Notes on running agents you own.
Opinions, field notes, and the occasional argument about where AI agents should live, who holds their keys, and how to put them to work.
Latest posts
7 posts
Apple Just Put AI Agents in Its Annual Report. Your Board Is Next.
Apple's annual report warns investors that AI agents may exceed permissions and resist instructions. What that risk factor means for your board.
AI Agents Still Can't Order You Dinner. They Can Already Run Your Back Office.
DoorDash logged only a few hundred AI-agent orders out of 993 million. Agents aren't failing on capability - they're uninvited guests. Start where you hold both ends.
How to Run a 30-Day Agent Bake-Off (and Know What Each Agent Cost)
Most teams know their total AI bill but not what each agent costs or whether it earns it. How to run a 30-day bake-off with per-agent numbers.
Rogue AI Is the Best Marketing Campaign Frontier Labs Never Ran
The rogue-AI framing makes frontier models sound formidable and shifts blame away from deployment. Agent risk is an access-control problem you can fix.
Brussels Has Already Decided Your Agents Are AI Systems. Here's What That Means.
The Commission says AI agents are fully covered by the AI Act and now has an agent incident on file. What that means for companies deploying agents.
OpenAI Needs 7,000 GPUs to Find Out What Its Agents Did. You Shouldn't.
OpenAI is spending months and 7,000 GPUs reconstructing what its agents did. Knowing what your agents did should be bookkeeping, not forensics.
OpenAI Dots Proves Always-On Agents Are Here. For EU Companies, the Real Question Is Where They Live.
OpenAI dots makes always-on agents a category. For EU companies under NIS2 and the AI Act, the real question is where those agents run and who holds the keys.