The AI Economic Doomsday Report That Shook Wall Street

0:05Ryan Knutson:Our colleague David Uberti was doom-scrolling on Sunday night, like a lot of us do, when a certain post caught his eye.

0:12David Uberti:I was reading in bed on my iPad this interesting subsec post from a financial research firm I hadn't heard of before.

0:21Ryan Knutson:It was published by a relatively unknown firm called Citrini Research. And it was written like a memo from the future, from June 2028, looking back at how artificial intelligence transformed the economy.

0:34David Uberti:And they basically framed this report as sort of like a post-mortem on what happened over the time between now and then and how the economy has changed. It read to me like really good science fiction. I didn't put it down, despite the fact that it was 7 ,000-plus words long. The picture the Citrini report painted of the future was bleak. We described it as a doomsday report, and very much so it was. In the scenario that they outlined, there was something like 10.2 % unemployment across the United States, which is worse than what it was at the depths of the Great Recession.

1:09Ryan Knutson:The report proposed that AI will become so good at writing code and replacing jobs that it could become very bad for the broader economy.

1:18David Uberti:Basically, the question is not whether, like, AI is bearish or bullish for the economy. Is it what if it's so bullish that it becomes bearish?

1:26Ryan Knutson:David wasn't the only one reading the Citrini post that night. It was going viral. And it was freaking people out.

1:34David Uberti:Pretty soon after the market opened on Monday morning, a lot of stocks that we follow in the software space, in particular, were all flashing red on our screens. And when something across the entire sector is moving in the same direction, it tends to mean something big is happening. The Dow dropping about 1.6%, losing more than 800 points. Salesforce, Snowflake, a few other software names. Syncredy and Capital One, those names plunging. So too is DoorDash and Uber. DoorDash, that stock was down over 6 % and all coming after that report we were talking about from Centrini Research, where they lay out the potential risks.

2:12Ryan Knutson:Why do you think this post about a hypothetical future almost like science fiction, led to such a big reaction in the stock market.

2:21David Uberti:I think it really articulated a lot of existing fears that people have about artificial intelligence. I think people who think a lot about this space and the uncertainty around it are looking for ways to understand it, and this definitely tapped into that vein.

2:39Ryan Knutson:Welcome to The Journal, our show about money, business, and power. I'm Ryan Knudsen. It's Friday, February 27th.

2:52Ryan Knutson:Coming up on the show, the AI Doomsday Report that shook the stock market.

3:38Ryan Knutson:At the end of last year, the biggest worry about AI in the stock market was that it was a bubble, that the technology was overhyped and the companies were spending too much money on it.

3:48David Uberti:There's a lot of questions about AI companies over-investing in AI, basically throwing billions upon billions of dollars into data center construction chips and more, and that all of that wouldn't actually pan out. But over the last couple of months or so, there's been a vibe shift of sorts. And it's shifted more from this idea that there's a bubble that might burst to this idea that AI might actually pan out.

4:13Ryan Knutson:This vibe shift was sparked by massive advancements in AI tools. Tools that allow for AI agents, or digital assistants, to actually do stuff for you. And there have been advancements in coding tools, like Anthropic's Cloud Code and OpenAI's Codex, which make it possible for just about anyone to write computer code.

4:32David Uberti:I mean, when you see people like me who are able to go into Cloud with no previous coding experience and do in a couple of hours what trained software engineers would take much longer to do, traditionally speaking. I mean, that's a pretty big development. And it raises the question of, you know, how quickly can people spin up new pieces of software? It wouldn't take the type of massive investment that you'd have from a traditional software company that takes billions upon billions of dollars to do this stuff. The idea being that it's actually much cheaper to do this now.

5:07Ryan Knutson:One AI founder compared this moment to February 2020, when we could see a pandemic brewing on the horizon in other countries, but didn't know exactly what was about to hit us.

5:17David Uberti:There's been this sort of deer-in-the-headlights moment on Wall Street of what people should do, and it has left the market in this very herky-jerky, trigger-happy mode. So it just sort of speaks to how this particular Substack post played into an existing trend, and in many cases exacerbated that trend.

5:40Ryan Knutson:Okay, well, let's dive into that Substack post. It was written by Citrini Research. First of all, what is Citrini Research?

5:48David Uberti:So Citrini Research is a small research firm. They do macro and stock research, which they have published on Substack for the last couple of years or so. They're not as widely known as a lot of the research outfits that we tend to follow, but they have a really sizable following on Substack. And in fact, they're one of the largest financial blogs on Substack. And this went more viral than anything else that they've done previously.

6:15Ryan Knutson:Tell me more about the future world that this report lays out. What do they think is going to happen?

6:20David Uberti:The scenario at a high level is that the sort of rapid advancement in adoption of AI will lead to this sort of cascading dynamic of job losses, disinvestment in a race to the bottom of prices. that ultimately leads to mass unemployment among white-collar workers.

6:38Ryan Knutson:Here's one of the authors of the report, Alep Shah, in an interview with Bloomberg.

6:43David Uberti:And so that's kind of the underlying thesis here, is that there's going to be significant replacement of jobs with AI, specifically AI agents. Those things only really came into fruition in the last few months. And so as it comes through and shows up in the productivity of different corporations, that's when things are going to get a little more interesting and where we've got to pay real attention.

7:05Ryan Knutson:The report lays out a potential downward spiral for white-collar workers. It starts with software. Because AI makes it so easy to create software on the cheap, big software companies might not need as many employees. Or might not need to exist at all. At the same time, the report suggests that consumers will also start using AI agents to complete tasks. Like for online shopping. As in, you'll tell your AI agent to go buy shoes for you rather than searching around for the cheapest price.

7:37David Uberti:They said these agents will just sort of exist in the background of all of the apps that you or I use, and they will basically make decisions either autonomously or semi-autonomously to reduce the friction of something like ordering food to your house to make sure that you get the lowest possible price at all times.

7:57Ryan Knutson:As AI agents take over more aspects of commerce, they'll likely look for efficiencies that could lead to more job losses. For instance, one of the decisions AI agents might make is to avoid credit cards and pay with something that doesn't have fees, like cryptocurrency. That could devastate the credit card companies, leading to more layoffs.

8:18David Uberti:And then their hypothesis is that it will actually snowball into the broader economy because if you have mass layoffs among white-collar workers, those folks tend to make a lot of money. There will be, in turn, a huge decline in U.S. overall consumption in the economy. And then all sorts of financial institutions, whether they are private lenders, mainstream banks, payment processors, mortgage lenders, all of the financial firms that sort of exist around that space, those firms will also suffer as well. So they raise the question that this could also sort of lead to a financial contagion of sorts, in addition to an economic malaise.

9:02Ryan Knutson:A key point of the report is that AI will create something the authors coin as ghost GDP.

9:08David Uberti:Basically, the ghost GDP idea is that all of this innovation will help create wealth in the form of new GDP, new stuff, new value that we create in the economy. but it won't actually go back into the quote-unquote real economy in any meaningful way.

9:26Ryan Knutson:Instead, the report suggests that the wealth created by AI will only benefit a small number of people.

9:32David Uberti:Basically, all of this wealth will accrue to people or companies who win this sort of winner-take-all situation, and the benefits of that, the financial benefits of that, won't be filtered elsewhere.

9:46Ryan Knutson:This doomsday vision of the future helped fuel Monday's huge sell-off in the stock market for all sorts of companies that were either named directly in the post or that have business models that AI agents could replace.

9:58David Uberti:You have DoorDash, which is your food delivery app. You have Visa and MasterCard, which are some of those payment processors that take a cut of each transaction. Salesforce, Zscaler, CrowdStrike, which are all different types of software firms. And then also some managers of private assets. Blackstone, KKR, Apollo, Blue Owl. These are sort of integral parts of the financial plumbing for the tech and software worlds. And this piece really sort of crystallized some of the fears people have around those firms that have already existed. It's also worth saying that a lot of these names have bounced back, at least to some extent since then, which I think sort of speaks to how crazy this market is and how no one really understands how to price this.

10:45Ryan Knutson:So how concerned should we really be about the future laid out in the Citrini report? That's next.

11:02Ryan Knutson:After Citrini's hypothetical vision of the future went viral, critiques of it started going viral too. Plenty of smart people have pushed back poking holes in this thesis. Most notably, there is no mention of purchasing power here.

11:17David Uberti:We have to think about what normal people can do and how their lives are going to get better. Software engineering postings are still rising. Data center construction is booming. Capital spending is accelerating. The apocalypse, they argue, is simply not in the numbers. The math doesn't matter.

11:32Ryan Knutson:David says the criticism falls into two broad camps. The first has to do with that ghost GDP concept. The idea that the value created by AI won't go back into the real economy.

11:43David Uberti:It goes against a lot of what mainstream economists think and how they think about GDP. A normal economist would tell you if you create more money through additional productivity, through something like AI, that money has to go somewhere. And also it overlooks the idea of political economy at all. The idea that the U.S. government and public policy would slow walk into this without any meaningful changes on how to tax people, for example. If you have all of these incredible benefits accruing to a tiny number of companies or individuals in ways that really fundamentally alter the economy in a bad way, it seems unlikely that the government would just let that happen.

12:25Ryan Knutson:The second big camp of criticism is that the piece focuses too much on the negatives of AI without getting into the possible upsides.