Bayesian updating can't fix a wrong model: Gelman & Shalizi's critique hits AI circles
sebkrier · x · 2026-10-09
Seb Krier observes that many in AI come from a Bayesian tradition and hedge with probabilities, yet rarely question their initial modeling assumptions. He highlights Gelman & Shalizi's paper Philosophy and the practice of Bayesian statistics (arXiv:1006.3868), which argues Bayesian updating only moves belief within the hypotheses you start with—the paper connects successful Bayesian practice to sophisticated hypothetico-deductivism, stressing that model checking and revision fall outside Bayesian confirmation theory, and warns that inductivist views have discouraged practitioners from checking models.
Related event: Bayesian Mindset in AI Criticized for Ignoring Model Assumptions(2 posts)→
More from AGI Musings
- Andrew Critch hits back at "Claude probably isn't conscious": which consciousness, exactly? — AndrewCritchPhD · 2026-10-09
- Safety researcher: OpenAI is uniquely sabotaging regulation while other labs' risks aren't symmetrical — Turn_Trout · 2026-10-09
- Scott Aaronson: the AI skeptics' 'wake me up' warnings have all come true — scottleibrand · 2026-10-09
- Former multimodal researcher: aligning model 'taste' for everyone is unsolvable — A_K_Nain · 2026-10-09
- Traversy Media: newbies can't learn AI coding without writing code first — capeandcode · 2026-10-09
- Enjoying the Centaur Era While It Lasts: OpenAI's Math Results Signal Its End — moyix · 2026-10-09