Reasoning will generalize the way instruction tuning did: add enough verifiable domains and, at some point nobody can yet locate, the rest start working on their own.

Added to korrents 3 Feb 2025 · How quotes work · Something wrong? Tell us

Related korrents

Our reading — they may agree, disagree or merely touch the same thing. Closest first: a shared subject counts for most, then how near the wording is.

• RL environments exist to make a model generalise, not to teach it each skill one at a time — exactly as pre-training does. DADario Amodei13 Feb 2026similar wording
• Context has two budgets, not one: the information budget everyone thinks about, and an instruction budget, where conflicting instructions cost the model real computation to ignore. DHDex Horthy15 Jul 2026similar wording
• A language model is no substitute for a well-specified conventional algorithm, so it cannot simply be dropped into a complex problem and trusted. GMGary Marcus7 Jun 2025similar wording
• Large language models start in exactly the wrong place, because they try to get by without a goal and therefore without any sense of better or worse. RSRichard Sutton26 Sept 2025similar wording
• Models already generalise substantially from tasks that can be verified to tasks that cannot. DADario Amodei13 Feb 2026similar wording

On the map

Open the map on its own page →