✳flâneur — a map of the web's best reading

Deep learning as program synthesis — LessWrong

lesswrong.com · 21,016 words · saved by 1 readers

This is also far from finished research - more a snapshot of a hypothesis that seems increasingly hard to avoid, and a case for why formalization is…

x Deep learning as program synthesis — LessWrong Interpretability (ML & AI) Singular Learning Theory AI Frontpage 2026 Top Fifty: 17 % 150 Deep learning as program synthesis by Zach Furman 20th Jan 2026 49 min read 33 150 Epistemic status: This post is a synthesis of ideas that are, in my experience, widespread among researchers at frontier labs and in mechanistic interpretability, but rarely written down comprehensively in one place - different communities tend to know different pieces of evidence. The core hypothesis - that deep learning is performing something like tractable program synthes

Explore this link on the map →

saved by

related reading