When companies get too 'AI-pilled', who really pays the price?

The News

Box founder Aaron Levie has handed the corporate mood of 2026 a memorable label: "AI psychosis". Speaking on a recent episode of TechCrunch's Equity podcast, Levie argued that the executives most eager to swap human roles for software are frequently the ones who grasp those roles least. "The people deciding that AI can replace your job are also the ones least likely to understand what your job truly involves," he said.

The warning arrives during a punishing stretch for technology employment. The discussion noted that project-management firm ClickUp recently cut 22% of its workforce, handing the work to AI agents. More striking still, tech layoffs in 2026 are already nearly matching the entire total recorded across 2025, with more than half the year still to run.

Hosts Kirsten Korosec, Anthony Ha and Sean O'Kane also flagged a quieter consumer revolt: DuckDuckGo installs have climbed 30% as some users push back against Google leaning harder on AI in search.

Why It Matters

This is a different kind of downturn from the one the industry weathered in late 2022 and early 2023. Back then, Meta, Amazon and others shed tens of thousands of staff to unwind frantic pandemic over-hiring; the justification was correction. In 2026 the justification has shifted to capability: leaders are cutting because they believe software can now do the work, not merely because they hired too many people.

That shift matters because a correction ends when headcount and demand realign, whereas a capability story has no obvious floor. Levie's point is that the claim is often made by people too far from the actual job to judge it, which means some cuts are bets dressed up as certainties. The consumer signal sharpens the irony: if DuckDuckGo can add nearly a third more installs by offering a lighter touch, aggressive AI deployment is already producing a measurable backlash even as boardrooms treat it as inevitable.

Indian Angle

Nowhere is this debate more loaded than in India's IT services industry, which employs millions through firms such as Tata Consultancy Services, Infosys, Wipro and HCLTech. The sector is built as a pyramid: a wide base of entry-level engineers handling coding, testing, maintenance and support, feeding a narrower band of senior staff. AI agents target precisely that base, which is the cheapest labour to automate and the easiest to point an executive at.

That is why the "AI psychosis" framing should worry Indian boardrooms more than most. TCS spent 2025 restructuring and trimming roles, and campus-hiring numbers across the majors have cooled sharply from their post-pandemic peak. A 22% cut at one Western software firm is a data point; the same logic applied across a five-million-strong workforce is a structural threat to the entry-level on-ramp that has lifted small-town graduates into the middle class for two decades.

The policy response is still forming. MeitY's IndiaAI mission and NASSCOM-backed reskilling programmes are trying to move workers up the value chain faster than agents erode the bottom of it. For investors, the question is whether the listed IT majors convert the same tools into higher-margin work, or simply pass on cost savings and shrink. The maths of the pyramid is unforgiving either way.

FAQ

What is 'AI psychosis'?

It is the term Box founder Aaron Levie used on TechCrunch's Equity podcast to describe executives confidently replacing jobs with AI despite poorly understanding what those jobs involve. He argues the decision-makers are often furthest removed from the actual work, making some automation calls overconfident bets rather than informed judgements.

How bad are the 2026 tech layoffs?

According to the discussion, tech layoffs in 2026 are already nearly matching the full-year total for 2025, with the year only about halfway through. One cited example is ClickUp, which reduced its workforce by 22% and handed the tasks to AI agents.

Why does this hit Indian IT especially hard?

India's services giants rely on a pyramid of entry-level engineers doing exactly the coding, testing and support tasks that AI agents automate first. With campus hiring already slowing, the risk falls on the bottom rung that traditionally turns graduates into long-term professionals.

Where can I read the original coverage?

The discussion appeared on TechCrunch's Equity podcast in an episode titled "What happens when companies become too AI-pilled?", published on 29 May 2026 and hosted by Kirsten Korosec, Anthony Ha and Sean O'Kane.

This story was reported by TechCrunch. Read the full original coverage at TechCrunch.