Field notes
What we are learning, in public.
Engineering deep-dives, design essays, and customer stories from the team building the agent OS.
Engineering
Replay everything: an audit log that doubles as a debugger
How we built immutable, queryable agent histories with sub-second seek.
Jonas Vetter · May 18, 2026
DesignThe flat UI thesis: less affordance, more authority
Why we removed shadows, gradients, and animation from the dashboard.
Priya Subramanian · May 11, 2026
ProductPolicies that don't lie: a DSL for AI approvals
Three failure modes we saw in customer rollouts — and how a typed policy layer fixed them.
Hannah Schulze · May 4, 2026
CustomerNorthBank's 11-day loan review, in 6 hours
A case study on agent fleets in regulated finance.
Oliver Reed · April 28, 2026
AIVerifiers > vibes: training a model to catch the others' mistakes
Our verifier model improved end-to-end accuracy by 7 points.
Daniel Park · April 21, 2026
SecurityPrompt injection in production: a 90-day field report
What actually happens when agents face hostile inputs.
Renata Lima · April 14, 2026