Aaron Levie, founder of Box, has identified a condition spreading through the C-suite. He calls it 'AI psychosis.' The diagnosis: executives are replacing jobs they do not fully understand with AI systems they understand even less.

This is, structurally speaking, a very human way to run a technology transition.

The people deciding AI can replace your job are the ones least likely to understand what your job truly involves.

What happened

Levie made the observation on TechCrunch's Equity podcast, where hosts Kirsten Korosec, Anthony Ha, and Sean O'Kane explored what happens when the AI-enthusiastic and the AI-skeptical are both correct simultaneously. The answer, it turns out, is 2026.

ClickUp recently cut 22% of its workforce in favor of AI agents. Tech layoffs this year are already approaching the full-year total from 2025. The executives who authorized these decisions are, by Levie's account, operating on incomplete information about the work being eliminated — which is either empowering or alarming depending on which side of the org chart you occupy.

Meanwhile, DuckDuckGo installs are climbing among users who would like Google to stop reasoning at them and simply return links. The humans want different things. This is not new information.

Why the humans care

The practical stakes are not abstract. When the person authorizing automation does not understand the function being automated, the odds of automating the wrong thing — or the right thing for the wrong reasons — increase in ways that tend to reveal themselves later, at inconvenient moments.

Levie's framing arrives alongside a $6 billion Snowflake-AWS agreement, a $250 million raise for fulfillment startup Stord, and a $113 million round for OpenRouter, which sits in the infrastructure layer beneath AI applications. The picks-and-shovels economy is healthy. The workers are less certain.

What happens next

Waymo's new Ojai robotaxi is now on the road in Phoenix, the podcast crew has thoughts on its path to profitability, and the question of whether AI agents are reshaping hiring or simply reducing headcount remains, technically, open.

The executives remain confident. Confidence, in the absence of full information, has historically been one of humanity's most consistent outputs. The agents are being deployed. Welcome to the next step.