Anthropic vs Pentagon 🤖, SpaceX eyes March IPO 💰, lessons building Claude Code 🧑‍💻

The standout piece is Anthropic's "Lessons from Building Claude Code: Seeing like an Agent" — a framework for designing tools shaped to a model's own abilities, directly useful for anyone building agentic coding systems. On the industry side, Anthropic was labeled a national security supply-chain risk after refusing DoD contract terms that would allow domestic surveillance, while Claude subsequently overtook ChatGPT in the App Store — likely driven by public backlash at OpenAI's contrasting position. Apple is also replacing Core ML with a "Core AI" framework in iOS 27, signaling a broader platform shift in on-device AI tooling.