Melanie Mitchell on whether LLMs truly reason or just mimic it
stevenstrogatz · x · 2026-08-22
The latest episode of Quanta Magazine's 'The Joy of Why' features computer scientist Melanie Mitchell discussing the nature of LLMs: when an LLM answers a question, is it actually reasoning like a human, or simply producing text that resembles reasoning? Janna Levin retweeted, suggesting that machine cognition might be better understood as an 'alien' intelligence rather than an emulation of human cognition.
More from AGI Musings
- Public Opposition to Local Data Centers Soars to 75% in One Year — MatthewBerman · 2026-08-23
- Dropping alignment gradients makes CoT tokens less aligned — davidad · 2026-08-23
- "I'm the API between Claude and my manager": the talent layer is shifting — sanderssays · 2026-08-23
- Economist Brynjolfsson: AI Perfectly Imitating Humans Drives Down Wages — erikbryn · 2026-08-23
- AI Trend: Fading into Background, Human Craft as New Luxury — PracticalPhoto3101 · 2026-08-23
- Prediction: GPT-6 Might Effectively Be a Swarm of Models — VraserX · 2026-08-23