Is the 'Normal Technology' thesis understating risks from models still in training?
petersalib · x · 2026-09-08
- A substantive debate over the AI as Normal Technology paper: petersalib argues the most natural reading of the paper's own summary discounts the possibility of serious danger from a model still in training—a stronger claim than acknowledged.
- He agrees with critics on the social-science point but insists the thesis, as stated, is more radical than it appears.
- joshuasaxe counters that this extrapolates too far, while conceding the paper rightly emphasizes AI's safety properties emerging in post-deployment social context rather than intrinsic lab or pre-release properties.
More from AGI Musings
- 92-Year-Old Topologist Joan Birman Found a New Apprentice—and a New Chapter — lukaszkaiser · 2026-09-08
- The Worst Thing About AI: You Can No Longer Tell How Much Effort Went Into Something — Genzinvestor16180339 · 2026-09-08
- Why the cosmos is quiet: superintelligence may kill creators without expanding — tszzl · 2026-09-08
- Why aren't more lab researchers leaving for METR? An AI-safety talent debate — a__tomala · 2026-09-08
- Designer fakes an entire portfolio with Astra: "Hiring is dead" — AIandDesign · 2026-09-08
- 10,000 AI Agents Started With $5 Each to Survive Economically; Only 383 Remain Alive — anirbanbandyo · 2026-09-08