Technology CEOs seem to suffer from AI psychosis

There’s a certain amount of brutality in the tech industry these days, both mimicking previous eras of big change, like cloud computing (runaway costs in the early days), and unlike anything we’ve never seen before (record revenues accompanied by mass layoffs).

There’s a widespread theory that attempts to explain this phenomenon: Technology executives, especially CEOs, are collectively suffering from delusions of grandeur thanks to artificial intelligence. At least one CEO has said it out loud: Box founder Aaron Levy.

“CEOs are uniquely vulnerable to AI psychosis because they are far enough away from the last mile of the business that they still have to accomplish to generate the most value using AI,” Levy said. Written on X.

“CEOs play with AI, develop a prototype, or enter into a contract to use Levy’s examples, and then jump to believing that agents can do the work.

But these high-level executives are not the people who have to review code, catch bugs, and identify calls to hallucinatory libraries before software is published. They’re not responsible for training AI models on a company’s contract terms, nor do they have to spend days combing through contracts to find deceptive terms, Levy points out.

In other words, Levy’s theory posits that CEOs don’t understand operations well enough to know what can and cannot be automated. But this lack of knowledge does not prevent them from acting on their beliefs.

It’s important to note that Levy is not an AI hater. Quite the opposite. He mostly spreads AI positivity on X to his 2.7 million followers, and writes blogs titled, “Headless software is the future” On how designing software for AI agents is the way forward. He also puts his money where his mouth is, backing AI startups as an active angel investor.

So what should CEOs do instead? Levy advises CEOs to use AI often to learn what it can and can’t do, “and come out the other side with an appreciation for both the upside and the real work.”

CEOs are uniquely vulnerable to AI psychosis because they are far enough away from the last mile of business that they still have to accomplish to generate the most value with AI.

So when they play with AI, they see the results of the happy path, and they often don’t take into account the next 10 or 20 things that… https://t.co/ne5mvJ4Rgx

-Aaron Levy (@Levy) May 24, 2026

I have enough faith in humanity to believe that there are CEOs out there trying to do this, but right now, they seem to be in the minority.

In just the first five months of 2026, the tech industry has already seen nearly as many layoffs as it did in all of 2025: 115,430 people have been laid off from 152 tech companies so far in 2026, compared to 124,636 people laid off from 275 companies in 2025, according to an industry layoff tracker. Layoffs. FYI.

The majority of companies have cited artificial intelligence as a reason for these job cuts. Many argue that the largest technology companies It is washed by artificial intelligenceor taking credit for past or future AI productivity gains, when other business decisions and metrics are actually driving the reductions.

However, some of these stories are surprising. Zeb Evans, CEO of project management and productivity software startup ClickUp, He announced proudly On

Evans swore this was not done to cut costs. Instead, he wants a workforce made up of people who manage AI clients and spend their days quickly reviewing the clients’ work. He believes this will create a “100x enterprise,” as he calls it.

While AI can be an extremely useful tool, data on AI and productivity do not support such assumptions. In miles.

A meta-analysis of other research published in October at the University of California, Berkeley California Department Review It found “no strong relationship between AI adoption and increased overall productivity.”

Research published in March by The National Bureau of Economic Research concluded Adopting AI has improved productivity, but pointed to a “productivity paradox, where perceived productivity gains are greater than measured productivity gains.”

After creating thousands of agents to work on tasks, Researchers at MIT He concluded that agents are not yet doing human-quality work in many cases. They predict that at the current rate of LLM improvement, models will be “able to complete most text-related tasks with success rates of, on average, between 80% and 95% by 2029 at a minimum sufficient level of quality.”

In other words, AI is on track to achieve basic proficiency in most tasks within about three years. These researchers believe that the agents will need a few more years to outperform humans.

Meanwhile, research published in Harvard Business Review He showed that when everyone uses AI to produce more stuff, the bottleneck simply shifts to the executives. Their work is waiting for the people who must license all the things everyone produces. If everyone is empowered to act, then than what OpenAI saw last yearWe can say that things may get out of control.

Are CEOs ready for this? If not, the most certain outcome of a CEO’s ongoing AI psychosis will simply be organizational chaos.