Interconnects

Interconnects is written by Nathan Lambert, a researcher who has worked inside major AI labs, which gives his commentary a level of technical grounding that's rare in the newsletter world. Issues dig into topics like model training techniques, reinforcement learning from human feedback, and the strategic decisions labs make — the kind of detail that's usually confined to research papers or internal memos.

It's grown quickly among people who want to understand the mechanics behind frontier model releases, not just the marketing around them. Expect long-form, opinionated essays rather than quick digests, aimed at readers who already know the basics and want the next layer of depth.