Palantir Chief Technology Officer Shyam Sankar has raised questions about the growing belief in artificial general intelligence (AGI) among Silicon Valley’s technology workers, claiming that some researchers are taking up smoking because they expect future AI systems to solve diseases such as cancer.
Speaking to CBS News, Sankar described what he sees as a culture of excessive faith in AI, calling it “mass psychosis.” He argued that some people developing advanced AI systems have become so convinced of the technology’s future capabilities that they are willing to take health risks today, believing future breakthroughs will protect them tomorrow.
Sankar’s comments have drawn attention to a broader debate in the technology industry: whether confidence in AI’s potential to transform medicine and extend human life is encouraging some people to underestimate present-day risks.
However, his claim about researchers smoking specifically because they believe AGI will cure cancer remains unverified in the information available. It is also important to distinguish his characterisation of Silicon Valley culture from evidence about how widespread the behaviour actually is.
What did Shyam Sankar say about AI researchers and smoking?
Sankar, CTO of data analytics and defense technology company Palantir, suggested that some AI researchers use smoking as a signal of their confidence in the technology’s future.
“So many researchers have started smoking as a way of signaling to other researchers how much they believe in AI,” he said in his CBS News interview.
He described the behavior as evidence of a broader “social delusion” surrounding AI development. According to Sankar, some technology workers appear to believe that the medical advances made possible by future AI systems will outweigh the consequences of unhealthy decisions they make today.
“I’m showing you my chips are all in. So there is kind of this mass psychosis and social delusion around it,” he said.
Sankar did not provide evidence establishing how many researchers have adopted the habit, identify specific researchers engaging in it, or demonstrate that belief in AGI was the reason they started smoking.
His comments should therefore be understood as a criticism of a perceived cultural trend, rather than a verified account of the behavior of AI researchers as a group.
Why do some people believe AGI could cure cancer?
Artificial general intelligence refers to a hypothetical AI system capable of performing a broad range of intellectual tasks at a level comparable to humans, or potentially beyond human capabilities, depending on the definition used.
Unlike systems designed for specific tasks, AGI is generally envisioned as being able to transfer knowledge across different fields, reason through unfamiliar problems, and adapt to new challenges.
Supporters of advanced AI argue that such capabilities could accelerate scientific research, including the development of new medicines and treatments.
How AI could change medical research
AI is already being investigated and used in several areas of biomedical research, including:
- Drug discovery: Machine-learning systems can help researchers identify promising drug candidates and analyze molecular interactions.
- Protein science: AI models can predict aspects of protein structure and help scientists investigate biological mechanisms.
- Medical imaging: AI tools can assist in analyzing scans and identifying patterns that may be difficult to detect consistently.
- Clinical research: AI can help researchers analyze large datasets, identify potential trial participants, and explore relationships between biological variables.
- Personalized medicine: Researchers are investigating how AI can help tailor treatment strategies to individual patients and their disease characteristics.
These applications offer reasons to study AI’s potential in medicine. They do not, however, establish that AGI will eliminate cancer or that a universal cure is imminent.
Cancer is not one disease but a broad category of conditions involving different biological mechanisms, mutations, tissues, and responses to treatment. Progress against one cancer type does not automatically translate into a cure for every other type.
AI may help researchers make discoveries more quickly, but those discoveries must still be tested for safety and effectiveness through appropriate laboratory studies, clinical trials, and regulatory review.
The distinction between accelerating research and guaranteeing a cure is central to evaluating the optimism described by Sankar.
Are AI researchers really taking up smoking because of AGI?
The scale of the alleged trend remains unclear.
The supplied report cites coverage by Fortune and Business Insider suggesting that nicotine products and smoking have become more visible in some technology workplaces. It also quotes Avi Schiffmann, the CEO of AI wearable startup Friend, discussing the idea that future AI-enabled medical advances might offset today’s health risks.
“By the time you do get lung cancer, our AI robots are supposed to have cured that,” Schiffmann told Business Insider, according to the supplied report.
The remark illustrates the kind of technological optimism that Sankar is criticizing. It does not establish that smoking is widespread among AI researchers or that the industry has collectively adopted this reasoning.
Several distinctions matter:
- Smoking and nicotine use are not interchangeable. Cigarettes expose users to tobacco smoke, while nicotine pouches deliver nicotine without combustion. Both involve health considerations, but their risks are not identical.
- Anecdotes do not establish an industry-wide trend. Reports about individual workers or particular offices cannot establish how common smoking is across Silicon Valley.
- Motivation must be demonstrated. Even if more workers in a particular company smoke, that alone would not show that they are doing so because they expect AGI to cure cancer.
- Workplace culture varies. Practices at a handful of startups or technology companies should not automatically be attributed to the entire AI research community.
To establish the trend more conclusively, reporting would need interviews with identified researchers, evidence about changes in smoking behavior, and independent information about their motivations.
Why does Palantir’s CTO call it “mass psychosis”?
Sankar’s use of “mass psychosis” is rhetorical language intended to convey what he sees as an unusually strong collective belief in AI’s future capabilities.
It should not be interpreted as a clinical diagnosis of AI researchers.
His criticism centers on an apparent contradiction: some people fear that advanced AI could eventually pose existential risks to humanity, yet may simultaneously believe it will deliver medical breakthroughs powerful enough to neutralize the consequences of present-day risk-taking.
In Sankar’s view, this combination reflects a culture in which expectations about future technology can override conventional judgments about personal responsibility and risk.
The argument raises a legitimate question about technological optimism: when does confidence in future innovation become a reason to discount risks that are well understood today?
However, the smoking example alone cannot answer that question for the industry as a whole. Researchers may have different views about AGI, medicine, personal health, and the risks of advanced AI.
A careful assessment of Sankar’s broader criticism would require examining the specific statements and behaviors he is referring to, rather than treating his characterization as an established description of Silicon Valley.
What is the “God-shaped hole” idea, and how does it relate to AI?
Sankar also invoked the idea of a “God-shaped hole” to explain why some people may place extraordinary faith in artificial intelligence.
The phrase is commonly associated with the broader idea that humans seek meaning, purpose, or something greater than themselves. It is often discussed in religious and philosophical contexts, although its popular use does not establish that all people share the same spiritual needs.
The concept is sometimes associated with the French mathematician and philosopher Blaise Pascal, whose writings explored the relationship between human reason, faith, and the search for meaning.
Sankar’s application of the idea to Silicon Valley suggests that some people may be treating AGI as more than a technological achievement—as a force capable of resolving fundamental human problems.
Under that interpretation, AGI becomes a source of hope for overcoming disease, scarcity, and even mortality.
This is an interpretation of the culture, not proof that AI researchers generally regard AGI as a deity or substitute for religion. Some researchers approach AI primarily as a scientific and engineering challenge, while others have more ambitious views about its potential effects on society.
The distinction is important because belief in AI’s scientific usefulness does not necessarily imply religious devotion or unquestioning faith.
Does the promise of future AI justify health risks today?
The central issue raised by the smoking controversy is whether confidence in future scientific advances can become a justification for ignoring current evidence.
Smoking provides a particularly clear example because its health risks are well established. The World Health Organization identifies tobacco use as a major cause of preventable disease and death. Future advances in medicine cannot be assumed to reverse damage caused by tobacco exposure, nor can anyone reliably predict whether a particular cancer will be preventable or curable when it develops.
Even substantial progress in AI-driven drug discovery would not eliminate the uncertainty involved in developing treatments, testing them, and making them available to patients.
A future AI system might contribute to cancer research in important ways. That possibility is different from knowing that it will produce a cure within a particular person’s lifetime.
There is also a broader lesson for technology forecasting. Predictions about transformative innovations often involve several separate uncertainties:
- Whether the technology will achieve the anticipated capability.
- How long development will take.
- Whether the resulting discovery will work in real-world conditions.
- Whether it will be safe and accessible.
- Whether it will address the specific problem an individual faces.
Confusing a plausible future breakthrough with a guaranteed outcome can lead people to underestimate risks that they would otherwise take seriously.
That concern applies beyond smoking. It can influence decisions about climate technology, financial risk, medical treatment, and other areas where people expect future innovation to solve problems that exist today.
What does this controversy reveal about Silicon Valley’s AI debate?
The disagreement reflects a larger tension in the AI industry between confidence in technological progress and concern about its limits.
Optimists argue that increasingly capable AI systems could accelerate scientific discovery, improve productivity, and help solve difficult problems. Critics warn that predictions about transformative capabilities can run ahead of demonstrated results and encourage unrealistic expectations.
Sankar’s comments focus on the personal consequences of that optimism. Rather than debating technical benchmarks or investment forecasts, he uses smoking to illustrate what he perceives as a willingness to trade present-day health for anticipated future benefits.
But the example also highlights the need for precise reporting. An anecdote can reveal a cultural attitude worth investigating without proving that the attitude is widespread.
The strongest way to assess the claim is to separate the underlying questions: Is nicotine use increasing in particular workplaces? Are workers explicitly connecting risky behavior to expectations about AGI? And how do AI researchers themselves describe their confidence in future medical advances?
Answering those questions would provide a clearer picture than relying on a dramatic label alone.
The bottom line
Shyam Sankar’s remarks offer a pointed criticism of what he sees as excessive faith in artificial general intelligence. His claim that some researchers have started smoking because they expect AI to cure cancer has captured attention because it connects an ambitious technological promise with an immediate, familiar health risk.
Yet the available information does not establish how widespread the behavior is or prove that belief in AGI is driving it.
AI is already contributing to scientific research, including efforts to understand diseases and develop treatments. Whether future systems will deliver cures for particular cancers—and when they might do so—remains uncertain.
The distinction between scientific possibility and guaranteed progress is the heart of the debate. Future breakthroughs may transform medicine, but they cannot be treated as insurance against risks people take today.
