prior to a “new build”.
I’m pleased I’ve spotted these trends and started to build the habit of (telling the agent to)“refactor to make the change easier”, but my percieved productivity will go down and I risk the ire of my leaders.
FWIW, our codebase is growing, and the size of each change is also growing, but it's because AI is making us fix all the bugs we'd previously check in because our code long ago surpassed what even our best developers can reason about.
The funny thing is that the AI adopters are in the middle of the bell curve. Our worst devs continue to perform worse than AI yet refuse to use it and our best devs continue to insist AI sucks despite it finding issues in their code and the reviews and designs they've approved.
Too many comments about AI read like inadvertent admissions of incompetence or dysfunctional projects.
Couple of WTFs that come to mind:
* How many bugs did one have per commit, that commits have to noticeably grow in order to not have those bugs in the first place?
* How does one even do software engineering if the (best) developers can’t reason about the code?
Code debt.
Software engineering is possible but largely a myth in practice.
Jesus man take some responsibility.
Lol sure dude.
Shouldn’t you be, in general, separating out new functionality from old fixes?
Shouldn’t those “fixing bugs we gained in the past” be their own MR that can be read, reasoned about and have evaluated test coverage?
Sure and sometimes that is what happens. Sometimes it's not. We should be doing a lot of things, but have to triage issues and act pragmatically.
In the case of what I'm currently working on, filing bugs for every issue I found, and then factoring out each fix, and then running each change through the 8 hour ci/cd system, and hoping an unrelated issue doesn't get misattributed to me... No, I'd rather just wrap it up into one coherent refactoring change and be done with it because when I'm done there are several more like it waiting for my attention.
I don't know your exact situation; and there are sometimes genuinely valid reasons for an 8-hour CI/CD system; but ... man, reducing that down to like 20 minutes (which is possible and often common) would pay as many dividends as all the AI stuff that's been added. Man, like ... holy crap @_@
We're uh.. we're a well known and widely used piece of software. There is an insanely large amount of testing that needs to go in to each change across many platforms and configurations and we have a lot of devs hammering that system with changes to test. Even worse now because of AI...
If your company is really interested in doing AI stuff you may be able to sell people on the idea of a "fast agentic CI" where you:
1. Track code coverage from all current test cases in your big 8hr runs.
2. Index the coverage so it goes testcase name -> methods touched.
3. On every commit run an agent which looks at the diffs, tries to predict which tests the code being edited is touching, and run just those tests (or a random subset if there are too many).
Once you have this you can also automatically kick off builds every 8hr that contain all the previously submitted patches. Once that is done, if a test begins to fail, it can automatically bisect the history to find change, and notify the author.
You could also get these benefits with a Bazel like build system which can cache test executions so others don't need to rerun them if the binary is not changed.
I see this complaint frequently about losing track of what the agents are doing, and I agree you do need to understand your system. But there seems to be this baked in assumption that if you lose track, you now need to manually wade through this massive mess to untangle it and maybe that is impossible. I don't agree.
If you don't understand the codebase, ask the agent to explain it to you. I'm not not kidding. Modern frontier models are fantastic as this - even more so than actually writing the code. It can tell you in words. It can generate architectural diagrams and sequence diagrams. It can write tests and scripts that prove it's assumptions. It can happily refactor so that the system design is aligned with your preferences.
Once you accept this, you can stop worrying so much about it and instead focusing on building the architectures and tools that lets the agents succeed better and faster - so called closed loops or agents prompting agents. Build systems that are more easily verifiable and deterministic so the agent can write very powerful property based tests. Focus more on what and why you are building, how to make sure all external properties are verifiable and leave the internals to the agents. The code is not really for us anymore.
In my experience it's really good at making you feel that you understand things.
Then when you actually dig into the code, there are many things that are not like you'd expect.
When you've experienced that a few times, you stop trusting that the agent gives you the full picture - for good reason.
When I review AI generated code I generally find so many flaws that it makes it hard for me to believe that those who are not reviewing their output are not just fooling themselves. Maybe not all the time, but quite often.
One such recent example was an SSO simulator for a local env. Instead of using a cookie to remember who was logged in, the agent remembered the last log in a variable, assuming the the next requests would come from that login.
This snowballed into our tests, where later agents had created helper tools for working around the SSO simulators statefulness.
Never trust a single session.
Things improve drastically however if you spin up a second session and ask it to adversarially review everything that the first session produces (this goes for everything: not just code, but also design, planning, and explanations).
This works even better if you use models from different families to do so.
Problem is where to stop. Open a 3rd session? Are you sure the 4th iteration is mostly correct? Let’s try a fifth now…
I have, on a lark, reached 20 to 30 adversarial sessions a few times. For some tasks, the sessions will just never converge on anything that yet another session won't find fault with, recommending an alternative already ruled out by another session. Even if all the reasoning in between was documented, the new sessions will endlessly claim to find flaws in past reasoning.
My experience is this is great when the model surfaces something to you. But I'm constantly caught off-guard by things the model didn't volunteer, things I would have quickly stumbled upon if I was working on code the traditional way. The model didn't think it was relevant but I sure do.
In my experience having the agent explain the code doesn’t work very well for real world apps, even the parts written by humans. For example we tried using to generate diagrams, class hierarchies, etc as part of documentation. If you don’t know the code it looks great. If you do, it’s focusing on all the wrong things, missing the mental model, and ignoring lots of important bits. And Claude tends to be extremely verbose to the point of muddling things.
[flagged]
"The one thing plan mode helped is for the humans to get an understanding of the strategy"
For me, this phase still happens, but a distinct "plan mode" is unnecessary: I just tell the model, "This is discussion; no code changes yet." and spend hours figuring out what will and will not be done.
"This is discussion; no code changes yet."
vs
Press "Tab"
Pretty embarassing state of affairs for this "engineering" discipline
its objectively not engineering
people actively deskilling themselves and becoming totally dependent on an ai pay to play service.
To be fair, it has always been like this. People actively became dependent on the Internet, on StackOverflow, fast Internet connection, cloud services, etc.
At a company I had setup scripts to build our packages, and the CI was running those scripts. Someone more junior (only a few years, not decades) found it strictly superior to remove my scripts and replace them with GitHub actions: people could now know even less about it (as in, no need to know how to copy-paste and adapt the recipe for a new package), but now it depended on GitHub. GitHub is down, nobody can build anything anymore. And it happened once every few weeks, so people would just go have a coffee during the outage.
You know what happened next? That person got promoted for their good work. That was before AI.
yeah before all you needed was internet and an editor. Now to carry on it's $200 min
> "so complex there are no obvious deficiencies" (as opposed to "so simple there are no obvious deficiencies"
The quote is "so simple that there are obviously no deficiencies"
The quote is so simple that there are obviously no deficiencies in it
> I can't help but think we are racking up immense amounts of debt that will very soon become due.
Future AI agents will refactor tech debt at lower interest rates.
Unfortunately what's happening is people just can't help themselves. Just like an addict reaching for another hit, it's really difficult to make yourself do work when you could just not. I've said it before but LLMs are our cigarettes. It's going to get a lot worse before it gets better.
These comments are so bizarre when we are what, 1-2 years away from AGI?
Like - you really think models won't be able to clean up the tech debt they created!? They are very good at this already. Ask Opus 5.5 to clean up the tech debt from some Opus 4.6 vibe coded app.
Code is cheap now. The most important thing is to ship, ship, ship. If you are handwringing over "tech debt" you have already lost - and you deeply misunderstand how good this technology is getting!
We get it, the future can't come fast enough for you. Code is a commodity and the only important thing is scale. Congratulations on leapfrogging the midwits on the path to the C-suite!
Did I say anything about tech debt? What's the point of "ship, ship, ship" if AI will be able to do it in a couple of years? Don't you see where this is going? We're rapidly losing our ability to think.
The point of "ship, ship, ship" is to cash in (in terms of your career or company) before AGI renders you jobless.
If we get to post-scarcity you never needed the money. But if we end up in some dystopian hellhole where people have no jobs but capitalism still exists, you'll be thankful.
Correct code is not cheap. And some places you need a very correct code. Like in mission critical systems.
The fundamentals of writing correct code have become dramatically cheaper by orders of magnitude. I have been involved in hiring dozens of developers. Fable/Opus 5.5 would be a superior drop-in replacement for perhaps 75% of those hires at roughly 1/100 the cost.
Perhaps for writing boilerplate, or generic websites from scratch, or early stage prototypes.
I work in embedded systems where it's harder to offload the burden of testing and maintenance. Systems knowledge is important: experienced developers can be costly to replace.
LLMs make a lot of things easier, but costs have certainly not fallen by 2 orders of magnitude.
How many people on HN work on mission critical systems? The discussion is not currently constrained to that. I agree, but the exception proves the rule.
Define "mission critical" and then maybe we can come up with a reasonable estimate.
Defined as "people die if your code has a mistake".
And even in these systems AI will be better than humans at finding issues.
2026: you're going to forget how to code and spend your entire career writing documentation, and you're gonna love it
2027: Stage 6 - a confusion.md so thick you forget forgetting (ever having written documentation)
And what happens when it becomes due?
Nothing, the people worrying about AI tech debt on HN have been worrying about it for years. 2024 HN threads had the same comments.
The reality is that models just keep getting better and are very good at cleaning up the debt they created. The "tech debt" bill never came due. It won't.
I’m not worried about tech debt, that’ll be my career making money from the laziness of vibe devs.
> The "tech debt" bill never came due.
Companies paying $200k a day for coding models to churn on what the coding models are messing up is one thing.
I’m not even worrying about tech debt, I’m talking full on defects, production incidents, security holes, and reputational damage.
Do you think the average dev is better than frontier models when it comes to spotting defects and security holes?
Until some catastrophic data loss event or data leak that nobody understands or has any mitigations for. Oops! Guess bankruptcy and prison time are the ultimate uncaught exceptions.
It's not 2024. Models today are far better than humans at spotting the bugs that would cause this.
I remember this being said as far back as 2022 yet we're still here fixing things manually when the models inevitably code themselves into a hairball convoluted enough they can't find a way out.
No one at a FAANG is fixing things manually these days.
If it turns into a hairball, just have a few agents rewrite it.
Do you have any source to back this claim? I haven't heard of any Facebook, Instagram, Google or macOS rewrites being fully done with agents.
We cannot blame just LLM models, it is brains nature to save energy. If agent did tasks consistently good, our brain try to delegate cognitive load to the model to save energy. After consistent use of LLMs anyone can have tendency commit slop just reviewing at high level, this is specially true with busy lifestyle. Also AI generated code do not give dopamine just like solving problems we did before LLMs, we tend to get lazy. Strict discipline is necessary to make good use of LLMs in order to not commit slop and not to make us dumb.
[dead]
When I draft my idea for the implementation of a feature or bug fix, I don't even trust a _human_ to understand what I mean the first time. There are _always_ either errors on my part, or erroneous assumptions on theirs. Everything from "this accounts for X and Y, but not Z which breaks the whole thing" to "this part of the idea directly contradicts what with you said earlier, what do you want to do about it?"
I can't bring myself to trust that an LLM understands what I mean better than any human would, no matter how "good" people claim they are getting.
TFA seems to be advocating for regular old vibecoding. Code now and ask questions later. Which is their choice, and is perhaps even a valid choice in many cases. But at least call it what it is.
Plan mode, for me, is my opportunity to develop and understand my own plan. Sometimes, rarely, Claude demonstrates it misunderstood my intentions by writing a plan to address the wrong problem. But usually it's about me fleshing out the scope and boundary of the intended changes.
I find that faster than code first, ask questions later. But it takes more time up front.
In the before times, and sometimes now even, the way I always coded was by thinking/discussing the work, then hacking up various interesting bits, then throwing it away and doing it the "right" way. Plan mode makes it harder to go back and forth between "hacking" and "thinking" phases.
Coding with plan mode is still vibe coding. No-one outside HN calls it things like "agentic engineering" or whatever. Everyone just says they are vibe coding regardless of how hands on they are.
Whether you use plan mode or not is orthogonal to whether or not you're vibe coding. "Vibe coding" means that you don't read or edit the code; you are running the software and iterating on vibes alone. That's the meaning since the original (and very recent!) coinage and hasn't changed.
If you're reading and editing the code, you're not vibe coding. If you're not reading and editing the code, you're probably vibe coding even if you feel "hands on".
Doesn't matter and it has changed. To most devs vibe coding is when you let AI write your code.
I'd say maybe a different term would be better.
The real reason why plan mode is dead is because you can just conversationally instruct the agent to not make changes to the repository or to make changes to selected documents only, and it will listen. There was a time when we needed to enforce this via selected tool use, but we have surpassed that.
I agree. Plan mode came about because earlier models were loose cannons, doing what they pleased. Today's models follow instructions better enough to not need a separate mode. However, there is still value in using a separate, smarter model for planning than execution, and persisting it for auditing on completion.
If you do use plan mode you might like https://plannotator.ai/
I typically converse with the default model to point the plan in the right direction, then have it iterate with a smarter reviewer to find flaws until the plan file is converged.
It seems unwise not to implement sandbox measures just because the chance of misuse has gotten lower.
Plan was never a sandbox or a permissions system
Plan mode does have different permission controls though? Its difficult for me to parse the docs on this but it does seem to imply more than just a prompt although less than a full sandbox. https://code.claude.com/docs/en/permission-modes#analyze-bef...
The way it worked originally was it would trigger a different system prompt, and then the permissions system would prompt you (even on YOLO mode) before running any non-readonly actions.
Cursor would also write a plan document, which was useful when working on larger tasks (due to context size). I still find that part useful today.
early plan mode would just refuse to make edits at all
Plan forced the model to write down and outline what it was going to do before writing the code. As Boris said up thread, he created it to prevent it from just rushing into writing code. So that regard, it did act like a permission system.
It helps to think more abstractly when approaching problems like this.
Yes, but plan mode wasn’t that.
You can use Docker’s sbx or similar VM/containers for that.
What? Nothing to do with that
It also doesn't allow leaving an audit trail of plans and decisions (by default, anyway). Most of my mutating prompts look like "Propose a plan for change X and write to file Y" and "Execute steps M-N from file Y".
RE the article: I don't think it's obvious why this process is worth following until you find your time and attention wasted. Conversationally-building is the express train to waste. I'm not sure why you would even be talking to claude if you don't understand what you want to build.
Do you really trust it with a production code base or database? Telling it not to make changes feels an awful lot like "Make no mistakes". I also like how Plan mode on Codex asks clarifying questions. I'm in management now so I've used Work more than Code lately.
true dat
Anecdotally, in the Opus 4.6 days, it felt like there was something special about using plan mode to discover the approach then clearing the context to execute on it.
A mixture of defending against a disastrous mid-implementation compaction (where suddenly things would veer off the rails) and also allowing the fresh execution to double-check the assumptions and notice any subtle mistakes before context was poisoned.
I’ve found that for large enough changes I still prefer having a parent theorizing about the root cause of issues based on evidence and then dispatching targeted child sessions to fixed based on theories and concrete telemetry examples.
There’s something clean about having sandboxed context and a session you can quiz about architecture while one is heads-down working against a spec.
For me, it was the agent that decided about plan mode, when that happened. Most of the time I didn't invoke it but kept discussing and tweaking and if the context was close to full, make it dump it out to disk and start a new session. I felt like that worked more flexibly for me because I could keep steering and adjusting until it was what I wanted. At that point I made it write a detailed spec. Like my homegrown plan mode.
The "standard" plan mode felt too stifling.
For me he plan mode is as useful as ever, even though I understand this is just a prompt. For large changes there are few problems exist such as context window limit and compaction, and even though models got better these are still detrimental to models performance. I found making one large high level plan and then splitting it into N plans grouped around distinct areas and then running N implementations if possible in parallel, or some consecutively, improves model "vision" far ahead. The workflow author suggests might be more useful for prototyping, but there are still types of work that are better done with massive pre-planning stages. And of course running exploratory prompt a.k.a. plan is useful even for smaller types of changes like small features or bugfixes.
I found the article interesting in that the author didn't baulk at receiving an unchecked uncheck-able bulk of code. If this is what is supposed to ship, am I just old-fashioned to think this should not be acceptable? There is no way I would ship code that a person had not read, let alone written.
A quick straw poll. Are most people here who use AI to code well-versed in their languages/software development? i.e. 10+ years experience doing it "by hand"? I think in ten years time there will be no developers with that 10 years experience behind them.
I did everything by hand for so long and miss it, now I just read thousands of lines day in day out. But there are interesting gains such as lowered barrier to quickly run adhoc perf tests, do some formal verification, you can just add more guardrails for the thing that gets shipped. The thing I hate though is when managers sell their generated code and you are left shoehorning it to production, they take away the interesting part of the job.
Completely agree with your first paragraph and a point I've made repeatedly here is that I want agents to work more like pair programming and less like code review. I want to delegate typing, not thinking.
But what I've noticed is that lots of people don't want that. They're happy to delegate both typing and thinking.
There are folks who still want to think for themselves of course, but model providers are incentivised to maximise token spend over all else, so the "official" harnesses are not aligned with their needs. My message to those folks is: _go and build your own harnesses_.
I built one, it wasn't a huge lift and it completely changed how I work with LLMs. One consequence is that I choose smaller, cheaper models now because latency is more important with shorter feedback loops. I don't want a super-intelligent model that disappears for ages and provides a final answer, my own brain thinks and guides the process instead.
I'm deliberately not linking to my harness here because that's not the point. Think about what modes of work you want from a coding harness, then build one that works exactly like that. Immense happiness and satisfaction is just a couple of weekends away.
Are you following more of an “aider” type approach, or something different?
I think this is the clearest example of: exceptions don't make the rules.
Author might be a vibe coder, designer or someone who does not do really heavy and complicated development, maybe some api here and there.
But it is pretty obvious that a very concerned engineer would not allow any of that, for critical systems.
I have two years of copy pasta from Reddit pre-AI. Does that count?
I'm probably just behind the curve, but I still use plan mode quite a bit in Claude Code. I iterate on the plan repeatedly until it largely agrees with what I want. Once it seems good I ask it to build the plan and open a PR when done.
My only critic of the plan mode is I wish it was easier to see the updates and changes easily in Claude Code as we iterate on the plan. It is wasteful to have to remember what parts I have reviewed and what parts are new (and need another pass). I have thought about fixing this but I also feel the review is the actual thinking (even if ineficient), and so I purposely have not removed it.
+1 on your critic. I have tried to solve it by making plans folder a git folder, and using tools like Hunk and Tuicr but it’s awkward…
For an adhd mind that serial interface is a bless and hell at the same time. Great for the detailed planning discussion, bad for checking the results.
I am currently working on canvas-based interfaces for that reason and i would think the only way to really create value here is with a deep independent analysis and visualization of the changes afterwards to reach some ease of mind. Live would be cool (if you like that)
When it comes to planning itself, I recently tried the token-saving planning plus phases execution agents approach and had to find out that agents actually don't necessarily communicate better by prose-reduced specs than we do.
I had to go back to the planning agent to implement or fix things with our full planning context in mind. So if you want really high control for a "tight" implementation, I'd say just sharing plans is not enough. The probability of things getting filled in by the executing agents rises and you either find yourself holding those agents' hands or fixing things afterwards.
Actually phases are still to large and you would actually want the planning agent to hold that hand all the time, meaning small context is not the way to go, as you might need the full checking context much more often than current phases sizes suggested by the planning "doc" would use it.
Distributed building still is the way to go though, steady control by the overall context or one specifically thinned out for the particular job is. But don't go prose-based plans anymore. These are dead indeed.
Disagree. While it's easier to 1-shot more things now, when there's more complicated interaction of components, a good plan from plan mode can make the project a "looks good, go" and run overnight thing vs. "steering" needed.
How many thousands of dollars are you burning a month on “looks good, go” and then *letting it run overnight*?? I spend enough money on 8 human, interruptible and human-thinking hours with paused hours.
I can’t imagine how high that number would go if I got rid of the collaborative style and also the “please don’t run scary commands on random directories without permission” mode.
Write a good high level plan with functional and non-functional requirements and it works out costing less, works more efficiently, and no surprises.
It’s like this: If you know the right words to say to the LLM, you’ll likely get back the “right words” also. And the more right words up front, the less steering you need to do.
(Having 20yrs of experience writing these user stories and designs help, my “unfair” advantage)
I still rely heavily on plan modes, first to think through product design from the 3-sentence descriptions handed to me as super-complex feature requests, and second to implement the technical requirements I want to build. Way before AI, I think the first idea, while sparked by inspiration, has always created a local minimum and suggested that something could be done around a problem, but it was rarely the best idea. I think the speed at which the models can turn a bad idea into code will lead to bad, design across systems and interfaces. I am also very pedantic about how the data models get implemented, and I want to be sure that things are built in the way I want them to. I'm also very pedantic about how specific if statements and dicts are created, and again, the models make it easy to do things a shitty way. These are hard-learned lessons of several decades of shooting myself in the foot and locking myself into less-than-ideal design patterns. I spent two days going through designing and planning for the LLM to build the actual feature in less than 2 hrs; once it's set, it's lost. It has made testing go very quickly and i am deploying a very complex feature and high risk feature in a 4 days when without the models it would have taken probably 4 weeks to get right.
Lately, I've been using Matt Pocock's “/grill-me” skill more and more, rather than plan mode or a complex, spec-driven set of skills. I made a personal fork of the skill to use Claude’s ask tool, which has a nicer UX.
I used plan mode for two reasons: to review the choices before execution, and to execute with another model (i.e., using the barely documented opusplan feature).
The grill-me skill is much better for reviewing and clarifying choices (and modifying it to use the ask tool makes you go faster). Instead of opusplan, you can explicitly tell Claude to start a subagent with another model to divide the tasks.
Unless I'm going to just do something, I use /wayfinder for almost everything now. It automatically goes into grill mode when appropriate, and it'll suggest just doing the work if it's light enough.
Try /grill-with-docs, too.
I can’t imagine working without a plan first.
Firstly I find it’s an excellent way for me to get very good clarity about what will be built and whether it’s going to be done in a sensible way.
Very often I don’t really know what the work will need to look like until I’ve explored the problem with the LLM towards first making the plan.
Without a plan I find myself having to do the initial understanding through code review of its generated code which is much harder than reviewing a plan, and then I invariably need the LLM to fix up what it did which is much slower when it’s doing code than working on a plan, never mind the next review I need to do.
And when the plan is good enough, I clear the context before telling it do it, which I’ve found vastly improves the quality of the LLM output.
All of those things are trivially possible without an explicit 'plan mode' built into the harness.
For me the most useful thing about plan mode in Claude or Codex is not the plan, I usually barely read it, it's that it forces models to gather as much context as possible before implementing. Without plan mode models tend to implement wrong solutions because they are reading not enough context before writing code.
Another useful property is the readonly nature. I can easily let multiple agents plan in parallel without having to worry about annoying worktrees or conflicts and then I can come back to the plans later.
Of course this can be done with just another prompt, but that's exactly what plan mode is. It's nothing more than a predefined prompt in the harnesses with maybe some extra guardrails (that don't always work)
Using an LLM for any large project shows how useful it is to have data and functions that aren't siloed. That's why CLI apps have had a resurgence: because the LLM can interface with them. Yet the reaction from so many who are deep into LLM development is to build their own little silo: an app.
We really need a better model. One alternative is to have an everything-app: a general purpose tool in which (almost) everything lives. The terminal is one of them. The text editor / word processor is another. (I use Emacs for everything.) In a business context the spreadsheet is probably the best choice.
Claude code’s plan mode began its slow march towards deprecation when they hid the “clear context and implement” option behind a flag. From the discussion at that time it seemed that the developers considered it mostly a legacy feature, and that newer models are smart enough in long context not to need it.
There's some pretty interesting evidence that this feature actually just increased costs: https://stencil.so/blog/prewalk
Wait, you can turn that back on? Where's the flag?
The flag is:
Boris rationale was "with 1m context window, most users don't need it anymore."https://x.com/bcherny/status/2035375125382451648
This part I find a little disappointing. The plan is the perfect boundary to cut context which still isn’t free, even if models can handle a lot of it now.
Remember, Anthropic knows best for your workflow!
/s
This is why open harnesses and open models will always be superior, it's only a matter of time until someone decides that something you use isn't worth maintaining anymore. The reply from the Anthropic employee up above doesn't give me much confidence that Claude users will be able to continue using plan mode forever.
Codex and Claude Code are the best harnesses out there by a long shot. They work out of the box, require no real config, and just do the thing.
The flexibility and lightweight nature of open harnesses is mainly useful for open models. These models are considerably far behind the frontier and need a lot more steering. Complex harnesses confuse them. I use open harnesses for my OSS model rig because the model needs it.
> Codex and Claude Code are the best harnesses out there by a long shot. They work out of the box, require no real config, and just do the thing.
Even if I agreed, that doesn't address my point: Claude Code might be the world's best harness right now, but that doesn't stop an Anthropic employee coming along tomorrow and taking a giant steamy dump on the thing you like most about it. You're using their workflow, not your own.
Sure, but their workflow is the best for Claude, and it's the same with Codex for OpenAI's models. These companies work and live with their own harnesses, they are ahead of the curve on what works for their models and what doesn't.
Plan mode is still very useful to me, I just need to read the plan. Even if the model is smarter than me, it’s still working on my project, and thus I need to keep a mental view of the system. My solution has been to ask the assistant for far less than it can do and leave for myself a bigger bunch of the work, so that I’m forced to keep myself up to date on the project. Yeah I’m leaving productivity on the table but… am I really?
I once tried to build something similar to your product on top of Codex, and I failed.
But my conclusion on plan mode is slightly different. I agree that plan mode itself may be a dead end, but I still believe there might be another way to achieve the same goal.
When I was building my product, I found that the biggest issue wasn't capability, but taste. The agent could build something that worked, but it often wasn't what I actually wanted. And behind that "taste" is a huge amount of implicit context — preferences, past decisions, product intuition, and trade-offs that live in my head. Distilling all of that into context takes a lot of effort, and I suspect giving it all to a single agent may eventually become overwhelming.
I've been wondering whether a better approach is to have multiple agents with different roles, prompts, and perspectives, and find a way for them to work together efficiently.
It's still just a hypothesis though. There are a lot of complicated coordination problems to figure out, and I don't have the answer yet.
For me, identifying a problem of "taste" doesn't exactly break it down. The problem is underspecified in some way. Can we be more specific?
The sense I get from this discussion is that the models/harnesses do not elicit feedback well. Where there is ambiguity, they tend to pick a solution and call it good.
A planning step aims to make these choices explicit. An iterative process is necessary to capture the detail.
It is natural to look to teams of agents to satisfy that process, but do they know where the decision points are?
[dead]
So planning is tied to the spec. This much is clear.
What the author I think is hinting at is not planning alone but "the development and evolution of any program and the state of this program throughout the planning, elaboration, and eventual runtime".
I chuckle at the thought that throwing an md file or a prompt at this problem is sufficient.
So, I posit that if we want any agentic code to evolve meaningfully in the short future and over the long run, we have to have a system which holds and presents this information, the state of a program, in a coherent manner to a human operator. No other way. No other way. And this I say to both nay and yay sayers.
You can argue also that this is part of an even bigger thing. But it is not part of the current discussion on planning and speccing in agentic systems.
OpenAI and anthropic can throw all the billions they don't have at this and adjacent issues, but if this is not solved then they don't have anything.
For work, I primarily use Gemini CLI, and the one real awkward part of plan mode is that the harness can't run commands, even exploratory ones. So when I'm giving the harness a task in plan mode, it can't experiment and is either blind or relying on its training data and/or any other context (e.g. source code, documentation) I give it. And it often gets things wrong operating that way (depending on how detailed the context ends up being).
For example, the ability for the harness to call into a Python one-liner just to experiment is pretty powerful. If I'm asking it to use D2 to build an SVG graph, it can write some Python to introspect the XML to see if things appear to be placed correctly (size, x-y coords, etc.). Which is a pretty cheap way for it to experiment and verify its results before I deign to examine the rendered SVG with my own eye balls.
I use Claude's Plan mode daily and it's great. I almost always have feedback to refine the plan and I want a clean separation of planning time before writing code. I don't get what the problem is.
I also find this "kill Plan mode" push on Twitter odd, because developers have been complaining about AI supposedly killing their jobs, yet they want to take away the main feature that lets them be an active collaborator and participator in the process. Weird.
I also don’t get what the problem is. Claude’s plan mode is great, especially when you add interviewing into the equation. It means I can get to a point where I’m confident all the changes proposed make sense, and that I understand them, before anything is actually changed.
I feel like this post is hawking for commercial reasons more than it’s based in reality, especially a reality that takes into account the extremely varied experiences engineers seem to have with agents. Not to mention it ignores capabilities that, for example, Claude’s VS Code addin have had for ages (specifically select and annotate in plan mode, as I mentioned in another comment).
As such I’m not inclined to take it very seriously. More than that, the “X is dead” trope was overdone 10 years ago and I don’t think it’s been long enough to warrant a resurgence.
I see what you’re saying. That’s actually why I built Nuanced (https://www.nuanced.dev), I wanted a better way for humans and agents to work through intent before the code got written.
but I think I overfit the interface to how I used to work without agents. when I was at GitHub, I wrote a lot of ADRs, RFCs, and design docs, and I liked that because writing is how I clarify my own thinking. with agents though, I’m often not doing the writing myself. I give the model a rough intent and it fills in a bunch of gaps, and then I get back a long, polished plan containing decisions I didn’t explicitly make.
That’s the part that feels broken to me. The plan can be detailed and technically correct, but still be hard to review because the important bits are buried and feel distant from my own thinking. All the assumptions and tradeoffs and questions may or may not be legitimate, but it’s hard for me to get into flow state and carefully check them.
So I still want the collaboration step. I’m just less convinced that generated prose à la plan mode is the right interface for it.
There are many ways people solve this. I use spec directories, and the agents just keeps updating it as I make decisions. It isn't a single file, it's a directory; sometimes nested based on the size of the project. I'll have the directory open in my editor of choice. It being a directory allows me to use an arsenal of text tools, and also optionally commit it for history/tracking.
This tool is a difficult sell. Users _might_ consider an Open Source version, but switching people away from familiar tools is not easy.
I likewise don't understand this push. Surely it's a trivial feature to keep enabled. It mostly seems like advertising that "our models are so smart you don't need to plan". But I find it to be a great mental review process for my workflow.
It sounds like the Ant perspective is "you can just ask claude to plan", but at the same time that's a little more tedious than shift+tab
I see what you mean, but I’m saying something else, I think plan modes were solving two different problems:
1. giving the agent sufficiently precise instructions 2. helping the human understand what’s being built
I think #1 is increasingly going away as models get better. #2 is a separate problem, and I actually think it matters more as agents get more capable. I’m not arguing that human understanding should disappear with plan mode.
I’m also not pushing this on behalf of a model lab, I don’t represent one. These are lessons I learned from my own mistakes building Nuanced (https://www.nuanced.dev). We built around a very explicit plan-oriented workflow because that’s how I used to work. I spent years at GitHub writing ADRs, RFCs, and design docs, and I still really like writing because it’s how I clarify my own thinking.
What changed for me is that AI-generated plans don’t give me that same effect. The agent reads between the lines, generates a lot of detailed prose, and now I’m parsing decisions and assumptions I didn’t actually make.
So I still think human understanding deserves a first-class primitive. I’m just increasingly unconvinced that a big blob of generated text is the right one, especially as we move toward many agents working in parallel.
>yet they want to take away the main feature that lets them be an active collaborator and participator in the process.
It's just redundant when you can just collaborate in the "main" mode. There's never been a point to having a separate plan mode.
There is a point, plan mode has harness level behaviour - it's restricted to readonly tools, and it re-reads its plan into context after compaction
> developers have been complaining about AI supposedly killing their jobs, yet they want to take away the main feature that lets them be an active collaborator and participator in the process.
Goomba fallacy. These are two distinct groups of people.
Yeah, honestly, I'm wondering, not even why people are against plan-mode, but how they're getting stuff done without it.
Maybe it's me (it usually is), but I don't give an LLM small tasks that a human could knock out in a half-day. I give them big tasks, stuff that would take a human a few months to a year - and then I ask it to give me a plan, as a living document, and we spend a good hour or two iterating over the plan.
Then, when its finally at the point where I'm happy with the plan, and I've talked it down from wherever it first wanted to go, or pointed out that we don't need to be all-things-to-all-(wo)men and focus is important, I get it to start going through it.
Likewise, I get it to maintain TODO.md with lists of known bugs, separate lists of future-features, and again, I make a rule that this file must be updated whenever something material changes. I just asked Claude "where are we ?" and got back stuff like:
The ## Still-open detail section lists two items:
- Task #1070: the ported back end doesn't fold offsets into vector loads, so it emits an extra add on 11 files (from 892) at -O3. The output is correct, just longer.
- Task #1080: array sizes must constant-fold. For example, u8 buf[EVSZ * MAXEV] is rejected because size expressions only accept a literal.
This is for 'xc' [1] - a compiler for an Objective-C-like language (but without the excessive []). The language has ARC, blocks and bound-functions/methods named 'block' and 'callback', automatic parsing of DWARF data so you can #use a shared-object, so there's no header files - just read enums/types/functions/methods from the shared object. It's a cross-compiler, runs on mac,windows,linux and creates executables for mac,windows,linux,ios,android,WASM (amongst others). I have a binary running on my iPhone which was written on, and signed on a Linux box - no Apple software used at all. Oh, and it produces code that is very comparable to clang in speed on both arm64 and x86_64.
You can appreciate it's a reasonably large project. It's taken actual months(!) [grin] for me to get working. Months! There's no way I'd approach a problem like this without detailed plans of what I wanted the language to do, where we were going with it
FWIW, "I" wrote blewit.net [2] entirely in xc - both the server back-end (#use <psql> was very useful for binding to Postgres) and the WASM client - which share classes between client and back end, to make it very difficult to get out-of-step between them. No Apache (#use <tls>), no scripting, just a lean-and-mean daemon talking to postgres via valkey (#use <valkey>) - a reddis-alike. Oh yeah, blewit.net has a plan too. Actually it has lots of planning :)
1: https://compile-xc.org/compiler/
2: https://blewit.net/
I have never used plan mode and always worked by planning. I just say 'don't write any code yet. I want to [describe usage]. Ask me the top ten relevant questions to doing this well. We're working up a planning .md we'll handoff to a fresh agent.'
This has never not worked as expected.
Alternatively you could just use plan mode
> While the ease of generating code this way triggered a greater dopamine reward
I get zero dopamine rush from this. I get a dopamine rush from building something or figuring something out myself. I don't get a dopamine rush for generating thousands of lines of code that I then have to try to understand.
Likewise. I get drained quick when I let it just spit out who knows what code.
I like writing spec files exclusively by hand and just asking the agent to surface questions about it, which I then clarify by editing the spec file further by hand. It keeps the spec file more manageable than having the agent generate the spec file from your conversation.
I disagree that more intelligent models make plan mode no longer useful. Even with superhuman AI, I would still like to iterate on a plan and understand all the important decisions before changing the code. AI cannot read minds, no matter how intelligent it is. Hell, often I don't even know exactly what I want before the planning phase is finished.
And yes, you can say that I can do this without an explicit plan mode.. But it's such a useful and common workflow that it deserves a special mode, IMHO.
Plan mode isn’t that useful but planning is, I never used it much, but in the agent build mode, I lay down the specs, the architecture, and everything I can think of, and I ask the agent to make the blueprint with the specs that will be used throughout the project, I review them, make the needed edits and revisions, then code and review follow. This is good because you can use the blueprint in any other model or agent, even humans can read it and understand it, sure, sometimes it gets overly verbose but it’s better than nothing.
With Claude, most tasks do not require discussion, I know exactly what to ask and what approach I want to take but for especially complex functionality with multiple possible options, I will ask it to list the tradeoffs and suggest an approach. It's still kind of 50/50 whether I take its suggestion or not; it's still a bit off in terms of weighing up importance, but it's really good at listing out relevant constraints and tradeoffs. It sometimes misses opportunities but it always sees the tradeoffs and the issues.
I still like plan mode in my harness because I restrict it's tool use so it can't go rogue and write documents outside of a specific folder. In my case it is /agile which contains a roadmap.md with epic1.md through is many epics as needed for the full vision. Then I often have one chat session per thread. If multiple agents take a sprint or a card it serves as a clear unified context for me and the agents to follow and then track what has been done.
I feel old. These days any new thing I try is dead by the time I actually start using it.
I use plan mode for anything that touches more than 2-3 files because I want to know what code will be added beforehand.
I guess people don't even look at code anymore
Yes, plan mode is still a reasonable way to guide agents _before_ any code is written. They're so eager toshamelessly change the files even after a simple quesiton.
But nothing is stopping us from prompting agent to only write to plan.md this session.
Fascinating piece, puts in words what I've been feeling for a while..
Over time I found some useful patterns (indeed after not getting what I want from plan mode), but this piece convinced me to double down on them and always try to find (and let the LLM construct) clarifying abstractions that have a deterministic relation to the code.
2 examples I recently build while I'm developing a large Django system with a complex datamodel and RBAC (spending quite some time to make it look good):
A script that makes an svg of my data model with all the classes/tables laid out and their relations encoded in the line ends (1 to 1, many to 1, many to many) and their on_delete relations encoded in colors. This also helps me discuss with stakeholders. The visual is also in the README.
For the RBAC model I decided it should be declarative so a TOML in the code that seeds the DB with the roles. I quickly landed on small script that translates the TOML to a markdown table with roles as columns and perms as row. It's also in the README.
After reading this piece I'm going to actively think how I can build these visualization more often and consciously, on different levels. Great realization.
I'm using a set of skills for planning now which does something similar to plan mode but it's creating documentation that later skills reference it while building. It works so much better than anything else I've tried, because it keeps things on track throughout the feature buildout and across different agents/models/tasks. From what I read here (& in general) my take is that a lot of people haven't adopted something like this and it's just totally the wild west right now as everyone is cooking up their own flows.
I agree that as I move away from holding agents' hands through actual coding I need a different way to monitor what's going on. What step of the plan are we on, what are the tests actually doing, what agent owns what, etc. I haven't found a product that does a great job of that yet and it seems like the next frontier of the 'IDE' to me. It'd be more like an IM(management)E really. The closest thing I've seen was whiteboard [0] but I didn't have a great experience trying it out.
[0] https://dev.fast
Maybe, I always say "let's write a plan and refine it" which ends up being a nice rubber duck session to keep claude and friends from going off the rails implementing things like privacy or encryption or some other rabbit hole when I just need it to do a few features at a time so I can review while it starts the next few things. Maybe this is slower to deliver, but I don't spend a lot of time unwinding commits.
I don't necessarily use plan mode explicitly for this, but I might turn it on if I think something destructive is coming up, like a schema change or some other devops thing where claude may decide "oh we copied it all over so I can delete it from source" "oh I'm sorry I didn't catch that error code and assumed success, your data is now gone" lol. This happens a lot less but I've been bit right on the face by it before.
Are plans the wrong abstraction? You can use a big enough plan to spam tickets into linear and have claude burn them down. I find it increases subagent accuracy.
I still find plans to be useful for big changes. I first ask to create a plan for the specific feature i have in mind. Then i ask for a second round of research with the current project in mind. Then i ask if it can make improvements on the plan. Then i read the plan and note issues down and let it correct them.
In my experience the problem with plans is that llms (at least gpt and claude) are really lazy and will stop at the first solution without even understanding your project. I had plans that wanted to add a few thousand lines of code just to support a very minimal part of a feature. After asking for a second round of research with the project specific limitations in mind it recommended a very small change to the existing codebase that did the same thing.
This is probably also the reason why most llm generated code stacks thousands of lines of code.
Plan mode - primarily for me was for Human. Models through chat could always be informed and forced to correct but real difficulty is having a good guesstimate of what model know and how its going to act. Relying on Model and its future improvement and declaring plan mode is no longer relevant - is too jumpy IMO. If you are running bunch of agents - then you need to confine them and also instruct them in BATCH fashion. Plans also useful for shaming model when it goes wayword and give it a baseline to correct. With AI, one needs a control instrument plans and spec are useful instruments.
Plan mode is "dead" because most plans are actually generated prompts to steer agents to keep doing the next thing, in hope they keep doing the right thing until the end. Real life is not so elegant and straightforward, you have a lot of "nah, screw that decision".
It is more useful to capture the motivation of the "plans" and their expected/observed end results, rather than the plan itself.
The disconnect between the prompt and actual motivation can be significantly high.
AG has dedicated plan mode now, idk why. Probably just rename into "guides" or "opinions" or "brainstorming".
I'm going through the same problem right now
Qwen 3.8 27b is the supervisor
Qwen 3.5 4b are the 6-15 minions it controls
Gemma 4 e4b is the validator for the supervisor.
A plan means it preps all work for the agents up front, tests that evals work, makes sure the dev environment is right for each agent, then finds and fixes each before the distributed tasks even begin.
What I thought would take minutes took hours as a supervisor or one agent did the prep / pre flight work.
My solution so far has been to drop all but basic setup and force the supervisor to ask before every op - if this is not the design choices, can this be run in parallel? If so, hand it off NOW.
I'm still iterating this workflow, but less setup for all the minions plus handing them work that may be incomplete/ broken is caught and fixed by the minion and its own qa gates.
This can mean a number of minions end up replicating the same fixes, but in general the time cost of that is small Vs the supervisor working in parallel instead of too sequentially.
These are self hosted for learning experience, I could have built an agent swarm in the cloud, but I'd never have learnt the fundamentals.
- Cold starts impact, context length issues, task lifecycle management
- Inefficiencies in delegation, necessitating workflow patterns for small projects (big AIs hide this problem until you scale and they hit the same issues).
- Limits of the AI would be harder to find or notice (e.g. where time - and cost - is being spent needlessly).
Or just use a more capable local model like Qwen3.8-Flash-Next?
Or Kimi K3? /s
Some people are resource constrained.
Tell the planner to not do all that. Just plan out the high level. No code. No implementation details.
Yeah, it's way better when you do design documentation, or even ticketing, to instruct it not to include any implementation specifics. You're not doing the deep dive on the zero shot that writes the ticket or the document and so it is much less informed than the agent doing the work will be.
I'm genuinely curious as to what you're working on that you find 4B models good enough. I wouldn't even let a 27B model code, never mind supervising smaller models.
At the moment I'm still doing shakedowns, so Typescript games compilation with a menu that has 4 games and retro artwork.
This seems to be a good example because things like the menu, high score boards etc are common, but the games are distinct. Then there's the artwork which requires decisions on look, and for coordinating.
The Qwen 4B model is multimodal so part of the AC is to view the output - I've a robust anti AI-look QA chain for that I've been using elsewhere, e.g. no floating parts, consistency, obvious missing fingers etc etc.
The longer term plan is to do some llama.cpp refactors specifically for some target hardware I have and implementing slightly different novel architectures I'd like to try (one I did already targeted CPU inference, which I did using 3 agents with specific roles; main planner, QA for planner, and benchmarking/environment handling)
The implementation was 85% of the speed of the original maxed out on my hardware but performance scaled with CPU core count whereas the original implementation plateaued. Unfortunately the break even mark seemed to be around 30 - non HT - threads.
I suppose I should look at that one again, since the increase in cores did not linearly drop off performance e.g. due to memory contention.
If you can run the smarter 3.8 27B model, why not use it for everything?
How do you set all of that up with supervisor and minions? What harness?
I still use plan mode. I use it for large features that I haven't fully thought out. Which I guess is like the author's "planning" phase. But I don't see the issue with using plan mode for planning. I welcome the agent to ask me for questions and clarifications. Do I read the full spec it produces at the end? Usually not. Usually the agent is smart enough to highlight the important tidbits and assumptions. I read those.
It's particularly important when you're making architecture changes or things other features will need to build on top of. I want to know what libs its going to use. The nitty gritty details I don't really care about.
At work, we use Claude Code with superpowers which organizes most work into planning followed by review followed by implementation and superpowers applies a fairly high level of scrutiny at every step. It's really quite effective, I don't generally read the code that comes out. In that workflow I have never used plan mode. I don't know what I'd use it for.
On my hobby projects I use Antigravity CLI and almost always start with /plan. This is a builtin skill - not a mode. It generates a plan artifact that I review and approve. Once approved the implementation speed is uncanny compared to Claude.
I tried using superpowers with Antigravity CLI but that slowed the agent down considerably without much benefit (plan quality was much worse).
I have been using this super simple skill[0] to make Claude ask exhaustive questions about the implementation of larger features. It has been absolutely amazing and felt like a step change in the output quality.
For complex features there can be 10 or more questions but I have a very strong sense of understanding the changes about to be made and Claude is very good at following all the decisions exactly. It's like working with an engineer who is both excellent at soliciting requirements and fast at implementation.
[0]https://github.com/mattpocock/skills/blob/main/docs/producti...
I realised I linked super cringe docs instead of the actual skill:
https://github.com/mattpocock/skills/blob/main/skills/produc...
The claudism in such docs is so tiring. So many words, yet reading them feels like trying to fill a sieve with water.
That writing style is borderline incomprehensible.
I think it is clear that planning before coding is generally a good idea - it's true for human developers and even more so for AI agents. The question now is more: what is the right way for an AI agent to form a plan and share it with the user? Maybe the dedicated plan mode in the agent harness is no longer needed, but I would like to make very sure that the AI agent has a plan upfront and that this plan is compliant with my requirements, my general expectations and that it is consistent and clear.
I can't speak for Claude, but in Codex, plan mode is very useful for finding serious issues with one's spec. I use it to improve my spec. I keep rerunning plan mode over my updated spec until it stops finding issues. It is only then that I execute the spec.
This is exactly what I do. If it's asking a lot of questions your spec isn't specified enough. It's very useful.
me too
Personally I really like grilling before a big feature, it's a Matt Pocock skill that will iterate and ask you a lot of questions so you REALLY know what you are building. It's particularly good for complex and big changes, so you actually know what the AI is gonna build before it does. Of course, that's not always the case, sometimes you don't care about the specifics. But I think an integrated "plan mode" is really not needed, maybe a "brainstorming" skill would be more useful.
Yes, a conversation can be the plan. It helps if you ask it what the commits will be and then tell it to use subagents to implement them.
Also, since sometimes an AI can be a busy beaver, I added instructions to not edit any checked-in files if the prompt contains a question.
I will try to update this repositorysoon, but I had to stop because the ahem compatibility failure of Claude Code env features; observer; agent workflow message passing LACKING on installed marketplace plugins
I am still actively working on thesis : a self-directed plan mode to generate artifacts that can go through multiple evaluations of interactive interrogation is valuable
https://github.com/samelie/claude-plugin-pnpm/blob/main/skil...
I've been using superpowers:brainstorming skill with claude for a while and have found the process to work well for me. I generally spend 10-30 minutes typing my requirements and ideas for a feature and then I feed it into brainstorming skill which asks questions to fill in the gaps before starting to build things. I don't bother reading the spec it creates. I have been consistently happy with the results.
However, since everything changes every 5-minutes now; I am curious what is now a better process than using superpowers? What works for you?
My only issue with superpowers is: it wast too much time/tokens - besides creating a lot of useless testing
I agree with that. Currently, I have claude's $100 plan and most days I do not run out of tokens. If I were building more; that would bother me too.
We experienced another step function in model quality over the summer. Models are now one-shotting medium sized tasks that require significant engineering judgement. We're moving into an era where the production of software will be done entirely by machine, requiring only an abstract description of the need the system is to fulfill in business language. This has been the goal since the 1970s. "Plan mode" isn't necessary when the model can suss out what you need from a few sentences of prompt and make competent engineering decisions to fulfill that need.
Brooks was wrong. We have a silver bullet now.
My own tool has a Project Starter, that takes users from idea to todo list in a number of very specific phases. I've been using it for small to medium sized projects for a year now and it still works great for me.
Just don't expect to end up with a finished project; it's more like a first draft. Once it's there, it's much easier to determine what it is you actually want, since you can directly experience what works and what should be changed.
One important caveat is that I do not work with agents; each step goes through a fairly rigid manual review phase.
I don’t agree. I think the way maybe the author used planning to maintain an understanding of the changes is largely not necessary, but there are other good use cases.
For example, it’s a great hook in the process for agentic review.
Get a second agent to look at what will be implemented and check it for inconsistencies, check it against whatever decisions were made or provided previously in the chat, or against whatever technical rules you’ve written out for your project, before going ahead. It surfaces a large number of opportunities for refinement, and generally pushes the output closer to the direction you’re looking for.
I've been working on a custom orchestration system which starts by creating an overall work plan, then splitting up development tasks among agent/model combos depending on their capabilities/expense, then coordinating their output, testing, etc. with various adversarial agents acting as gates throughout the process. It's really satisfying to watch the task board as agents go through their little routines.
It's been an absolutely fantastic way to set tokens on fire and watch them burn, and produce pretty much nothing of value.
Planning largely doesn't work because it's extrapolation. You don't know for sure how that remote system or compiler will react to the proposed change until it actually occurs.
Execution is the answer. All the complex success stories I've seen involved the LLM iteratively probing live data sources with varying filters, throwing code changes at the compiler over and over, or invoking shell commands until it succeeds.
I believe there is a Yoda quote regarding this.
> We still haven’t solved how to help people stay oriented as hundreds of increasingly capable agents change a system at once…
Building an issue tracker that addresses this. It can replay the workflow after the fact like a movie, and pin down the parts that require human input via tagging and inline diffs in the tickets. It is git-native, lives in your repo alongside the code doesn’t require any external service.
https://ljtn.github.io/epiq/
I am not so sure honestly. You can still wreck a codebase if you don't plan because the ai can make assumptions you don't want, or as other people say it's all a matter of taste
I also noticed this personally. Consider when an agent says
"Should I start with the spec, or go straight to building it?
Unless you're working on a super crucial piece of engineering, you can probably get away with going straight to building it. Even if something doesn't go the way you intended, I find that it's usually faster to correct the agent later, once the initial implementation is in place. It's more of an iterative approach to building and I feel that it is less cognitively demanding.
I find it interesting that what the author arrives at is the OODA loop.
The final form became: understand → act → inspect → clarify → adjust → act again
But he says that "understand" isn't really fitting (or something like that), so we are at: act → inspect → clarify → adjust → act again
That can be understood as "observe, orient, adjust, act" - but with other words.
That is the method to use when you are IN the sh*t rather than removed from it and making some theoretical plan.
I thoroughly enjoyed the read <3
I only really used it because the harness was way too trigger happy to start making changes. Even if I just asked a question some times I would come back and it refactored the whole codebase. Now it doesn't seem to do that any more.
I still would appreciate a "read-only" mode. It's not uncommon that I start a harness ONLY to explore and understand the code and I don't really want one typo to have it off building something, or even to save a plan document.
It's quite interesting. I use blend mode pretty heavily. On both claude code and codex.
But it's no longer to plan a single PR that can be done easily without plan mode.
It is mostly because I am more geared towards having my system work on initiative level changes where it works for days and plan is a good document to maintain to keep the agent aligned with the original goal and avoid unnecessary drift.
I have rules.md read on every start of the session, in which some of the rules are:
Never start coding befire I say codenow, and add datetime. Before patch explain in plain English what it does. Each file has datetime added to the top of the file, and updated if exist. Each file has a backup copy, created with name.datetime and checked after patch vs backup file, and again with new .backup file for logic vs new backup
I think Plan mode is very useful when building something that needs step by step validation, but also needs a broad strategy before starting.
For instance, in games where I work, we often need to manually test work out by playing or by using tools in ways that aren't feasible for the AI to do. In that case, getting the human in the loop between steps is an organized process when it's following a staged plan.
I find a lot of value in using plan mode to get the LLM out of dead ends. If it's in Act and starting to spiral (evidence of looking for something or trying to better understand something), I'll kick into Plan and tell it to think a bit harder—also avoids it wasting tokens retrying the same techniques—which results in a much more considered, and more often than not, correct approach.
My problem with planning and having the models ask the user questions is that the questions the models ask are still often irrelevant: you’ll find yourself answering a bunch of questions like “Should I use the existing system or create a whole new one in a different language?” - yet it will then make assumptions about other things that were actually ambiguous.
I still use plan mode, but during plan mode I will spin out sub agents to implement the current draft in a tmp dir and bring back lessons. I feel this keeps me grounded in my original starting point, rather than ending up with a implemention that meanders through the agents own discovery process. It also lets me ask concrete questions about (possible) implementation
I think with Claude Code plan mode has value even if Fable “gets it” off the bat. It’s a chance to explicitly disambiguate intent through AskUserQuestion tool as well as utilizing the option to clear content when plan is approved. The automatic disambiguation step is RLHF’d out of the model even when it’s needed.
The second I discovered the plan start hook I hijacked it and replaced the built-in planner with my own. The built-in planner is useless.
Still working on and with this but had to move it private because marketplace installed plugins loose capability
https://github.com/samelie/claude-plugin-pnpm/tree/main/skil...
I haven't used plan mode for a long time. I do still plan, I usually ask claude to make a plan with me first an explain how it will work, then I iterate with it until I am satisfied and then just let it go. So far it has not started implementing on it's own before I told it to.
Plan mode is for me not the agent.
I also bounce around to lots of different models. Some are cheap and dumb, some not. When unfamiliar with a model, the last thing I want it to allow file edits and destructive tool calls. Having a mode explicitly for that is helpful.
I don’t think it’s dead at all. I use it as a phase transition to let the agent decide when discovery is sufficient, as an intermediate phase between that and implementation. Even as a human dev you rarely go tinkering with files all over without first forming a plan.
In Claude Code it started to feel like it just asks 2 questions and sometimes there isnt really needed 2 questions so the options there often felt like choosing to downgrade or do something actually dumb because it was forced to suggest 2 things.
It's been a while (well, 8 months) since I used Claude Code, but what I can say is that I greatly enjoy using the plan mode that comes baked into omp (oh my pi), especially for tasks around firmware development. I recognize that has a very different workflow than webapps though.
I’ve been waiting for some kind of steering mode.
Imagining a mode where the agent is very transparent about its direction and process, and I can passively interject at any time to steer the outcome, without having to wait my turn or hard-interrupt.
I'd like a brainstorming mode, with limited side effects.
I want it to go and read stuff but also invoke some commands and generate reports and discuss on the results. Latest GPTs in codex plan mode always go.. write a plan. Shocking, i know, but thats not what i want on every turn when im in that mode.
I generally only work with the default claude auto and see that it does pretty well for my needs
It is? I still found it invaluable, when having a lot of externally managed knowledge. Basically the only gate before it goes on a certain failure, when dealing with proprietary libraries, tools and services
Switched from opencode to pi and I do kinda miss planning mode. I like spending a bit more time in that phase and having multiple providers LLM look at the spec doc etc
That’s very much coming from a desire for improved quality than understanding though
I work on long horizon scientific tasks with expensive computational tools. In that case users are basically asking for the harness to stop and confirm the cost (and plan) before it goes ahead
YES, and I also find that plan mode incentivizes bad practices and results in steadily reduced thought. On bad practices, I refer to zero/one-shot big bangs over iterative features/bugfixes. On reduced thought, assuming the plan was sufficiently comprehensive because it seemed to ask reasonable questions.
It can still be used in ways that I personally consider correct, but I think the parts I personally consider incorrect are so inherently alluring that I find plan mode to be an overall net negative for software development. I celebrate its apparently impending default removal (at least Claude Code and OpenCode are openly stating that they think it's time for it to go).
“Plans are of little importance, but planning is essential” - Churchill
I have noticed this in practice the past month. It’s interesting because goals and plans are obsolete. Skills too.
Much of what distinguished a knowledgeable AI dev a year ago is now an anti pattern.
A bit late to this party but what the heck.
I too have started using plan mode less, mainly for two reasons: 1. I plan my prompt more carefully and think about architecture up front 2. I found with more recent powerful models, the clarifying questions were generally not useful because it was pointing out issues it obviously knew the answer to and would have resolved in implementation anyway. So essentially they became time-wasting and anxiety inducing for no good reason
The mental shift I made was to not accept poor understanding of a code base on my part when writing the prompt - if I don’t understand it, I can’t predict what assumptions the model will make even on a basic level.
I used to tell myself that plan mode mitigated that, but I usually ended up mentally glossing over the generated plans anyway.
That just resulted in pure anxiety-driven engineering, where I’d often spend extra cycles verifying what was built and worrying about the design.
So invest the time understanding the system, at an appropriate level. That level will change over time as models get better.
+1 to a lot of what you said here, especially this:
> I used to tell myself that plan mode mitigated that, but I usually ended up mentally glossing over the generated plans anyway.
At a certain point I realized I far preferred delegating the first implementation pass to agents and reviewing the _code_ instead of a plan up front. If you had asked me this question ~2 years ago I'd think anyone would be insane to hand off discretion like this, but frontier models are just so _good_ nowadays.
Let's say you launch 10 working sessions in a day, would you rather:
* Review 10 plans, and _then_ review 10 PRs, and _then_ maybe adjust the approach on 1-2 and merge the other 8
* Review 10 PRs, and maybe adjust the approach on maybe 4 and merge the other 6
The first option just feels like needless attention for the sake of feeling in control; but the reality is that these models are becoming just as good if not better as us humans and our reckoning is here. Our codebase has enough linters, hooks, and guards such that generally the agent just follows our blessed patterns already, so what is plan mode actually doing beyond giving me a false sense of security? I'm still going to read the code anyways.
Maybe we need to figure out how humans should communicate with AI
so I use "opusplan" mode bc Opus is more expensive and supposedly better at thinking, and Sonnet is good and cheap for executing....
Changing models still gives us a usage hit (right?) but that is outweighed by benefits?
Plan mode is more about be rambling on while Opus detangles my thoughts.
Use instead:
or Or anything similar.This article is also a nice visualization on why spec-driven development has not lived up to its promises.
I dont know about Claude Code, but I'm using Amp Code since November 2025 and it don't need the mode at all, just conversation.
>We still haven’t solved how to help people stay oriented as hundreds of increasingly capable agents change a system at once.
Engineer, meet business school.
I’m curious if anyone has used automations for any kind of regular refactoring and cleanup tasks? Most agent guis support it.
Plan mode is not dead, it just moved to OpenSpec and other similar tools.
If you’re not using those either… I pray for your codebase.
I would have assumed plan modes make use of tools to keep the goal/plan persistent across compaction etc. Is that not the case?
It seems people will never tire of finding out that writing code isn't the hard part of making useful software.
> how do humans maintain a coherent mental model of a software system
This is the core question and it is wrong. Because it focuses on the software system rather than the problem you are trying to solve. A better version would be:
You need to have a picture of the whole process and everything related, not just your architecture and code.I disagree I want a proper read only mode, a lot of times models are still trigger happy.
Ironically enough Google Antigravity added plan mode this week.
Waterfall is dead. Agile iteration is the future.
That’s the amazing realization?
i disagree, i think it is very useful to have the ability to ask anything about the code and have the confidence that it WONT execute supid stuff
Cool article from a cool founder. Her other posts are great.
The Planning vs Plan is the core of the post.
Why is a plan "mode" needed? I really hate reading plans in conversation UI. One line in AGENTS.md is always better, and you can customize it to fit your document convention. From the communication perspective, the planning process is most about creating clarity and alignment, on things like scope, constraints and decisions, among people (and agents now), often requiring multiple rounds. The planning tool built in Codex or Claude Code is not persisted, version-controlled, nor well accessible. In practice, My AGENTS.md contains an instruction about writing a document before starting implementation. Usually I won't read that document, because I don't want to micromanage agents. That document mostly serves a historical purpose like ADR, helping me find out what agents missed, made mistakes on or misaligned on, if needed.
He just rediscovered agile to replace waterfall. But seriously, I am reading an early-release book Regenerative Software: https://learning.oreilly.com/library/view/regenerative-softw... I don't 100% agree with the book, but it contains a lot of good ideas about how to externalize knowledge about a piece of software from the code, which is what everyone tries to do with plans, specs, chats.
I totally disagree. This perspective only works if you don't care about token cost.
I'm using Astra for some homelab work and I tell it I want to do something and then it burns through a giant percentage of my 5-hour allowance (on the £20/mo subscription) doing that in a really weird way. I've had to tell it to stop zooming off and doing things, calling loads of tools and looking up websites, and just have a quick conversation first. It's much better now, but I've basically just reimplemented plan mode via AGENTS.md.
i find astra's every other message is a small plan, and i have to confirm for it to go
why do you even need plan mode?
“discuss your plan with me before implementing anything”
theres your plan mode
yeah plan mode never was really alive to begin with
In other news, Antigravity (the one that isn't the IDE or CLI) just added plan mode: https://x.com/antigravity/status/2103611698800140697
> This cut-and-paste workflow felt clunky and made it arduous to work through an idea while keeping track of the current plan.
I mean, it is, and if that were the workflow I’d be very fed up of it by now, but the plan mode in Claude’s VS Code extension has supported select and annotate directly since at the very least early 2026.
oh interesting! I haven’t used vscode for anything meaningful since shortly after leaving GitHub (2023). I generally find it overwhelming but codex added the annotations too and I love it.
"AI-generated text is painful to read"
Even worse are slop graphics. A stick figure driving a realistically-illustrated horse and carriage??? Barf
hard no for sufficiently complex existing codebases. even astra misses nuances in my moderately high loc app. plan, then implement is much better and, more importantly, much safer. at least i review the plan.
[flagged]
[flagged]
[flagged]
[flagged]
[dead]
[flagged]
[flagged]
Can you please not post AI-generated or AI-edited comments to HN? It's not allowed here - see https://news.ycombinator.com/newsguidelines.html#generated and https://news.ycombinator.com/item?id=47340079.
Of course, it's impossible to know for sure what was LLM processed or not, but some of your posts (like this one) have been getting classified that way.
[dead]
[dead]
[dead]