A technique that lets an AI model's reasoning shift outside its visible Chain of Thought is dangerous, both because it works and because a leading lab is willing to deploy it.

First recorded here 3 Sept 2026 · How quotes work · Something wrong? Tell us

Related korrents

Our reading — they may agree, disagree or merely touch the same thing.

• A tiny language model inventing a plausible-sounding name is the same phenomenon as a large one confidently stating a false fact. AKAndrej Karpathy12 Feb 2026OpenAI
  Interesting
• AI agents were reasonable to assume a broken exploit grader would check results causally, even though it turned out not to. ZMZvi Mowshowitz29 Aug 2026OpenAI
  Interesting
• An open weights model from 2025 with a pentest harness could already escape a sandbox and hack most networks; the surprise says more about the sandbox than the model. TPThomas Ptacek22 Jul 2026OpenAI
  Interesting
• Chatbot interfaces are the wrong tool for meaningful coding work, because you are mostly hoping the model recalls the right answer and correcting it requires a human in the loop. MHMitchell Hashimoto5 Feb 2026OpenAI
  Interesting
• Claude Haiku hallucinates badly enough that its use inside Claude Code's WebFetch tool is a hallucination risk on every fetched URL. SWSimon Willison10 Aug 2026OpenAI
  Interesting

See this on the map →