Oct 9, 2026 / Systems, Latency, Reliability
Queueing Theory in One Picture
Spare capacity is what lets a bottleneck absorb variation without making every request wait.
Read postSoftware, systems, and the p99
It is not enough for code to work.
“The best way I’ve found to have a positive impact on code is to write it.”
Engineering notes
2026
Oct 6, 2026 / Systems, Reliability, Overload
Retries as a Load Multiplier
A retry gives one request another chance. Across a fleet, it can keep a failing dependency from recovering.
Read postOct 1, 2026 / Systems, AI, Reliability
Skill Is a Dependency
The silent failure in AI-assisted engineering is you
Read postSep 23, 2026 / Systems, Delivery, Reliability
Ship Fast. Recover Faster.
Fast delivery depends on small changes, canary rollouts, production telemetry, and a tested rollback path.
Read postSep 7, 2026 / Systems, AI, Reliability
Synthetic Resilience
A coding agent can turn a dependency failure into silent, incorrect behavior by inventing a fallback your system was never designed to use.
Read postAug 25, 2026 / Systems
Conformance Is Not Interoperability
Devices do not always behave according to their advertised capabilities. Reliable integrations verify behavior at runtime.
Read postAug 23, 2026 / Meta
Why p99log
What this is, what the name means, and why the tail matters more than the average.
Read post