To be fair, his Midas touch is a result of consistency and a lot of hard work.
It's like the gardener at one of the Oxford colleges said - it's really easy to create these perfect lawns, just turn up every day and trim and water it - for a couple hundred years.
there is so much more subversive marketing out there than any of us can really fathom. i try not to be too paranoid but it's getting a lot harder every day.
i know someone who worked in what we might call the 'astroturfing' space within the entertainment industry. after having a few discussions with him and with things like this[0] becoming more known, it's really difficult to afford any assumption of organic intent when money is on the line - especially at the scale that microsoft works at compared to something as comparatively quaint as the music industry.
It is not good for text to speech (TTS) as well. I am trying it for few days. First of all 1.5B model documentation is not there. 0.5B realtime is shit model. I was converting text, line by line and it was randomly adding music and couldn't handle special characters like "…".
I really disappointed with this model to say the least.
The OSI's take on this is that an open source model can be modified through fine-tuning etc, even if you can't rebuild it from scratch.
The problem with requiring "build from scratch" for open source models is that the number of interesting models with training data that can be openly licensed is close to zero.
If you trained your model on an unlicensed scrape of the web you can't release the data under an open source license!
I would personally disagree slightly with this take. Freely being able to use means IMHO, that this can be done for all applications in a legal (and ideally ethical) fashion. Regulation often requires to prove the quality or provenance of data. Open source has IMHO often a very libertarian view on things focusing on the rights of the user an not society in general.
With free libre software, where freedom and liberty are about what the end user is empowered with actually, the software is mostly metonymic. Free software, free society, because there are free people in the middle of course.
Right, as I said elsewhere, maybe let's just let "open-source" have it.
"Open-source" can be "anything you can go out and grab a copy of and use" but doesn't give you much legal certainty about any of it, and reserve "free software" for the other, better thing.
But, free software lost it's way around GPLv3. From the end user's perspective, GPLv3 says that you can only use the software if it's either a cloud service, on hypothetical open firmware devices, or if you install it yourself.
AGPLv3 partially solves the issue by blocking people like Google from using it to build proprietary cloud services that take away their users' freedom. (It still doesn't solve the problem where providers use network effects to achieve the same end game.)
> From the end user's perspective, GPLv3 says that you can only use the software if it's either a cloud service, on hypothetical open firmware devices, or if you install it yourself.
The anti-tivo clause bans things like Apple pre-installing GPLv3 software on macs, but allows them to let you use exactly the same software as long as they do not give users access to the binary. AGPLv3 blocks both use cases, GPLv2 blocks neither.
On the spectrum of "things that take away user freedom", withholding the source code is bad. Withholding the source code, the binaries and physical access to the computer is obviously much worse! This latter business model is heavily subsidized by GPLv3.
I don't understand this either. The GPL doesn't address end users and their use of software at all, to be technical. It only addresses what terms of copyright redistributors of GPLed software are allowed to apply in-turn to subsequent end users.
The point of the Free in free software was always to protect the users of the software, not the vendors or the redistributors. (This is why the license focuses on the redistributors -- the mechanisms of the license limit their rights in order to protect others' rights.)
The first sentence of the GNU manifesto says this, and a few sections later in the document elaborate on the point:
Note, in particular, footnote [1] which explains that its OK for distributors to ask for payment, but that it's never OK for users to have to ask for permission to use the software, and the section "Why I Must Write GNU".
Since then, software service monopolies became common, and all of the most end-user-hostile systems on earth rely heavily on the GNU system. At this point, we're paying for permission to use those services with our money, our data, our democracy, etc.
I certainly cannot give you permission to use any of the GPLed services that I have used, or that I've been paid to extend. Therefore, I say the free software movement has lost its way.
Devils advocate here: I can give you a binary of my open source MIT code and never phone you the code. The code is still MIT licensed, and open source. You just have no access to it.
That said, I entirely agree that MS is misrepresenting their openness here, which isn’t in the least surprising.
In their defense, most everyone else does the same thing. They still shouldn't do it, but at least they're not the trendsetter here (though they are contributing to the ongoing problem)
I don't disagree, but it is perfectly acceptable per the MIT license, which is an OSI approved license. MIT doesn't require source distribution with the binary (which is why from the developer perspective, it's a more "permissive" license)
The license describes what users are allowed to do with the source code, it doesn’t (and shouldn’t) define what a creator has to do to make the source code open.
> inventor of GIF didn't begin with a document clearly laying out what is and isn't to be called a "GIF”*
Neither did the inventors of AI. A third party published a document after corporations went with open weights = open source and a spoiler block in FOSS wanted all training data published.
> it's right to push back whenever a huge tech corporation tries to build goodwill by falsely using terms like "open source
I think it’s counterproductive. Most people only see a squabble, which makes any ensuing points from the open-source community seem silly. Those who care can continue using the more-precise language they choose to.
Put another way, there is a difference between using terms like cracker and fully spelling out cryptocurrency, and telling people who use hacker and crypto more loosely that they’re wrong. They aren’t wrong and that isn’t meaningful feedback. At the same time, the person using the precise language isn’t wrong either.
There's a big difference between correcting some random commenter on an internet forum and correcting Microsoft.
> think it’s counterproductive. Most people only see a squabble, which makes any ensuing points from the open-source community seem silly.
Only to people that truly don't care whether something's open source. In which case, Microsoft using the term (correctly or incorrectly) won't change their perception.
But the people who do care won't like to be mislead by Microsoft. There's a reason the term is right in the headline: people respond to it.
I wish I had time to come up with a better example, but it's like if a AAA game company says they've released "native Linux build," but really they're just packaging the Windows build with Wine.
99% of people won't care, neither about the news nor the deception. But for that last 1%, any goodwill garnered with the headline would be gone, and the game company are the ones who look foolish, not the people calling them out.
I've been pronouncing both of them as /dʒis/ like hiss and not /dʒɪz/. I however am not a native english speaker of English. I wonder if native speakers gravitate towards the z more?
I would end both with the S sound, but I'm operating under the assumption that the person I was replying to either pronounces their Ss as Zs or can't tell the difference between the S and Z sounds.
Because the other assumption I could have gone with is the less charitable take that they know GIS with a soft G doesn't sound like jizz, but they were just looking for a crude way to mock the soft G.
Same way I pronounce my first name btw ;) but I think of "gif" as "gift" and this is probably the subconscious association people make without realizing it.
Which is why I find it fun to bring up that in Old English "gift" hadn't yet picked up the "t" and was spelled "gif", but in Old English "g" was most commonly "HY". I like the Old English pronunciation of "gif" as "HYEEF", which is a "compromise" position that often makes some of both soft-g and hard-g "gif" pronunciation fans angry.
I sometimes just pick the opposite of whatever everyone agreed to just for fun. I do the same when people cry about vim or emacs since I have used both. ;)
Some men just want to watch the world burn. At least it's mostly harmless fun anyway. It's even funnier when they bring up how my name is pronounced in defense of "jiff" and I tell them, so you're calling me the expert in "Gi" pronunciation then? :)
Basically hallucinations are false external things, and delusions false internal things. You hallucinate a pink elephant, you delude yourself into thinking trump won 2020.
I mean, you have "AI" which means just about anything in marketing speak, "Agentic" is kind of becoming similar, hopefully they don't goof that one too badly, would be nice to know what you are trying to sell me. Used to be "Cloud" meant storage not just hosting (I guess it still does).
Then there's "Smart" in front of Car, Phone, TV, and so on... Meaning different things.
I do think "Open Weight" should be more commonly used. There's definitely communities that spring up that build the training infrastructure and inference infrastructure around open models on the other hand.
Open weights is not exactly right either because we do get source of the software that uses those open weights.
Maybe open inference?
But we often also get source code for fine tunning the model.
So maybe it's closer to open source than to anything else?
Isn't it a bit like not calling a game open source because engine tooling used to made it isn't open source and they didn't publish .psd files with asset designs?
I'm genuinely torn on this one; I get technically why not, but why I think I have no problem with it is the wishy-washiness of "open source" generally.
As I teach this stuff to people newer to this tech, it's probably just easier and more helpful to refer to the wide array of "stuff you can just download and use yourself" as "open-source" and then after that, go deeper and talk about why Stallman was right, how "Free Software" was first. etc.
> “This means a future of abundance. A future where there is no poverty, where people can have whatever they want in terms of goods and services.” – Elon Musk
> “I think we see a path now where the world gets much more abundant and much better every year.” – Sam Altman
In my mind, Vibe-anything means "some slop carelessly thrown together to ship as fast as possible." Wild that it's being used in a serious product name!
"get offended" is just what the clickbait news cycle made of it. It was based on the post at [1], and this is all it said:
> We need to get beyond the arguments of slop vs sophistication and develop a new equilibrium in terms of our “theory of the mind” that accounts for humans being equipped with these new cognitive amplifier tools as we relate to each other
When a CEO says "We need to get beyond the arguments of X" it is universally a polite, PR-scrubbed way of saying, "Please stop talking about X, it is hurting our business" which is how the media interpreted it.
Look at the "News" section in the readme - The original TTS model is gone from this repo (you can still find it other places), but the SST/ASR, long form TTS, and streaming TTS models are newer.
When explanations get posted directly in HN comments, I imagine someone somewhere in the world is able to learn in spite of their Internet restrictions/firewalls
People will also post their own interpretations in response to comments, and quickly find out they missed something.
… But if you try to automate it, like include a summary under every HN post, you encourage laziness too much and are pre-chewing too heavily. Some balance here.
[on topic]
(OK I’m done making excuses, time to read the article… thanks for the encouragement!)
I thought this was not explained in the readme directly but in fact I missed it. I wasn’t going to read Microsoft entire changelog! But it was substantive, thanks to sibling commenter:
“2025-09-05: VibeVoice is an open-source research framework intended to advance collaboration in the speech synthesis community. After release, we discovered instances where the tool was used in ways inconsistent with the stated intent. Since responsible use of AI is one of Microsoft’s guiding principles, we have removed the VibeVoice-TTS code from this repository.”
I've been using VibeVoice's ASR (speech to text) model quite intensively for the past month and have found it to be a lot more reliable and out-of-the box functional then Whisper, parakeet and other models. The fact that is has diarization built into to the model is a huge win in my book. Without that you have to run a different model just for that which adds significantly to the overall processing time vs VibeVoice which gives you reliably great results. Big fan.
Note that this just covers the Speech-to-Text/Speech-Recognition aspect (a-la whisper), there's also models for long-form Text-To-Speech and steaming Text-To-Speech.
I'd be willing to bet it will be "Word of the Year" for 2026. Merriam-Webster had 'slop' for 2025, and 'polarization' for 2024. Is there a prediction market for this?
I took a look into local options for ASR and diarization some months ago, I missed that VibeVoice now has this feature.
My conclusions back then (which only came from a shallow research on the topic and 0 real experience mind you) was that Whisper + Pyannote was the "stable" approach.
Have the VibeVoice, Voxtral, Qwen or the Nemo solutions caught up in segmentation and speaker recognition?
Local? No idea. Cloud? Eleven Labs, probably. But it's described as "cloning" not "training". Not sure what the distinction is or why it matters if the end result is you can to generate any TTS that sounds like you. There might very well be an important one, I just don't know it.
Seems quite heavy for a STT model, Parakeet and Whisper are much smaller and perform great for quick dictation and transcription of longer files. I guess that's due to additional accuracy and speaker diarisation?
The TTS example clip in the repo of 'spontaneous singing' is creepy as fuck
This is not a new model. Also, it hallucinates a lot. Also, it's very heavy and slow in inference. It's also bad in multilingual.
Edit: I'm talking purely about speech to text (STT). Not sure about the other things this can do.
you saved us a lot of time here.... i unstarred the repo
moving on....
You just saved me an afternoon.
Yeah, I don't get why it is suddenly getting so much attention today, it is all over twitter too
Simonw (who has a bit of a Midas touch for posts here) just posted about it https://simonwillison.net/2026/Apr/27/vibevoice/
To be fair, his Midas touch is a result of consistency and a lot of hard work.
It's like the gardener at one of the Oxford colleges said - it's really easy to create these perfect lawns, just turn up every day and trim and water it - for a couple hundred years.
there is so much more subversive marketing out there than any of us can really fathom. i try not to be too paranoid but it's getting a lot harder every day.
i know someone who worked in what we might call the 'astroturfing' space within the entertainment industry. after having a few discussions with him and with things like this[0] becoming more known, it's really difficult to afford any assumption of organic intent when money is on the line - especially at the scale that microsoft works at compared to something as comparatively quaint as the music industry.
[0] https://www.wired.com/story/geese-chaotic-good-marketing-ind...
well duh, they updated the news section
https://github.com/microsoft/VibeVoice/commit/e73d1e17c3754f...
which is microsoft for "we removed two dead links". AI innovation knows no limits!
Interestingly that seems to be in response to [1], which might indeed be the trigger for this.
[1] https://doublepulsar.com/microsoft-vibing-capturing-screensh...
It is not good for text to speech (TTS) as well. I am trying it for few days. First of all 1.5B model documentation is not there. 0.5B realtime is shit model. I was converting text, line by line and it was randomly adding music and couldn't handle special characters like "…".
I really disappointed with this model to say the least.
I think this was all covered when they said it was released by Microsoft?
The nuance is lost on LLM agentic dominant partakers.
I think in this category, Voxtral by Mistral is a lot better. It also happens to be small enough to run on webGPU https://huggingface.co/spaces/mistralai/Voxtral-Realtime-Web...
I think we should stop calling this type of models open source. They are indeed "open weight." The training code is proprietary and never revealed.
https://github.com/microsoft/VibeVoice/issues/102
I'm reserving that complaint for "open source" models which are released under non-open-source licenses.
I care that I know what I can DO with the project when I see it described as "open source".
> I care that I know what I can DO with the project when I see it described as "open source".
Yes, the first of which is that you should be able to build it from source. Which requires the source code, and in this case data.
The OSI's take on this is that an open source model can be modified through fine-tuning etc, even if you can't rebuild it from scratch.
The problem with requiring "build from scratch" for open source models is that the number of interesting models with training data that can be openly licensed is close to zero.
If you trained your model on an unlicensed scrape of the web you can't release the data under an open source license!
The Open Source Initiative have a bunch of their thinking around this in their FAQ for the "Open Source AI definition": https://opensource.org/ai/faq#isn-t-training-data-required-t...
I would personally disagree slightly with this take. Freely being able to use means IMHO, that this can be done for all applications in a legal (and ideally ethical) fashion. Regulation often requires to prove the quality or provenance of data. Open source has IMHO often a very libertarian view on things focusing on the rights of the user an not society in general.
They’ll never reveal the data, because that would reveal this is all built on stolen work.
Some of the models DO reveal the data, and it's still built on "stolen work" in that it's unlicensed scrapes of the Web. Here's an example:
https://huggingface.co/allenai/OLMo-2-0325-32B
Here's one of their training mixes: https://huggingface.co/datasets/allenai/dolma3_pool - which includes 8 trillion tokens from Common Crawl.
That would be “permissive license”
Maybe we should have a little cue card for models: vendor/name, size, open weights, open source, permissive license.
It’s simple enough an idea.
Indeed. We now live in a world where freeware is named open source. We are very sorry, Stallman.
If you're going to apologize to Stallman, you should apologize for conflating open source with software freedom. ;D
I totally get you, but this is yet another thick layer away.
With free libre software, where freedom and liberty are about what the end user is empowered with actually, the software is mostly metonymic. Free software, free society, because there are free people in the middle of course.
Right, as I said elsewhere, maybe let's just let "open-source" have it.
"Open-source" can be "anything you can go out and grab a copy of and use" but doesn't give you much legal certainty about any of it, and reserve "free software" for the other, better thing.
But, free software lost it's way around GPLv3. From the end user's perspective, GPLv3 says that you can only use the software if it's either a cloud service, on hypothetical open firmware devices, or if you install it yourself.
AGPLv3 partially solves the issue by blocking people like Google from using it to build proprietary cloud services that take away their users' freedom. (It still doesn't solve the problem where providers use network effects to achieve the same end game.)
> From the end user's perspective, GPLv3 says that you can only use the software if it's either a cloud service, on hypothetical open firmware devices, or if you install it yourself.
What in the world do you mean?
The anti-tivo clause bans things like Apple pre-installing GPLv3 software on macs, but allows them to let you use exactly the same software as long as they do not give users access to the binary. AGPLv3 blocks both use cases, GPLv2 blocks neither.
On the spectrum of "things that take away user freedom", withholding the source code is bad. Withholding the source code, the binaries and physical access to the computer is obviously much worse! This latter business model is heavily subsidized by GPLv3.
I don't understand this either. The GPL doesn't address end users and their use of software at all, to be technical. It only addresses what terms of copyright redistributors of GPLed software are allowed to apply in-turn to subsequent end users.
The point of the Free in free software was always to protect the users of the software, not the vendors or the redistributors. (This is why the license focuses on the redistributors -- the mechanisms of the license limit their rights in order to protect others' rights.)
The first sentence of the GNU manifesto says this, and a few sections later in the document elaborate on the point:
https://www.gnu.org/gnu/manifesto.html
Note, in particular, footnote [1] which explains that its OK for distributors to ask for payment, but that it's never OK for users to have to ask for permission to use the software, and the section "Why I Must Write GNU".
Since then, software service monopolies became common, and all of the most end-user-hostile systems on earth rely heavily on the GNU system. At this point, we're paying for permission to use those services with our money, our data, our democracy, etc.
I certainly cannot give you permission to use any of the GPLed services that I have used, or that I've been paid to extend. Therefore, I say the free software movement has lost its way.
Devils advocate here: I can give you a binary of my open source MIT code and never phone you the code. The code is still MIT licensed, and open source. You just have no access to it.
That said, I entirely agree that MS is misrepresenting their openness here, which isn’t in the least surprising.
In their defense, most everyone else does the same thing. They still shouldn't do it, but at least they're not the trendsetter here (though they are contributing to the ongoing problem)
? Do you know what “source” means in open source? Like, what is the source of the binary? It’s the code. That’s the source in open source.
I don't disagree, but it is perfectly acceptable per the MIT license, which is an OSI approved license. MIT doesn't require source distribution with the binary (which is why from the developer perspective, it's a more "permissive" license)
The license describes what users are allowed to do with the source code, it doesn’t (and shouldn’t) define what a creator has to do to make the source code open.
> we should stop calling this type of model open source. They are indeed "open weight”
This ship has sailed. It’s now in the same category as hacker/cracker and the pronunciation of GIF.
The inventor of GIF didn't begin with a document* clearly laying out what is and isn't to be called a "GIF."
I think it's right to push back whenever a huge tech corporation tries to build goodwill by falsely using terms like "open source."
*https://opensource.org/osd
> inventor of GIF didn't begin with a document clearly laying out what is and isn't to be called a "GIF”*
Neither did the inventors of AI. A third party published a document after corporations went with open weights = open source and a spoiler block in FOSS wanted all training data published.
> it's right to push back whenever a huge tech corporation tries to build goodwill by falsely using terms like "open source
I think it’s counterproductive. Most people only see a squabble, which makes any ensuing points from the open-source community seem silly. Those who care can continue using the more-precise language they choose to.
Put another way, there is a difference between using terms like cracker and fully spelling out cryptocurrency, and telling people who use hacker and crypto more loosely that they’re wrong. They aren’t wrong and that isn’t meaningful feedback. At the same time, the person using the precise language isn’t wrong either.
There's a big difference between correcting some random commenter on an internet forum and correcting Microsoft.
> think it’s counterproductive. Most people only see a squabble, which makes any ensuing points from the open-source community seem silly.
Only to people that truly don't care whether something's open source. In which case, Microsoft using the term (correctly or incorrectly) won't change their perception.
But the people who do care won't like to be mislead by Microsoft. There's a reason the term is right in the headline: people respond to it.
I wish I had time to come up with a better example, but it's like if a AAA game company says they've released "native Linux build," but really they're just packaging the Windows build with Wine.
99% of people won't care, neither about the news nor the deception. But for that last 1%, any goodwill garnered with the headline would be gone, and the game company are the ones who look foolish, not the people calling them out.
I think you mean GIF.
It's the same as GIS, you wouldn't say jizz now would you?
I absolutely do, every single time it comes up.
The developer of the format declared the pronunciation 30+ years ago. It has always been jif.
Yeah, but society overruled them.
I hadn't thought about how to pronounce GIS, but do you have a problem with the pronunciation of the Japanese Industrial Standards: JIS?
I've been pronouncing both of them as /dʒis/ like hiss and not /dʒɪz/. I however am not a native english speaker of English. I wonder if native speakers gravitate towards the z more?
I would end both with the S sound, but I'm operating under the assumption that the person I was replying to either pronounces their Ss as Zs or can't tell the difference between the S and Z sounds.
Because the other assumption I could have gone with is the less charitable take that they know GIS with a soft G doesn't sound like jizz, but they were just looking for a crude way to mock the soft G.
I think it depends on region. Related, many speakers pronounce chips and salza, Tezla, Wezley.
i am absolutely going to from now on
I take it that you haven’t met the Arcgees people…
How do you pronounce giraffe?
Same way I pronounce my first name btw ;) but I think of "gif" as "gift" and this is probably the subconscious association people make without realizing it.
Which is why I find it fun to bring up that in Old English "gift" hadn't yet picked up the "t" and was spelled "gif", but in Old English "g" was most commonly "HY". I like the Old English pronunciation of "gif" as "HYEEF", which is a "compromise" position that often makes some of both soft-g and hard-g "gif" pronunciation fans angry.
I sometimes just pick the opposite of whatever everyone agreed to just for fun. I do the same when people cry about vim or emacs since I have used both. ;)
Some men just want to watch the world burn. At least it's mostly harmless fun anyway. It's even funnier when they bring up how my name is pronounced in defense of "jiff" and I tell them, so you're calling me the expert in "Gi" pronunciation then? :)
I have never heard this third option before but I love it!
How do you pronounce gift?
gorge = george
And "hallucination" which should have been "delusion".
Way early on (spring 2023) people tried to stop it, but no luck.
Why would it be delusion? It’s making something up which isn’t there and describing it.
A hallucination is a false sensory experience.
A delusion is a false mental belief.
Basically hallucinations are false external things, and delusions false internal things. You hallucinate a pink elephant, you delude yourself into thinking trump won 2020.
At least it's MIT licensed! As much as non-open training data irks me, restrictive licensing irks me more!
what is problem with restrictive licensing? Most of them starts if you have 1M users etc?
What you said makes a lot of sense. Free software should not be confused with open source
I mean, you have "AI" which means just about anything in marketing speak, "Agentic" is kind of becoming similar, hopefully they don't goof that one too badly, would be nice to know what you are trying to sell me. Used to be "Cloud" meant storage not just hosting (I guess it still does).
Then there's "Smart" in front of Car, Phone, TV, and so on... Meaning different things.
I do think "Open Weight" should be more commonly used. There's definitely communities that spring up that build the training infrastructure and inference infrastructure around open models on the other hand.
Open weights is not exactly right either because we do get source of the software that uses those open weights.
Maybe open inference?
But we often also get source code for fine tunning the model.
So maybe it's closer to open source than to anything else?
Isn't it a bit like not calling a game open source because engine tooling used to made it isn't open source and they didn't publish .psd files with asset designs?
I'm genuinely torn on this one; I get technically why not, but why I think I have no problem with it is the wishy-washiness of "open source" generally.
As I teach this stuff to people newer to this tech, it's probably just easier and more helpful to refer to the wide array of "stuff you can just download and use yourself" as "open-source" and then after that, go deeper and talk about why Stallman was right, how "Free Software" was first. etc.
Openwashing is the new greenwashing, which, coincidently, seems to have gone out of fashion a few hundred datacentres ago.
it was replaced with abundancewashing
What is "abundancewashing"?
> “This means a future of abundance. A future where there is no poverty, where people can have whatever they want in terms of goods and services.” – Elon Musk
> “I think we see a path now where the world gets much more abundant and much better every year.” – Sam Altman
https://www.diamandis.com/blog/elon-sam-abundance
Interesting to see "vibe" enshrined by the likes of Microsoft as an AI product word.
Maybe they were trying to make a pun on "Via Voice", the cursed IBM STT from the 90s?
Especially when "vibe coded" can have a negative connotation meaning quickly put together without understanding.
In my mind, Vibe-anything means "some slop carelessly thrown together to ship as fast as possible." Wild that it's being used in a serious product name!
I’m just surprised they put the name of the e-waste slop company in their product
I'm honestly more surprised that they could resist the temptation to call it Copilot
Which makes it even more weird they get offended when people use Mircoslop. They are the ones leaning into the marketing
"get offended" is just what the clickbait news cycle made of it. It was based on the post at [1], and this is all it said:
> We need to get beyond the arguments of slop vs sophistication and develop a new equilibrium in terms of our “theory of the mind” that accounts for humans being equipped with these new cognitive amplifier tools as we relate to each other
[1] https://snscratchpad.com/posts/looking-ahead-2026/
Are you sure you have the correct reference?
I think everyone else is relating to
https://futurism.com/artificial-intelligence/microsoft-bans-...
When a CEO says "We need to get beyond the arguments of X" it is universally a polite, PR-scrubbed way of saying, "Please stop talking about X, it is hurting our business" which is how the media interpreted it.
This is a very good model, but can it be run on the web?
Isn't this project the one Microsoft published but then soon after pulled it for security/safety reasons? What has changed since then?
Look at the "News" section in the readme - The original TTS model is gone from this repo (you can still find it other places), but the SST/ASR, long form TTS, and streaming TTS models are newer.
It’s confusing (at least for me) because the project covers a number of things including what you are mentioning.
[off topic]
When explanations get posted directly in HN comments, I imagine someone somewhere in the world is able to learn in spite of their Internet restrictions/firewalls
People will also post their own interpretations in response to comments, and quickly find out they missed something.
… But if you try to automate it, like include a summary under every HN post, you encourage laziness too much and are pre-chewing too heavily. Some balance here.
[on topic]
(OK I’m done making excuses, time to read the article… thanks for the encouragement!)
I thought this was not explained in the readme directly but in fact I missed it. I wasn’t going to read Microsoft entire changelog! But it was substantive, thanks to sibling commenter:
“2025-09-05: VibeVoice is an open-source research framework intended to advance collaboration in the speech synthesis community. After release, we discovered instances where the tool was used in ways inconsistent with the stated intent. Since responsible use of AI is one of Microsoft’s guiding principles, we have removed the VibeVoice-TTS code from this repository.”
Interesting story about this repo/product/author by cybersecurity researcher Kevin Beaumont: https://cyberplace.social/@GossiTheDog/116454846703138243
Still waiting for the open weights model that conclusively beats the multi-year old Whisper in accuracy, features, and performance.
It's crazy that a lot is happening in open models for stt, but there's very little progress when it comes to results, esp multilingual.
Shouldn't it be called something like "Copilot Voice"?
I've been using VibeVoice's ASR (speech to text) model quite intensively for the past month and have found it to be a lot more reliable and out-of-the box functional then Whisper, parakeet and other models. The fact that is has diarization built into to the model is a huge win in my book. Without that you have to run a different model just for that which adds significantly to the overall processing time vs VibeVoice which gives you reliably great results. Big fan.
Great post last night from Simon: https://simonwillison.net/2026/Apr/27/vibevoice/
Note that this just covers the Speech-to-Text/Speech-Recognition aspect (a-la whisper), there's also models for long-form Text-To-Speech and steaming Text-To-Speech.
“VibeVoice can only handle up to an hour of audio”
Why?
Microsoft Store App Vibing.exe Accused of Harvesting Screens, Audio, and Clipboard Data:
https://cyberpress.org/microsoft-store-app-vibing-exe-accuse...
So we've really just settled on Vibe as the verb for AI then?
I'd be willing to bet it will be "Word of the Year" for 2026. Merriam-Webster had 'slop' for 2025, and 'polarization' for 2024. Is there a prediction market for this?
it'll probably be something we're not even talking about yet - we still have 7 months in which to make the world even worse
Why use precise technical language when you can just vibe with your AI system?
Holy moly, a Microsoft AI product that isn't named Copilot!
Missed opportunity to call it Vopilot
Slopilot
You have selected Microsoft Sam as the computer's default voice.
My friends and I had fun in the computer lab with Microsoft Sam, inputting long strings of characters to create funny sound effects. Sususususususu.
I took a look into local options for ASR and diarization some months ago, I missed that VibeVoice now has this feature.
My conclusions back then (which only came from a shallow research on the topic and 0 real experience mind you) was that Whisper + Pyannote was the "stable" approach.
Have the VibeVoice, Voxtral, Qwen or the Nemo solutions caught up in segmentation and speaker recognition?
It would have been better if they provided not just weights, but also some frontend where it is usable as is.
Microsoft has historically made poor choices in product naming, but this has to be a new low.
I the past month or so, I added 2 models to my app Whisper Memos (https://whispermemos.com):
- Cohere Transcribe (self hosted)
- Grok Speech To Text (they provide an API, only $0.10/hr!)
They are both excellent. I'm not sure about this one. Would you like to see it in a consumer speech to text app?
I've had good experiences with the Mistral Voxtral models (I've used the API, but some of the model-variants are open weight)
Does Cohere work with longer transcripts? Do you have to do some magic to merge recordings over 35 seconds long?
Have you tried qwen?
Any non-Musk alternatives that are comparable in quality and cost?
Voxtral competes on price ($0.003/min) and quality. Speechmatics has best in class accuracy but is a bit more expensive ($0.004/min)
Our default is still OpenAI Whisper. Grok is just a choice for users who might prefer it.
What’s the current state of the art, for each of training locally and in the cloud, for learning my voice?
Locally maybe https://voicebox.sh/
Elevenlabs in the cloud.
Local? No idea. Cloud? Eleven Labs, probably. But it's described as "cloning" not "training". Not sure what the distinction is or why it matters if the end result is you can to generate any TTS that sounds like you. There might very well be an important one, I just don't know it.
open weights i would say S2: https://github.com/rodrigomatta/s2.cpp
Maybe Microsoft’s real strength was never making the best model, it was knowing you don’t need to, as long as you own the platform everyone builds on.
https://www.youtube.com/watch?v=d_AP3SGMxxM
looks like this offers ASR support in GGUF https://github.com/CrispStrobe/CrispASR -- haven't tested
For me its giving me very poor results
English only?
Previously:
Sept 2025 https://news.ycombinator.com/item?id=45114245
That was about the text-to-speech model, the speech-to-text one was release in January.
Seems quite heavy for a STT model, Parakeet and Whisper are much smaller and perform great for quick dictation and transcription of longer files. I guess that's due to additional accuracy and speaker diarisation?
The TTS example clip in the repo of 'spontaneous singing' is creepy as fuck
Microsoft is famous for choosing terrible names but how could they be this terrible.
lol they rug-pulled the 7B for our own safety some months ago