About this episode

The episode discusses the implications of superintelligent AI and the potential for international agreements to manage its risks.

Read the full transcript here. What are the world’s leading AI companies actually trying to build when they talk about superintelligence? Is the goal merely better chatbots, or systems that could outperform all humans across every cognitive task? Why would such a system be so alluring if it could accelerate medicine, science, education, abundance, and human flourishing? Why would it also create an unprecedented concentration of power for whoever controlled it? If intelligence includes not only abstract reasoning but persuasion, strategy, manipulation, planning, and technological invention, what happens when those capacities are automated at superhuman scale? How seriously should we take AI CEOs when they say the technology could go catastrophically wrong, and how should we interpret the tension between their public concern and their continued participation in the race? If we cannot reliably inspect their goals, motives, reasoning, or learned objectives, how could we know whether apparent obedience is real safety or just surface behavior? Even if alignment were solved, who should be trusted to steer a superintelligence? Could compute governance, chip tracking, training thresholds…

More episodes of Clearer Thinking with Spencer Greenberg

• The Hidden History of Evidence-Based Everything (with Helen Pearson) · June 12, 2026 · 1h 17m
• Is cash a better form of charitable aid? (with Nick Allardice) · June 5, 2026 · 1h 16m
• When painful thoughts feel true but aren't (with Christine Padesky) · June 1, 2026 · 1h 20m
• Is patriarchy gone or hiding in plain sight? (with Kate Manne) · May 13, 2026 · 1h 38m

Explore listener stats, chart rankings, contacts and more on the Clearer Thinking with Spencer Greenberg podcast page.