AlphaGo insider: Don't be fooled—today's LLMs still can't truly reason
luislamb · x · 2026-10-12
Thore Graepel, who helped build AlphaGo and witnessed the 2016 match against Lee Sedol, argues in an MIT Technology Review opinion piece titled Don't be fooled—LLMs don't reason.
- He recalls Move 37 in game two—a play so odd commentators suspected a glitch, yet emblematic of AlphaGo's creativity; Lee Sedol admitted afterwards that "AlphaGo is creative."
- He contrasts Deep Blue's 1997 win: brute-force search over 200 million positions per second with human-coded rules. Go's complexity made that impossible; AlphaGo had to grasp the position at a glance and invent novel moves.
- His thesis: ten years on, today's LLMs still don't tap into the machinery that made AlphaGo's win possible, and we shouldn't mistake their surface abilities for genuine reasoning.
More from AGI Musings
- Software Is Not Dead: AI's Real Disruption Isn't the Apps You Know — bennash · 2026-10-12
- Billion AI scientists will tackle aging, predicts AI commentator — Dr_Singularity · 2026-10-12
- Steve Yegge's favorite essay in 3 years: adaptive complexity as the supervalue — Steve_Yegge · 2026-10-12
- A chatbot's overly familiar greeting reveals why humans shouldn't default to trusting LLMs — relic_radiation · 2026-10-12
- AI's Real Disruption Is the Invisible World of Custom Software, Not Photoshop — bennash · 2026-10-12
- New Yorker Cartoon: 'We Need to Rethink Our Strategy of Hoping AI Will Just Go Away' — c_valenzuelab · 2026-10-12