Swyx digs into MIT CSAIL's viral 'Weight-Space Programming': cool, reproducible, but temper the hype
rschu · x · 2026-11-05
MIT CSAIL's "Weight-Space Programming" post is exploding on timelines, with claims that LLMs will soon replace gradient descent and AdamW. Swyx (rschu retweeting sitzikbs) says he dug through the code and post: the work is very cool and reproducible, but the hype needs tempering — analysis expanded in the rest of his thread.
Related event: MIT's Weight-Space Programming Goes Viral, Sparking Debate(3 posts)→
More from Models
- Ricursive releases Ripple-65B Preview, claiming US open-weight SOTA on agentic coding — DZhang50 · 2026-11-05
- Liquid AI ships LFM 3 models (1.7B/2B) running locally on iPhone Duo via MLX — helloiamleonie · 2026-11-05
- particle-1 tested on real agent work: near-Opus quality for under $2 a day — heyneighbor · 2026-11-05
- Speculative decoding makes Qwen3.6/3.8 much faster, latest draft model adds ~10% — teortaxesTex · 2026-11-05
- webAI open-sources 8.8B TwIL-8B hitting 93% of frontier coding performance — dr_cintas · 2026-11-05
- Open-source TTS: Phonon models under 50M params run on-device with 7ms latency — alex_verem · 2026-11-05