Palantir CTO Shyam Sankar: Some AI Researchers Start Smoking to Prove Their Belief in AI Curing Cancer
Palantir CTO Shyam Sankar stated on CBS's "Face the Nation" that some AI researchers have started smoking to demonstrate their belief in AI to their peers.
He mentioned, "Many AI researchers have started smoking as a way to show other researchers how much they believe in AI," explaining the logic as, "I believe so much that AI will cure cancer that I'm betting everything on it." Sankar did not name specific individuals or provide evidence.
He described this as part of a "collective madness" and "social delusion" in Silicon Valley, criticizing some practitioners for viewing AGI as a "machine god," citing Pascal's notion of the "God-shaped hole" in the human heart, stating that some people use AGI to fill this void.
Robinhood CEO Vlad Tenev later posted, "Hearing more and more about AI researchers starting to smoke... superintelligence is near." Previously, tech founders like Avi Schiffmann had stated that by the time AI robots should cure lung cancer, it would be too late for those who develop it.
In the interview, Sankar also criticized the AI safety views influenced by effective altruism, believing that most of the approximately 2,000 AI researchers in the U.S. hold this worldview. He emphasized that the industry should prove that AI can enhance national prosperity and security while incorporating safety from the design phase.
Market Mechanism: Sankar publicly linked the smoking behavior of AI researchers to a quasi-religious belief in AGI, driven by the high expectations in Silicon Valley for the timeline of superintelligence coexisting with a culture of health optimization; attention shifts from model capabilities and safety debates to personal behavioral signals, benefiting narrators with extreme optimistic positions (gaining visibility) while pressuring those lacking validated belief expressions (facing the pressure of being described as delusions).
ABAB AI Insight
As the CTO of Palantir, Shyam Sankar has long criticized the influence of effective altruism on AI safety discussions. In this CBS interview, he uses the example of AI researchers smoking to illustrate how some individuals' belief in AI curing cancer has become so strong that they use health-risk behavior as a signal, placing it within the frameworks of "collective madness" and "machine god."
In terms of capital and narrative pathways, he transforms personal behavior (smoking) into a public bet on the AGI timeline, aiming to highlight Silicon Valley's quasi-religious expectations of superintelligence while pushing the industry to focus on actual capabilities and national security applications rather than abstract doomsday or omnipotent narratives. This aligns with Palantir's business positioning in government and defense data platforms.
Analogously, this resembles extreme personal signaling behaviors that emerged during historical peaks of technological optimism or cases where specific industries showcased health risks as belief demonstrations. Currently, we are in a phase where AI is transitioning from a capability race to deployment and regulation, amplifying the gap between belief intensity and verifiable progress.
Essentially, this reflects a cognitive reconstruction under regulatory and narrative changes: when AGI is viewed by some as an imminent omnipotent tool, health-risk behaviors are used as belief signals, with the mechanism being to discount future technological commitments into present behaviors, prioritizing faith expression over current evidence, thus shifting the industry discussion focus from actual progress to extreme expectations.
ABAB News · Cognitive Law
- Belief strong enough to use health risks as signals, superintelligence narrative is near.
- Machine god fills the void, smoking becomes an all-in bet.
- When collective delusion occurs, actual progress is obscured.