Every post filed under architecture.
Before goldengoose had a team-status card, the lead drew one in Markdown. Turning that repeated report into deterministic UI gave the model and me one runtime-owned view of the team.
I didn't plan to build a process manager. But the two providers under goldengoose disagreed about everything terminal-shaped, neither gave me programmatic control or observability, and both wasted agent turns on polling. Owning that layer changed how I ship daily.
goldengoose has a native Rust runtime underneath it, and everyone assumes that's where the hard engineering lives. It isn't. The frontend, rendering twenty concurrently streaming agents without guessing, freezing, or lying, took more out of me than the runtime ever did.
I climbed every rung of the LLM abstraction ladder, from raw API wrappers through my own unified interface to the Vercel AI SDK, and every rung was the same completion API underneath. Then the providers open-sourced their harnesses, and the right foundation for agent products finally existed.
Everyone keeps answering 'how do agents coordinate?' with more hierarchy: spawn trees, sub-agents, orchestration layers. After building a multi-agent coding product, I think the answer is two primitives and a flat org chart.