Ex-OpenAI & DeepMind Scientist Geoffrey Irving Details Plan to Align Superintelligence
dhadfieldmenell · x · 2026-08-12
Geoffrey Irving (formerly at OpenAI and DeepMind, and ex-Chief Scientist of AISecurityInst) recently appeared on a podcast to discuss his plan for aligning superintelligence (ASI) through his new venture, @resolutionorg.
Key topics from the interview include:
- Misalignment Risks: What misaligned ASI will actually look like.
- Industry Optimism: Why current AI labs are more optimistic than he is.
- Slowing Down: When and how to appropriately slow down AI progress.
- Post-ASI Tech: The technological landscape after the arrival of ASI.
- Character Training: Whether good character training can carry over to superintelligence.
- Past Predictions: Reflecting on his accurate 2018 prediction that RL + LLMs would lead to AGI.
More from AGI Musings
- There's no formal route to frontier AI research: 19-year-olds are advancing tech from their basements — repligate · 2026-10-16
- 'Will you end humanity?' AI's candid answer becomes a clown meme — dermflork · 2026-10-16
- Stanford professor: AI-done homework scores high, students fail in-person finals — FinanceYF5 · 2026-10-16
- Data brokers pay $100K–$2M for SaaS datasets that may train their own replacements — FinanceYF5 · 2026-10-16
- tszzl: Pretraining science is like steam engines before thermodynamics, slouching toward interpretability — tszzl · 2026-10-16
- eigenrobot: AI acceleration will collapse human capital investment — eigenrobot · 2026-10-16