A Former AI Doomer Reveals a Dark Truth

Summary

Cal Newport interviews researcher Emile Torres about the ideologies Torres says have influenced Silicon Valley and the race to build advanced AI. Torres’s central argument is that the debate is not simply between people who want AI and people who oppose it: many prominent AI critics and AI developers share a longer-term vision of a posthuman future, but disagree about how quickly and safely to pursue it.

Key ideas

• TESCREAL: Torres uses this acronym, developed with Timnit Gebru, for a cluster of ideas: transhumanism, extropianism, singularitarianism, cosmism, rationalism, effective altruism, and longtermism. He describes them as relatively marginal in wider society but influential in parts of Silicon Valley.
• The technological vision: In Torres’s account, these movements share an ambition to use advanced technology—especially artificial superintelligence (ASI)—to transform or replace human limitations, potentially upload minds, expand into space, and create a vast posthuman civilization. “Singularity” commonly refers here to a possible feedback loop in which AI helps develop increasingly capable AI, leading to superintelligence.
• Disagreement within the AI movement: Torres distinguishes between AI “pessimists” or “doomers” and accelerationists. He argues that both camps may support the ultimate goal of creating superintelligence and a posthuman future, while disagreeing over timing, AI safety, and whether systems can be aligned with their developers’ values. In his view, objections to companies such as OpenAI or Anthropic can reflect disputes within this broader ideological camp rather than opposition to its underlying aims.
• Torres’s critique of “humanity” and “extinction”: He says discussions of AI risk can use “humanity” broadly enough to include future posthuman successors. As a result, preventing “humanity’s extinction” may not mean preserving Homo sapiens indefinitely. Torres distinguishes terminal extinction—the disappearance of our species, even if successors remain—from final extinction, in which no posthuman successors remain. He argues that some AI-risk thinking prioritizes avoiding the latter, potentially treating the disappearance of biological humans as acceptable.
• Risks of utopian reasoning: Torres contends that assigning immense value to a hypothetical future civilization can make present-day people seem expendable. He cites extreme hypotheticals involving catastrophic harm to current populations to illustrate what he sees as a danger in expected-value arguments and total utilitarianism. He also criticizes the lack of democratic input and inclusion in defining the future these ideologies seek.
• Why the ideas appeal: Torres describes his former involvement as partly driven by the promise of purpose, immortality, and transcendence—ideas he compares with religious narratives. He also characterizes effective-altruist and rationalist communities as socially cohesive, jargon-rich, and backed by substantial funding. These are his interpretations and recollections, not independently established findings in the interview.
• What drives the AI race: Torres says the race began with ideological ambitions, while market competition and geopolitical rivalry—especially between the United States and China—have become additional pressures. He portrays arguments for continuing development despite safety concerns as often rooted in fear that a rival lab or country will build a less-aligned system first.
• Self-fulfilling prophecy: Torres argues that some figures’ apparent predictions about AI risk are not simply forecasts: in his view, people connected to the relevant movements helped create and accelerate the race they now warn about. He points to historical and institutional connections among AI researchers, conferences, funders, and labs.
• Newport’s takeaway: In his recap, Newport says the interview clarified how some versions of “avoiding human extinction” may entail a transition away from biological humanity. He does not claim that superintelligent AI or the proposed utopia is feasible; he agrees that utopian projects can be dangerous when present lives are treated as insignificant relative to an imagined future. Newport says generative AI could produce useful tools, but is concerned that the technology is being shaped by a well-funded, anti-human utopian philosophy.

Reviews, guides, or tutorials

None. This is an interview and analysis, not a product review or tutorial.

Recommended sources mentioned

Torres recommends his article “Against Longtermism,” described as appearing in Aeon, and “Half a Pro-Human Manifesto.” He also mentions his newsletter, “Technological Apocalypse in Real Time,” and podcast, “Dystopia Now.”

Main speakers and sources

• Cal Newport: Host and closing commentator.
• Emile Torres: Guest and primary source for the ideological critique.
• Timnit Gebru: Mentioned as Torres’s collaborator in developing the TESCREAL framework.