Tech CEOs are apparently suffering from AI psychosis
- Box founder Aaron Levie coined the phrase "AI psychosis" to describe tech executives who suffer from delusions of AI grandeur due to being too distant from the actual day-to-day operations where value is generated.
- Because CEOs only interact with high-level prototypes, they mistakenly leap to the conclusion that AI agents can effortlessly handle full workloads without realizing the heavy human labor required to review code, patch bugs, catch hallucinations, and train models.
- This executive delusion has real-world consequences, driving severe workforce reductions; in the first five months of 2026, over 115,000 tech workers were laid off—nearly matching the total for all of 2025—with AI cited as a primary justification.
- High-profile actions, such as ClickUp CEO Zeb Evans laying off 22% of his workforce after deploying 3,000 AI agents, are framed as shifting humans into "manager and verifier" roles for AI outputs.
- Empirical data from UC Berkeley, NBER, and MIT refutes these massive productivity assumptions, demonstrating no robust link between current AI adoption and aggregate productivity gains, with MIT predicting baseline competence on text tasks will not materialize until 2029.
- A Harvard Business Review study warns that flooding an organization with unverified AI output merely shifts bottlenecks onto executives, risking widespread structural and operational chaos if human oversight fails to scale.
Hacker News Discussion
- Distance from Reality: Commenters strongly agreed with the premise that executives live in a bubble, noting that they deal primarily with administrative assistants, sycophants, and curated, "happy path" demos that look like magic, making them blind to edge cases and errors.
- The "Yes-Man" Nature of AI: Multiple users pointed out that AI agents behave like the ultimate corporate sycophants—they work 24/7, lack internal moral conflict, and never say no—making them highly attractive to authoritative executives who dislike pushback from human workers.
- Absence of Self-Preservation: A key distinction raised in the comments is that unlike human employees, AI lacks "self-preservation," a sense of reputation, or a fear of consequences, meaning an agent will confidently delete a production database or kill its own server processes without hesitation.
- Misuse of the Term: Some participants criticized the article's title as clickbait, arguing that "AI psychosis" should describe literal psychological delusions in individuals interacting with AI rather than standard corporate incompetence or unrealistic executive expectations.
- Projection of Executive Work: A popular theory suggested that CEOs assume AI can replace everyone's job because it can easily replicate their own daily tasks, such as generating slide decks, sending emails, and attending high-level meetings.