Nathan Lambert's Congress Testimony: China Has Led Open-Weight Models Since April 2025
natolambert · x · 2026-09-21
Nathan Lambert published the full version of his Congressional briefing on open models: open-weight models (Llama, Qwen, Gemma, DeepSeek) vs. true open-source ones (AI2's Olmo), with the core claim that Chinese companies have clearly led open-weight models since about April 2025, plus data and predictions on US-China competition.
Related event: Nathan Lambert Briefs US Congress on Open-Source AI and China Competition(2 posts)→
More from Models
- Alsop slams Claude and Codex subscription shrinkflation, routes tasks via classifier to cut tokens — StewartalsopIII · 2026-09-21
- Paying $1,400/mo across 7 AI subscriptions, he says free Antigravity wrote the best copy — aryanXmahajan · 2026-09-21
- Building a homework-correction pipeline in 30 minutes with Ling-3.0-flash-VL — nikola_mr64990 · 2026-09-21
- Qwen confirms you own the outputs generated with Qwen 2.1 — Metallic_Madness · 2026-09-21
- Confident Nonsense: why polished AI answers are the ones you should trust least — nikola_mr64990 · 2026-09-21
- Qwen3.8-27B in native 8-bit hits 37-55 tok/s on Apple Silicon, avoiding the 4-bit reasoning cliff — SnooPredictions515 · 2026-09-21