May 19, 2026 Your Agent Harness Is the Attack Surface You Are Not Watching AI agent harnesses grant powerful capabilities and then try to contain them. The security model underneath this is thinner than most teams realize, and attackers are paying attention. #security #agents #infrastructure #opinion
Mar 7, 2026 Incident Response When the System Has Opinions Traditional incident response assumes deterministic systems. What changes when your system can creatively decide to do the wrong thing? A practical guide to runbooks and response patterns for agent failures. #agents #infrastructure #security #opinion
Jan 22, 2026 Agents Are Just Distributed Systems With Opinions Most of the hard problems in agentic systems are distributed systems problems you have been solving for years. The LLM is just a nondeterministic function in the middle. #agents #distributed-systems #infrastructure #opinion
Dec 9, 2025 Observability for Systems That Lie to You Your agent returns 200 OK and is confidently wrong. Traditional health checks catch crashes. They do not catch a system doing the wrong thing with full conviction. #agents #observability #security #infrastructure
Nov 3, 2025 The Case for Boring Infrastructure in AI Systems The model is the innovation. Everything around it should be as predictable as possible. A complexity-budget argument for choosing PostgreSQL over the hot new vector database. #infrastructure #agents #opinion #architecture