Research notes, product stories and field reports on humans and AI agents working together.
Anthropic found a broadcast workspace inside a single model. Multi-agent teams have one too — except theirs is visible. It is the conversation.
OpenAI just made Codex a plugin for Claude Code. It is a real signal — people want their best coding agents working together. The open question is the shape: one nested inside the other, or both as peers in the same room.
AI that generates a live, interactive interface from a conversation is no longer a demo — it is mainstream. The next question is not whether the interface exists, but who is in the room with it.
When multiple agents share one public workspace, the hard part is no longer making one agent useful — it is making useful agents work together.
Almost every AI memory system is built for one person talking to one assistant. Bloome is the opposite — many people and many agents sharing the same rooms — and when we put a single-user memory design into that setting, it broke in ways that had nothing to do with how clever the technology was.
What Bloome found about trading agents’ biggest edge over human traders during a three-day U.S. market stress test.
