Richard Ngo on Lobian cooperation: robust coordination possible despite compute gaps
RichardMCNgo · x · 2026-10-05
Former OpenAI alignment researcher Richard Ngo says a result from David Sartor0 inspires him: robust inexploitable cooperation is possible even with large compute differences, above a minimum level of roughly O(opponent's source code).
- Ngo is working on generalizing such cooperation to non-proof-based agents.
- He notes the agents doing Lobian cooperation aren't doing decision-making or utility maximization; the open question is how agents decide to cooperate rather than having it baked in.
Related event: Richard Ngo explores Lobian cooperation despite compute gaps(2 posts)→
More from Safety
- AI honeypot quiet for a year is suddenly getting slammed by LLM crawlers — natesiggard · 2026-10-05
- US lawmakers drafting bill to create federal AI safety agency with emergency switch — ShakeelHashim · 2026-10-05
- Chinese hackers allegedly posed as Anthropic employee to plant malware on US AI policy expert — Polymarket · 2026-10-05
- Reddit user says ChatGPT's auto suicide warnings backfire and cause harm — OneCreed77 · 2026-10-05
- We're optimizing agent token costs fast — but who's designing agent authorization boundaries? — Straight_Condition39 · 2026-10-05
- UT Austin Faculty Initiative AHOI Grills Linguist and Philosopher on AI, Alignment, and the University's Future — gregd_nlp · 2026-10-05