Weight-space programming result reproduced, but 'emergent' claim disputed
anshulkundaje · x · 2026-11-03
Alex Krentsel reproduced the MIT blog post's weight-space programming results and added that this is a step-change emerging capability appearing only in the latest models.
@probablybots pushes back: hypernetworks have existed for years and building an RL env for weight generation/modification in smaller models is straightforward—it likely isn't emergent, and may just reflect math/code diminishing returns as labs move on.
More from Research
- ARC Prize 2026 closes with 18.5K entrants and 67K submissions, winners due December 4 — GregKamradt · 2026-11-03
- Should AI read cardiac MRI first? Lancet Digital Health study on ejection fraction — EricTopol · 2026-11-03
- "Emergent capability" claim mocked; Fleet hiring researchers to automate mech interp with frontier models — sebkrier · 2026-11-03
- Goodfire's Neural Geometry series: steering models along manifolds — burny_tech · 2026-11-03
- Prediction: an open model rated for 10M context within 3 months — teortaxesTex · 2026-11-03
- Why powerful AIs could cooperate in one-shot prisoner's dilemmas: open-source code proofs — tszzl · 2026-11-03