Csaba Szepesvári: AI Risk Stems From Unknown Unknowns, Not Ignorance of How Models Work
CsabaSzepesvari · x · 2026-09-21
- Reinforcement learning authority and DeepMind research scientist Csaba Szepesvári pushed back on debate about whether we understand how AI models work.
- His take: at some level of abstraction we already know a lot about how models work; the real source of risk is the unpredictability of unknown unknowns—a potential "magic formula" waiting to be discovered.
- Risk exists with or without AI, he argues; AI raises perceived risk only because it seems to boost the chance of finding such a formula (setting aside takeover scenarios).
More from AGI Musings
- Agents' bottleneck isn't intelligence but goal commitment — and that's good for alignment — mimi10v3 · 2026-11-12
- Google Cloud survey: 78% say legacy enterprise vendors may not lead the AI era — rseroter · 2026-11-12
- Safety researchers blast OpenAI/Anthropic plan: automating alignment research with possibly misaligned AI — trevposts · 2026-11-12
- NYT essay: resisting Big Tech and AI takes collective action, not willpower — TuhinChakr · 2026-11-12
- Ethan Mollick: 'talk to your documents' RAG solutions already feel dated — emollick · 2026-11-12
- Anthropic's pitch hasn't changed in 5 years: 'AI is winner-take-all, we will build god' — menhguin · 2026-11-12