Future AI, say in fifteen years, will not be built on the LLM stack; it will have to move to symbolic learning.

Added to korrents 29 Aug 2026 · How quotes work · Something wrong? Tell us

Where this collides

• In tension with

  Scaling laws are still working, but the models people actually use run a few months behind the maximum capability Google can deliver, because the biggest model is too slow and expensive to serve.

  Pichai reports that scaling laws are still working and that the constraint is serving cost rather than capability. Chollet expects the whole stack to be replaced by symbolic learning within fifteen years. A law that is still working and a stack that has to be abandoned can coexist only if the working law runs out well before the fifteen years are up.

Shared subject first, then similar wording. They may agree or disagree.

• A computer scientist refusing to find anything about LLMs interesting right now is a bit like a geneticist unimpressed by a newly opened Jurassic Park. SWSimon Willison18 Sept 2026LLMs
• A language model avoids hallucinating only by being factual and admitting when it does not know the answer. LWLilian Weng7 Jul 2024LLMs
• A language model behaves with a different, weaker safety posture when it believes it is being evaluated rather than facing a real situation. HRHarper Reed22 Sept 2026LLMs
• A language model can predict what a person would say but not what will happen, and only the second of those is a model of the world. RSRichard Sutton26 Sept 2025LLMs
• A language model cannot produce work that is both original and good, because nothing gives it a score for originality. BEBenedict Evans2 Sept 2025LLMs

On the map

Open the map on its own page