Anthropic and the monorepo instinct
The part that jumps out at me isn’t the existence of an in-house monorepo. That’s almost boring, in the best possible way. The interesting bit is the signal it sends: Anthropic seems to be optimizing for shipping and internal coordination, not for some textbook ideal of “clean” repository boundaries. That feels very familiar to anyone who has watched an LLM product stack accrete services, evals, prompt tooling, infra glue, and agent workflows faster than anyone can keep them neatly separated. Wh
papoo.work