Lambert's testimony: China has clearly led open-weight models since April 2025
natolambert · x · 2026-09-21
Nathan Lambert (Interconnects) was invited to brief US Congressional members and staff on open-weight models amid US-China competition, and expanded his prepared remarks into "The current balance of power in open models." Key points:
- Taxonomy: closed models (GPT-4, Claude — API/product access only), open-weight models (weights public under licenses, e.g. Llama, Qwen, Gemma, DeepSeek), and true open-source models (weights + license + inference code + full training code and data, led by Ai2's OLMo).
- Core claim: since around April 2025, Chinese AI companies have been the clear leader in open-weight models, while genuinely reproducible open-source models remain a mostly US endeavor led by the Allen Institute for AI.
Related event: Nathan Lambert Briefs US Congress on Open Models and China Competition(6 posts)→
More from Safety
- Researcher says Anthropic dodged questions on Mythos false positive rate; 37,966 findings, 2 exploited — OwariDa · 2026-11-12
- Who reads the queries your agent writes? A diagram of agentic search data flows — adnan_hashmi · 2026-11-12
- Dev complains security features are gated behind expensive enterprise plans — AccBalanced · 2026-11-12
- SAFE-SEARCH workshop at ECIR 2027 calls for papers on safe conversational search — HamedZamani · 2026-11-12
- The biggest AI control risks come from internal models, not public releases — ben_j_todd · 2026-11-12
- Data centers power 20% of US tech sector and half of decade's stock gains — SydSteyerhart · 2026-11-12