The scenario seems broadly similar to one where efficient, low-cost foreign competitors undercut domestic firms and gain market share. Some domestic workers are out of a job either way.
It's not impossible but it will take time and is easier said than done.
Was writing software really the bottleneck, or are there others? Is Stripe really in that much better a position to compete with other payment rails now that there is AI? Or maybe a new competitor comes out of nowhere and a huge number of businesses switch? The scenario where markets become hyper-competitive because of AI seems a bit dubious.
I think this scenario drastically underestimates the amount of "moat" many of these software companies have on the consumer side.
Not sure about Stripe, but I think it would be extremely dubious for vibe-coded competitors to Uber to serve as real competition like described in the substack post, because the business benefits so much from network effects. A new upstart would not only need to replicate the technical aspects but also the physical vehicle network and years of built-up fraud prevention efforts (including things like rider/driver ratings and features that targeted at edge cases) of the existing incumbent firms.
Back in the 2015-2020 self-driving investment boom I recall all kinds of dire predictions about the impact that robo-taxis and trucks would have on the US economy. There was an immediacy to these predictions. That once a viable driverless truck became widely available all truckers in the US would need to find new jobs the next day.
We can now see that it's going to take a lot longer for the technology to become viable and widespread. The impact is going to be felt in small pieces over many years, likely decades.
I wonder if AI will force people to consider necessary deregulation to create jobs; in other words, many jobs are regulated out of existence and artificial scarcity is created which AI will exacerbate (hence AI isn't the problem, over-regulation is and always was the problem)
> The scenario imagines every consumer deciding to use their own personal agent to transact and conduct business. This completely sidelines companies that monetise “friction” in the economy, such as travel and estate agencies that operate as middlemen in processes such as booking holidays or buying property.
The threat to the middleman businesses is their elimination. For that to happen, the service providers have to open up APIs for the users' agents' to use directly, so that then those agents effectively become the middlemen.
If that doesn't happen, then the user agents have to deal with the friction, and that works in favor of the middlemen because agents are willing to deal with much more friction than people.
In this scenario, only the rank-and-file friction workers and their management have to worry about being replaced by friction-generating AI; the business as such is safe.
My prediction for 2026 with respect to AI confidently included a proliferation of doomer content. At least the "Citrini scenario" didn't have AI causing nuclear war.
> Speculative as it is, the scenario has unnerved investors. The S&P dropped more than 1% on Monday, and the software component of the index fell to its lowest level since Trump’s “liberation day” tariff announcement in April. Doubtless some of the wobble is attributable to Trump’s latest tariffs, but Uber, American Express, Mastercard and DoorDash, specifically named in Citrini’s report, all lost between 4% and 6%.
So how exactly does the article attribute the movements of the market to this Substack post? What is the proof that this was the cause? Journalism usually deals with proof of what it says but I don't see any in articles like this.
AI doomsday scenarios all have the same flaws. They systematically confuse social constructs with laws of nature.
Everything that starts with mass unemployment and don't finish with strikes, riots and the technology being massively regulated can safely be disregarded. If people have to chose between AI and society, they will chose society.
It's important to remember that things like the economy, the market or even money are social constructs which hold because people collectively decide to believe in them. Watch what happens to data centers when the people controlling the electric grids, manufacturing parts, growing the food, or driving the trucks carrying it stop doing so.
People in power know that. They also know they would be collateral damages if it reaches this point. Generally speaking people in power like to stay in power. This power depends on society staying an acceptable choice to most. Except regulations to happen far before AI starts rocking the boat too much.
It depends on whether people wake up to the threat before or after there is a robot army that can crush them, doesn't it? If humans are economically and militarily useless, it won't matter what they choose.
This is a valid point. However, the conclusion depends on the fact that the people in power have the information and wisdom to make the choices required stay in power. Which has been mostly true (so far) in US history.
Historically this has very much not always been the case (citation: any revolution)
But "oppressive power", via information control or military/police control is also a factor. If AI is able to give the ruling incumbents more power, that shifts the equilibrium between popular sentiment and people in power.
The steel belt was thoroughly fucked by decades of deliberate policy choices and yet people just accepted the decay while killing themselves with meth and opioids.
The starving don’t start revolutions, they die of starvation
The scenario seems broadly similar to one where efficient, low-cost foreign competitors undercut domestic firms and gain market share. Some domestic workers are out of a job either way.
It's not impossible but it will take time and is easier said than done.
Was writing software really the bottleneck, or are there others? Is Stripe really in that much better a position to compete with other payment rails now that there is AI? Or maybe a new competitor comes out of nowhere and a huge number of businesses switch? The scenario where markets become hyper-competitive because of AI seems a bit dubious.
I think this scenario drastically underestimates the amount of "moat" many of these software companies have on the consumer side.
Not sure about Stripe, but I think it would be extremely dubious for vibe-coded competitors to Uber to serve as real competition like described in the substack post, because the business benefits so much from network effects. A new upstart would not only need to replicate the technical aspects but also the physical vehicle network and years of built-up fraud prevention efforts (including things like rider/driver ratings and features that targeted at edge cases) of the existing incumbent firms.
Back in the 2015-2020 self-driving investment boom I recall all kinds of dire predictions about the impact that robo-taxis and trucks would have on the US economy. There was an immediacy to these predictions. That once a viable driverless truck became widely available all truckers in the US would need to find new jobs the next day.
We can now see that it's going to take a lot longer for the technology to become viable and widespread. The impact is going to be felt in small pieces over many years, likely decades.
I think the impact of LLMs will be similar.
I wonder if AI will force people to consider necessary deregulation to create jobs; in other words, many jobs are regulated out of existence and artificial scarcity is created which AI will exacerbate (hence AI isn't the problem, over-regulation is and always was the problem)
> The scenario imagines every consumer deciding to use their own personal agent to transact and conduct business. This completely sidelines companies that monetise “friction” in the economy, such as travel and estate agencies that operate as middlemen in processes such as booking holidays or buying property.
The threat to the middleman businesses is their elimination. For that to happen, the service providers have to open up APIs for the users' agents' to use directly, so that then those agents effectively become the middlemen.
If that doesn't happen, then the user agents have to deal with the friction, and that works in favor of the middlemen because agents are willing to deal with much more friction than people.
In this scenario, only the rank-and-file friction workers and their management have to worry about being replaced by friction-generating AI; the business as such is safe.
what was that other post? some dude on twitter wrote something to spook everyone and got a 100 million views, i dont remember
https://archive.is/TN9Ks
My prediction for 2026 with respect to AI confidently included a proliferation of doomer content. At least the "Citrini scenario" didn't have AI causing nuclear war.
> Speculative as it is, the scenario has unnerved investors. The S&P dropped more than 1% on Monday, and the software component of the index fell to its lowest level since Trump’s “liberation day” tariff announcement in April. Doubtless some of the wobble is attributable to Trump’s latest tariffs, but Uber, American Express, Mastercard and DoorDash, specifically named in Citrini’s report, all lost between 4% and 6%.
So how exactly does the article attribute the movements of the market to this Substack post? What is the proof that this was the cause? Journalism usually deals with proof of what it says but I don't see any in articles like this.
AI doomsday scenarios all have the same flaws. They systematically confuse social constructs with laws of nature.
Everything that starts with mass unemployment and don't finish with strikes, riots and the technology being massively regulated can safely be disregarded. If people have to chose between AI and society, they will chose society.
It's important to remember that things like the economy, the market or even money are social constructs which hold because people collectively decide to believe in them. Watch what happens to data centers when the people controlling the electric grids, manufacturing parts, growing the food, or driving the trucks carrying it stop doing so.
People in power know that. They also know they would be collateral damages if it reaches this point. Generally speaking people in power like to stay in power. This power depends on society staying an acceptable choice to most. Except regulations to happen far before AI starts rocking the boat too much.
It depends on whether people wake up to the threat before or after there is a robot army that can crush them, doesn't it? If humans are economically and militarily useless, it won't matter what they choose.
This is a valid point. However, the conclusion depends on the fact that the people in power have the information and wisdom to make the choices required stay in power. Which has been mostly true (so far) in US history.
Historically this has very much not always been the case (citation: any revolution)
But "oppressive power", via information control or military/police control is also a factor. If AI is able to give the ruling incumbents more power, that shifts the equilibrium between popular sentiment and people in power.
The steel belt was thoroughly fucked by decades of deliberate policy choices and yet people just accepted the decay while killing themselves with meth and opioids.
The starving don’t start revolutions, they die of starvation
Discussion on the report:
https://news.ycombinator.com/item?id=47114579