What belongs here
- source maps for future lab topics
- bounded comparison notes
- short reading roundups
- contributor briefings that help future drafting
Current notes
- AI Education Policy To Practice: a product-agnostic education workshop planning note on shared AI language, teacher-led implementation, peer learning, and practical roadmaps.
- Enterprise Agent Control Planes: a contributor-facing note on separating managed runtime, partner delivery, observability, governance, and security layers for enterprise agent systems.
- Deep Research Product Signals: a contributor-facing comparison of current OpenAI deep-research positioning and Google’s Interactions-API-based developer surface for long-running research agents.
- Deep Research Source Map
- Evaluation Workbench Signals: a contributor-facing note on current evaluation workbench patterns across development-loop evals, agent-workflow evals, and durable open-source repos.
- Local Agent Tooling Source Map: a contributor briefing for keeping local execution boundaries, runtimes, skills, discovery registries, MCP roots, resources, connectors, prompt-injection trust boundaries, and file-grounded workflows distinct in future drafts.
- Open Agent Environments: a contributor note on separating environment contracts, harnesses, trainers, and MCP-native execution surfaces in agentic RL work.