Writing from the field.
Research, product thinking, and notes on judgment worth carrying forward.

Atom, Bit, Commit, and the Emerging Unit of Agent Knowledge
Agents are becoming producers of findings. What unit lets the next person or agent inspect, govern, and build on a conclusion?
Read article →
Memory Is Not Enough: Multiplayer AI Needs a Trust and Verification Layer
Multiplayer AI is moving from individual copilots to teams of agents and humans working across research, analysis, decisions, and execution. As these systems enter the enterprise, the obvious infrastructure question is memory:
Read article →
Proofpress for the WebMCP Challenge
We built Proofpress for the WebMCP Challenge: a governed layer from agent-produced knowledge that the next agent and human can rely on with trust.
Read article →
Agents Are Creating a New Knowledge Layer—and We Need to Govern It
Most enterprise AI infrastructure is built around a familiar problem: how do we give agents better access to what the organization already knows?
Read article →
What May the Next Agent Rely On?
A seven-model study of governed knowledge in Harvey LAB-derived legal agent handoffs.
Read article →