May 27, 2026
Prompts, pink slips, panic
Tech CEOs are apparently suffering from AI psychosis
Workers get cut, bosses get hyped, and the comments are absolutely not buying it
TLDR: Box founder Aaron Levie says many tech bosses get carried away after trying AI and forget how much human work still has to happen behind the scenes. Commenters were split between mocking CEO hype with memes and accusing the article itself of clickbait for using the word "psychosis."
The real spectacle here isn’t just Aaron Levie saying some tech bosses have "AI psychosis" — it’s the internet instantly turning that phrase into a full-blown roast session. Levie’s point was fairly simple: chief executives often test shiny new artificial intelligence tools, watch them do one impressive trick, and then wildly assume they can replace huge chunks of real human work. Meanwhile, the people actually checking mistakes, fixing broken code, and catching made-up answers are left cleaning up the mess. Add in a brutal wave of layoffs — more than 115,000 tech workers cut in the first five months of 2026 — and commenters were in no mood to clap for the hype.
That’s where the drama kicked off. One camp said the article title was pure clickbait, arguing this was really just a warning that bosses should use AI more carefully and learn its limits, not a diagnosis. Another camp didn’t care about the nuance because the mood was too relatable: one commenter joked about "AI vampire" executives sleeping four hours a night for their next prompt-fueled high, while another compared CEO obsession with AI to everyday people spiraling over rent. Ouch. The biggest subtext in the thread? People think executives are acting like AI is magic while workers pay the price. Even ClickUp’s CEO bragging about cutting 22% of staff after deploying 3,000 AI agents landed less like innovation and more like a community-triggering villain origin story.
Key Points
- •The article centers on Aaron Levie’s argument that CEOs often overestimate AI because they are removed from the detailed work needed to make AI outputs usable.
- •key operational tasks such as code review, bug detection, checking hallucinated libraries, and adapting models to company-specific data are presented as limits on full automation.
- •The article reports that 115,430 employees were laid off at 152 tech companies in the first five months of 2026, versus 124,636 layoffs at 275 companies in all of 2025, citing Layoffs.fyi.
- •ClickUp CEO Zeb Evans said the company cut 22% of its workforce after deploying about 3,000 AI agents for internal work.
- •The article cites research from California Management Review, the National Bureau of Economic Research, and MIT indicating that measured AI productivity gains remain limited and agent quality often still falls short of human work.