Plan A: Slowing Down Development for Superintelligence
Don't Worry About the Vase (Zvi) · rss · 2026-07-12
This long post introduces and reviews "Plan A"—a new proposal from the AI 2027 team that advocates slowing down AI development, collaborating with China, monitoring major compute sources, and considering mechanisms like "Mutual Assured Compute Destruction (MAIM)" to avoid dangers from superintelligence.
The author emphasizes that this proposal is not an endorsement or call for immediate implementation, but a serious discussion of its details and counterarguments. The post addresses typical criticisms: some believe superintelligence won't arrive soon, making high costs unnecessary; others worry that controlling compute could lead to authoritarianism; some question technical feasibility, overly optimistic economic predictions, or argue that such scenarios confuse real predictions with visions.
Overall, the article centers the debate on key judgments: whether superintelligence will emerge in the near future, whether it's worth accepting real-world costs to reduce risks, and whether governance/coordination schemes can be designed to be effective without being overly authoritarian in an international competition context.
More from AGI Musings
- François Fleuret: Only Two Long-Term Futures — No Super AI, or Staying Fully Human With It — francoisfleuret · 2026-09-11
- IG reel debunking the 'winning the AI race against China' fallacy hits 500k likes — louisvarge · 2026-09-11
- Post-AI World Leaves No Room for Learning on the Job — rachittshah · 2026-09-11
- Researcher questions AI safety eval firm, citing 'blatantly sloppy' security and monitoring — Kyrannio · 2026-09-11
- AI researcher memes agent-swarm tinkering with He Jiankui's embryo-editing quote — dejavucoder · 2026-09-11
- nabla_theta: happy to be wrong if the AI utopia arrives with little ex ante risk — nabla_theta · 2026-09-11