As of: 2026-11-19 22:11 CET. Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/ Canonical page: https://postcutoff.com/person/nathan-lambert/ # Nathan Lambert Co-founder and executive director, Trillium Labs; author of Interconnects (as of 2026-10-04; source: https://blog.trilliumlabs.org/p/introducing-trillium-labs). 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. ## Accounts - X: https://x.com/natolambert - Website: https://www.interconnects.ai ## Role changes - 2026-10-02: Launched Trillium Labs (co-founder and director), a nonprofit for open post-training recipes and open frontier-AI science. Source: https://x.com/natolambert/status/2106060179985019085 - 2026-06-02: Left Ai2 (post-training lead, co-led Olmo and Tulu). Source: https://www.interconnects.ai/p/farewell-ai2 ## In the news (8 events, newest first) - 2026-11-14: [Nvidia CEO search; petition to reinstate Huang](https://postcutoff.com/e/2026-11-14-nvidia-ceo-search-petition-reinstate-huang/), Business. Nvidia is the main supplier of AI accelerators and a financial backstop for many AI companies; who leads it, and whether its investment and lending web holds, affects compute supply and the pace of the build-out. - 2026-10-17: [DeepSeek-V4.1-Flash-1017 update: big agentic-coding gains, same price](https://postcutoff.com/e/2026-10-17-deepseek-v4-1-flash-1017/), Model releases. V4.1-Flash was already the open-weight model many developers used as a cheap coding agent. - 2026-10-17: [OpenCode breach leaks stored provider API keys](https://postcutoff.com/e/2026-10-17-opencode-console-breach-provider-keys/), Policy & safety. Labs have described distillation campaigns run through fake or resold accounts (see Anthropic's reports). - 2026-10-16: [China proposes export controls on its own AI models and agents](https://postcutoff.com/e/2026-10-16-china-proposes-export-controls-ai-models-agents/), Policy & safety. Chinese labs (DeepSeek, Qwen, Zhipu, Moonshot, StepFun) are the main source of open-weight frontier models used worldwide. - 2026-10-15: [Meta releases Muse Spark Open, an 816B open-weight model built with Apple to run on Mac Studios](https://postcutoff.com/e/2026-10-15-meta-muse-spark-open/), Open source. It is the largest open-weight release by a US lab in this dataset. - 2026-10-02: [Nathan Lambert and Tom Zick launch Trillium Labs, a nonprofit for open post-training recipes and open frontier-AI science](https://postcutoff.com/e/2026-10-02-trillium-labs-launch/), Open source. Fully open post-training recipes (as opposed to open weights alone) have been scarce. - 2026-09-29: [Anthropic: open-weights GLM-5.3 nearly matches Mythos Preview at exploit development](https://postcutoff.com/e/2026-09-29-anthropic-glm-5-3-spread-of-cyber-capabilities/), Policy & safety. It is the first time a frontier lab has published evidence that an open-weights model reached the level of exploit capability it had judged too risky to release widely. - 2026-09-29: [Meta FAIR and collaborators propose 'Context Language Models' that edit their own context as a file](https://postcutoff.com/e/2026-09-29-meta-fair-context-language-models/), Research. Context management (compaction, memory files, sub-agents) became a central engineering problem for long-running agents in 2026. ## Posts (5) - 2026-10-17: [“There’s something very odd going on with ‘distillation’ if a cyber attack to steal…”](https://postcutoff.com/p/x-natolambert-2111549984373911572/), Nathan Lambert (@natolambert), X. Cited in [OpenCode breach leaks stored provider API keys](https://postcutoff.com/e/2026-10-17-opencode-console-breach-provider-keys/). - 2026-10-15: [“The Meta Muse Open License is trash.”](https://postcutoff.com/p/x-natolambert-2110787899784601698/), Nathan Lambert (@natolambert), X. Cited in [Meta releases Muse Spark Open, an 816B open-weight model built with Apple to run on Mac Studios](https://postcutoff.com/e/2026-10-15-meta-muse-spark-open/). - 2026-10-02: [Nathan Lambert unveils Trillium Labs, a nonprofit for the open science of frontier AI](https://postcutoff.com/p/2026-10-02-lambert-trillium-labs-launch/), Nathan Lambert (@natolambert), X. 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. - 2026-09-29: [Nathan Lambert criticizes Anthropic's GLM-5.3 cyber report](https://postcutoff.com/p/2026-09-29-natolambert-glm-5-3-report-critique/), Nathan Lambert (@natolambert), X. Most prominent critique (70K views) of the Anthropic report from an open-model researcher. - 2026-08-07: [8 Predictions for the Era of Continual Learning](https://postcutoff.com/p/2026-08-07-dwarkesh-era-of-continual-learning/), Dwarkesh Patel (@dwarkesh_sp), Blog. 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) - 2026-05-02: [The New Claude Opus 4.7 Feature Developers Are Obsessed With](https://postcutoff.com/v/yt-mervin-praison-the-new-claude-opus-4-7-feature-develope/), Mervin Praison, YouTube, community, 3,512 views as of 8 October 2026. Profile checked 2026-10-04.