Conscious Turing Machines Model Pain but Don't Feel It, Argues Dietterich
tdietterich · x · 2026-10-08
In the ongoing AI pain debate, Thomas Dietterich adds: painkillers interrupt or dampen pain, so one criterion is whether an artificial system reacts the same way (he's only half kidding). Sharper still, he cites the Blums' Conscious Turing Machine: it models pain yet clearly doesn't experience it—so any criterion satisfiable by a CTM cannot prove experiential pain in a computing system.
Related event: Turing Award Winner Debates Whether AI Can Feel Pain(3 posts)→
More from AGI Musings
- Meme: mathematicians face existential AI crisis while everyone else just works — SuB8u · 2026-10-08
- Norvig's Classic Essay on Chomsky and the Two Cultures of Statistical Learning Still Reads Fresh in the LLM Era — 3scorciav · 2026-10-08
- Redditor calls out AI leavers for vague existential-risk talk with no specifics — IShootThemSteve · 2026-10-08
- Cryptographer Matthew Green: AI labs employ cryptanalysts, disclosure must be cautious — matthew_d_green · 2026-10-08
- Peter Yang: AI solved images, music, video — gaming is next — petergyang · 2026-10-08
- 'Corporations are superintelligence' takes get mercilessly mocked — tszzl · 2026-10-08