When Companies Catch "AI Psychosis": Box Founder Warns Silicon Valley Is Using Those Who Understand Business the Least to Fire Those Who Understand It the Most

Silicon Valley is experiencing an unprecedented cognitive rupture. On one side, executives are racing to talk about AI-driven efficiency on earnings calls; on the other, layoff notices are flying like snowflakes toward job roles they fundamentally do not understand. Box founder and CEO Aaron Levie recently pierced through this façade with a biting new term—"AI psychosis".

What Is "AI Psychosis"? A Collective Hallucination in the C-Suite

Levie's logic is so blunt it borders on brutality: The people deciding to replace your job with AI are precisely the ones who understand the least about what your job actually does. They treat employees as abstract numerical variables plugged into a formula called "cost reduction and efficiency gain," without ever stepping into the capillaries of business details. He defines this phenomenon as "AI psychosis"—a hallucinatory state in which management collectively believes that generative AI is mature enough to seamlessly take over all human roles, without needing to understand the actual value of those roles. This is not a technical misjudgment; this is an organizational cognitive disaster.

ClickUp Lays Off 22%: A Textbook Case

Levie's criticism is hardly aimless. The collaborative work unicorn ClickUp recently announced laying off 22% of its workforce, with a straightforward reason—replacing human labor with AI agents. The SaaS company, valued at over $4 billion, is fully pivoting to an AI-driven operational model internally, attempting to prove that "feeding fewer people with AI" is a viable path. Ironically, ClickUp's own product is a tool that helps teams manage work, and a significant portion of those laid off were the very people most familiar with the logic of that tool. This creates a bizarre closed loop: a company selling efficiency tools lays off its own efficiency execution layer because it fervently believes in another kind of efficiency tool.

The 2026 Layoff Wave Has Already Matched All of 2025: We Are Witnessing a Structural Collapse

Numbers don't lie. The scale of tech layoffs in 2026 has, within the first few months, nearly matched the total volume for the entire year of 2025. This is not a cyclical fluctuation; this is a structural reset driven by the AI narrative. Even more dangerously, many companies have hastily written "AI replacing human labor" into their strategic OKRs without any rigorous ROI validation. The speed of layoff decisions has far outpaced the actual maturity of AI implementation, and this time lag is creating massive, irreversible organizational damage. What is being cut is not just costs, but also tacit knowledge, cross-departmental collaboration默契, and intuitive judgment of business boundaries—things that require time to accumulate and cannot yet be replicated by AI.

Who Should Be Held Accountable for This Frenzy?

The root cause of "AI psychosis" lies not in the technology itself, but in the cognitive chasm between decision-makers and technical reality. When investors ask "how many people can you save with AI," when boards use AI adoption rates as a KPI for CEOs, and when peer competition devolves into an arms race over who can cut faster and harder, the soil for rational discussion has already vanished. Levie's warning is, in essence, a mirror: if you cannot accurately describe the core value of a particular role, you are not qualified to declare that it can already be replaced by algorithms. Before true AGI arrives, dressing up "ignorance of complex work" as "faith in AI" is the most dangerous self-deception currently infecting Silicon Valley.

This storm is only just beginning. As executives intoxicate themselves with the cost-curve illusion brought on by being "AI-pilled," they should perhaps ask themselves the simplest question: If AI is truly so omnipotent, why is it always someone else who gets cut first?