Palantir CTO uses an unlikely smoking anecdote to argue against AI hype

Palantir Technologies’ chief technology officer, Shyam Sankar, entered the debate over artificial intelligence with an analogy that quickly drew attention: he said AI researchers have “took up smoking” to prove a hypothesis. The comment was used to argue against what he described as a collective, irrational focus on AI, which he characterized as a form of “mass psychosis.”

The remarks, reported by Yahoo Finance through, were framed as a critique of the broader conversation around AI. Sankar’s underlying message was not about cigarettes or health, but about what he suggested is a willingness to build narratives around a preferred explanation without enough skepticism, verification, or humility about uncertainty.

Palantir, a company known for building software platforms that help organizations integrate data and put it to operational use, has often positioned itself as pragmatic in how AI is deployed. In this instance, Sankar’s point landed on the human side of research and adoption, suggesting that the AI industry can be swept into groupthink as much as any other technology community.

In the reported framing, Sankar’s smoking reference was meant as a provocative example of hypothesis testing, but also as a rhetorical counterpunch to what he sees as distorted hype cycles. “Mass psychosis,” as he used it, implies a widespread social or professional condition in which people can lose their ability to evaluate evidence independently.

The comment also highlights how leading AI executives are increasingly competing over narrative, not only model performance. Beyond benchmarks, companies are trying to shape how outsiders interpret AI risk, readiness, and the credibility of claims about what the technology can do, and how quickly it should be trusted in real environments.

Palantir has not, in the material available here, disclosed additional details tying the remarks to a specific product roadmap, internal study, or customer deployment. The report does not provide follow-up clarification on what “hypothesis” Sankar was referencing, whether the smoking anecdote was historical or illustrative, or what concrete takeaway he expects the AI industry to adopt.

Still, the broader theme is clear enough to track. As AI adoption accelerates across software, government, and enterprise operations, leaders like Sankar are using media moments to argue for caution, deliberate evaluation, and skepticism toward consensus narratives. For Palantir investors and customers, the message reinforces the company’s ongoing emphasis on controlled, application-oriented use of advanced computation rather than purely celebratory forecasts.

What to watch next is whether Palantir or Sankar expands on the comment in a fuller interview or in a company forum that provides specifics, such as how the company evaluates AI claims, what safeguards it emphasizes, and how it separates experimental work from operational deployments.