Nathan Lambert on the state of US open-weight models after Meta's big week
lmoroney · x · 2026-10-19
Laurence Moroney highlights Nathan Lambert's (Ai2, Olmo co-lead) reflection on Meta's big week, the state of US open-weight models, and the outlook for competing with Anthropic and OpenAI. His practical advice for ML learners: pick an open model, fine-tune it on a small dataset you care about, and measure before/after — one loop that teaches much of the craft.
More from coding & agent
- 4 Gemini 4 Argon agents team up for 27 hours to beat Minecraft's Ender Dragon — robertarail · 2026-10-19
- Hermes Agent official mobile app teased; dev built ChatGPT Voice 2.0 summarizer — alexcovo_eth · 2026-10-19
- DeepSeek 4.1 flash can near-match Opus 5.5's code-built Greek world with the right harness — victormustar · 2026-10-19
- Stardock CEO's agent benchmark run: only 2 of 5 tasks pass, agents not ready for professional work — draginol · 2026-10-19
- Developer builds open-world game vertical slice with Claude Opus, shares first screenshots — No_Criticism_2908 · 2026-10-19
- Dev experiments with open-source game where players vibe code their own characters — trq212 · 2026-10-19