Why AI hasn't yet proven useful deep learning theory: toy models are too hard to design
jasondeanlee · x · 2026-11-21
@Yann77 notes we're still waiting for AI to prove a useful deep learning theory result, arguing the blocker is designing toy models tractable enough to analyze yet rich enough to capture what happens at scale — and humans themselves lack much data on this. jasondeanlee endorses with a "Yep."
More from AGI Musings
- NEJM Analysis of $45B in Medicare Payments: Pathology 99% AI-Exposed, Dermatology 7% — HealthcareAIGuy · 2026-11-21
- VraserX: AI Should Let Ordinary People Build Entire Companies Without Hiring Anyone — VraserX · 2026-11-21
- Ex-A NSA Council cyber director: AI-driven exploitation and rushed patches are causing outage wave — R_D · 2026-11-21
- Opinion: stateful AI breaks the headcount scaling law — one operator running 9-figure conglomerates — aryanXmahajan · 2026-11-21
- Arcfront launches to teach AI judgment; 4 of 5 frontier labs already customers — alifcoder · 2026-11-21
- Vibe coding as revenge: users ditch Adobe and Microsoft, while veteran coders mourn a lost craft — SydSteyerhart · 2026-11-21