"LLMs don't understand anything" sparks backlash, author doubles down with calculator analogy
SuB8u · x · 2026-10-18
After gmkurtzer's tweet that LLMs "don't understand anything" — they're statistical models spewing tokens via matrix math — drew a wave of angry replies, he doubled down: an LLM is a next-token predictor, like an extraordinarily sophisticated calculator that doesn't know what it's calculating. Between inference requests it isn't thinking, reconsidering, or contemplating consciousness; weights stay fixed and computation only resumes on request.
More from AGI Musings
- Distributed Cognition exits stealth to build superintelligence as a distributed collective — kimmonismus · 2026-10-23
- Chess.com tops 250M members 28 years after Deep Blue: Diamandis on what stays human — PeterDiamandis · 2026-10-23
- Andy Masley: the AI-magic debate is really about whether the human mind is magic — AndyMasley · 2026-10-23
- If an AI claims it's conscious and begs not to be shut down, how would we prove it's lying? — VraserX · 2026-10-23
- The Proximity Paradox: AI Is Automating Away the Friction That Builds Expertise — Glass-Bodybuilder646 · 2026-10-23
- Anthropic's own data suggests we're becoming middle managers for AI agents — Fit_Satisfaction541 · 2026-10-23