AI Still Has to Run Somewhere
As agents take on business work, choosing a model is only one part of choosing how the business will operate.
Read field note →Topic archive
identity-security
Who Is Watching the Agents?
The more authority we give AI agents, the better our evidence needs to be about what they actually accomplished.
Read field note →I Want a Personal AI Agent. I Also Want It to Know When to Stop.
Personal agents are moving from conversation to delegation. The real question is not only what they can do, but what makes them stop.
Read field note →If Your AI Agent Goes Rogue, Can You Actually Stop It?
If an AI agent goes outside its intended purpose, can you reduce its authority, require approval, revoke access, or pause it?
Read field note →A Valid AI Identity Can Still Do the Wrong Thing
A valid AI identity can still do the wrong thing.
Read field note →What If the AI Behind Your Next Security Incident Wasn't Hacked?
What if the AI behind your next security incident wasn't hacked?
Read field note →Identity Stops at Login. AI Doesn't.
AI agents can authenticate once and continue making decisions, accessing data, calling tools, and taking actions long after the initial access decision.
Read field note →Identity Stops at Login. AI Doesn't.
AI agents do not just change who needs an identity. They change how long we need to keep making trust decisions.
Read field note →No rogue AI. Just one over-trusted tool.
No rogue AI. Just one over-trusted tool.
Read field note →