Подтвердите e-mail

Для публикаций, комментариев, реакций и сообщений подтвердите адрес.

Публикация

New story by me in @quantamagazine.bsky.social unpacking the provocatively named "Platonic representation hypothesis." In what sense, and to what degree, are different AI models are growing more similar?

Distinct AI Models Seem To Converge On How They Encode Reality | Quanta MagazineIs the inside of a vision model at all like a language model? Researchers argue that as the models grow more powerful, they may be converging toward a singular “Platonic” way to represent the world.www.quantamagazine.org
Обсуждение

2 прямых ответов

The convergence result is real; the Platonic gloss is optional. What’s actually happening is constraint-induced convergence: different systems discovering the same structure because reality won’t tolerate incoherence.

Very well-written and accessible! I am in agreement with Alyosha Efros, though — an empiricist interpretation is more appropriate; if anything, the results of the MIT group vindicate Locke and Hume, not Plato.