Box CEO calls out 'AI psychosis' driving Silicon Valley layoffs

In a candid X post, Box CEO Aaron Levie diagnosed a growing Silicon Valley malady: 'AI psychosis.' He argues that corner-office executives are too distant from the 'last mile of work' required to generate real value from AI. 'When they play with AI, they see the happy path results, often not considering the next 10 or 20 things that have to happen to get sustainable results from agents,' Levie wrote. His critique highlights a chasm between the polished demos CEOs see and the messy reality of hallucinations, bugs, and token optimization—or 'tokenmaxxing'—that frontline employees grapple with daily.

Levie's diagnosis aligns with hard data. A 2025 survey from AI firm Rev revealed that heavy AI users experience three times the rate of hallucinations and spend nearly ten times longer refining outputs compared to light users. Despite these inefficiencies, many tech CEOs are forging ahead with mass layoffs, betting that AI can replace thousands of roles. Levie's warning underscores a dangerous feedback loop: executives basing strategic decisions on incomplete, optimistic snapshots of AI performance while ignoring the substantial human effort still required to make AI work reliably in production.