Reports

Long-form writing on AI systems work: what breaks after the demo, what to measure, and how to leave the system operable.

1 report in Enterprise AI|LLMOps|AI agents

Latest in this domain

Enterprise AI|LLMOps|AI agentsSep 6, 20267 min read

Enterprise LLMOps Control Stack

rogue agent breach

Design for: rogue agent breach
Put kill switch, sandbox/ACL, and run ID reconstruction in place before write tools.
Measure task success and incident reconstructability, not only model latency.