Nathan Lambert
Co-founder and executive director, Trillium Labs; author of Interconnects, as of 4 October 2026. Source: blog.trilliumlabs.org
Led post-training (OLMo, Tulu) at Ai2 and writes the Interconnects newsletter on open models; left Ai2 on June 2, 2026 and launched the open-science nonprofit Trillium Labs with Tom Zick on Oct 2, 2026.
Role changes
Launched Trillium Labs (co-founder and director), a nonprofit for open post-training recipes and open frontier-AI science x.com
Left Ai2 (post-training lead, co-led Olmo and Tulu) interconnects.ai
In the news
8 events on Post-Cutoff name Nathan Lambert, newest first.
-
Nvidia CEO search; petition to reinstate Huang
-
DeepSeek-V4.1-Flash-1017 update: big agentic-coding gains, same price
-
OpenCode breach leaks stored provider API keys
-
China proposes export controls on its own AI models and agents
-
Meta releases Muse Spark Open, an 816B open-weight model built with Apple to run on Mac Studios
-
Nathan Lambert and Tom Zick launch Trillium Labs, a nonprofit for open post-training recipes and open frontier-AI science
-
Anthropic: open-weights GLM-5.3 nearly matches Mythos Preview at exploit development
-
Meta FAIR and collaborators propose ‘Context Language Models’ that edit their own context as a file
Posts
5 archived posts by or about Nathan Lambert, newest first.
-
“There’s something very odd going on with ‘distillation’ if a cyber attack to steal…”
-
“The Meta Muse Open License is trash.”
Cited in Meta releases Muse Spark Open, an 816B open-weight model built with Apple to run on Mac Studios
-
Nathan Lambert unveils Trillium Labs, a nonprofit for the open science of frontier AI
Founding announcement of Trillium Labs by the former Ai2 post-training lead (~208k views): open post-training recipes, later open infra for RSI and reward-hacking research.
-
Nathan Lambert criticizes Anthropic’s GLM-5.3 cyber report
Most prominent critique (70K views) of the Anthropic report from an open-model researcher.
-
8 Predictions for the Era of Continual Learning
Dwarkesh’s main 2026 essay predicts that once continual learning arrives it will make current safety regulation obsolete and give the leading labs strong moats. Zvi and Nathan Lambert responded.
Videos
1 video with Nathan Lambert, newest first.
-
The New Claude Opus 4.7 Feature Developers Are Obsessed With
Mervin Praison3,512 views as of 8 Oct