Notes from the Lighthaven Daniel Kokotajlo & Dean Ball debate

timely thoughts on regulation, centralization, & AI risk

event was on 3/9/26, ran by anti-debate! edited with opus 4.6. i did my best to accurately represent the arguments — let me know of any misrepresentations! TLDR and summary first; my takes at the end. Debate Question: Given advanced AI, is it time for government to step in and guide development, or should we leave it to the market? TLDR: Kokotajlo thinks regulation is necessary because concentration of power (both corporate and governmental) and loss-of-control risks are too severe to leave to the market. Dean thinks government intervention will likely make things worse — the executive…

related reading

• Notes from the Lighthaven Daniel Kokotajlo & Dean Ball debatejasminexli.substack.com
• The case for ensuring that powerful AIs are controlled — LessWronglesswrong.com
• The Case Against AI Control Research — LessWronglesswrong.com
• Coasean Bargaining at Scaleblog.cosmos-institute.org
• Radical Optionality — Governing Transformative AI Under Uncertaintyradical-optionality.ai
• Ex-OpenAI Researcher: Superhuman AI is 0–4 Years Awayyoutube.com
• The current bottleneck is political will, not research — LessWronglesswrong.com
• Executive Summary - AI Now Instituteainowinstitute.org
• Introducing Plan A - by Scott Alexander - Astral Codex Tenastralcodexten.com
• Can You Nationalize a Frontier AI Lab? - by John Allardjhallard.substack.com
• Kernel | “Democracy-Affirming” AI Could Make Things Even Worsekernelmag.io
• "Be It Enacted" - by Dean W. Ball - Hyperdimensionalhyperdimensional.co