Tech CEOs and the AI Psychosis: A Wild Ride in the Valley of Uncertainty

There's an undeniable wildness in today’s tech industry that mirrors previous eras of dramatic change and innovation but also presents unprecedented challenges. This volatility is both awe-inspiring and unsettling, with record revenues coexisting alongside massive layoffs. A theory circulating suggests that the root cause might be a shared delusion among CEOs known as 'AI psychosis.'

Understanding AI Psychosis

According to Aaron Levie, founder of Box, top-level executives are uniquely susceptible to this condition because they're often far removed from the nitty-gritty work that truly generates value with artificial intelligence. In his own words on X, he wrote: “CEOs are uniquely prone to AI psychosis because they’re sufficiently distant from the last mile of work that still has to happen to generate most value with AI.”

Levie explains how CEOs might develop a prototype or see immediate results from an AI tool and leap to believe it can fully automate tasks. However, these top executives aren't responsible for reviewing code, training models, or combing through contracts—tasks that require a deep understanding of the underlying processes.

Aaron Levie's Perspective

Levie is not an AI skeptic; quite the opposite. He advocates for headless software and has invested in numerous AI startups. However, he warns that CEOs must confront both the potential benefits and the reality of AI limitations. He suggests using AI extensively to gain a deeper appreciation for its capabilities and shortcomings.

The Reality Check

Despite these grand visions, data from UC Berkeley's California Management Review and other research cast doubt on the productivity gains claimed by many tech companies. A meta-analysis found 'no robust relationship between AI adoption and aggregate productivity gain.' Similarly, a study by the National Bureau of Economic Research noted that while AI can improve productivity, there’s often a discrepancy between perceived and actual gains.

Moreover, research from MIT indicates that current AI models will likely reach basic competence in most tasks by 2029. Even so, these tools are not expected to outperform humans for several more years. This reality poses significant challenges, as executives face the bottleneck of authorizing all this new output.

Organizational Chaos?

The current landscape is far from reassuring. In just five months of 2026, nearly 115,430 people have been laid off across 152 tech companies—only slightly fewer than the 124,636 layoffs in all of 2025 (according to Layoffs.fyi). Many point to AI as a key factor driving these cuts. However, some CEOs like Zeb Evans from ClickUp are boldly laying off employees to make room for AI agents, arguing this will create a '100x org.' Yet the data suggests that AI's impact on productivity is far more nuanced than many executives realize.

Conclusion

The ongoing CEO AI psychosis may indeed lead to organizational chaos unless leaders take a more grounded approach. The key lies in understanding both the potential and the limitations of AI, ensuring they don't become victims of their own optimism. As we move forward, it’s crucial for tech executives to stay humble and realistic.