'The Illusion of Thinking' Paper Rekindles Debate: Do LLMs Actually Reason?
CGMthrowaway · hn · 2026-10-30
A Hacker News thread revisits the arXiv paper The Illusion of Thinking, which argues that large language models don't genuinely reason — their apparent reasoning is closer to pattern matching. The paper, known for controlled puzzle experiments showing accuracy collapsing past a complexity threshold, once again sparked debate over whether LLMs truly reason.
More from AGI Musings
- RL 大佬 Szepesvári 力挺学术界：你的研究是托举产业的水位 — CsabaSzepesvari · 2026-11-01
- Sam Altman: Using latest AI tools feels like managing a factory inside a computer — haider1 · 2026-11-01
- Beyond "pets or uploads": a case for a third path where humans are empowered but not augmented — luke_drago_ · 2026-11-01
- "We're not plateauing": KOL points to strong unreleased internal models at every frontier lab — kimmonismus · 2026-11-01
- Erik Voorhees slams Anthropic's Claude self-assessment: machine chats create no victim — juanbenet · 2026-11-01
- a16z: AI jobs panic overblown — engineering postings up 45% vs pre-pandemic — a16z · 2026-11-01