//Writing
Notes on how systems misbehave.
Short pieces on measurement, reliability, and the failure modes that only make sense at the level of the whole machine.
- Pieces
- 8
- Latest
- 23 Sept 2026
- Feed
- RSS
8 pieces
-
23 Sept 2026
Method
I had five benchmark harnesses and they did not agree with each other
Every paper grew its own set of scripts, and by the fifth one I could no longer tell whether a difference between two results was the hardware or the harness. Rebuilding them as one framework was the only way to find out.
ReproducibilityToolingBenchmarkingMethod
-
23 Sept 2026
Measurement
Your local language model is waiting on memory, and one command will prove it
Text generation on a CPU is a memory problem wearing a compute problem's clothes. Here is the one-line rule that predicts it, the evidence from seven models on one laptop, and a tool that measures it on your machine.
llama.cppMemory bandwidthToolingRoofline
-
29 Aug 2026
Measurement
→Kubernetes did not slow my benchmark down. One line of YAML took 64 percent.
I moved an inference benchmark into a pod expecting to pay for the container. Across fifty runs I could not find a cost at all. What did cost me was a CPU limit set below the thread count, which is a thing you can write by accident and never notice.
KubernetesContainerscgroupsBenchmarkingllama.cpp -
09 Aug 2026
Colophon
→How this site is built
A static site with no CMS, no build-time magic, and a deliberate rule that the content is data rather than markup.
AstroTailwindDesign systems -
22 Jul 2026
Systems
→Reliability is a design property, not an operational outcome
How well a system survives contact with reality is mostly decided at design time, not during the incident.
ReliabilityDesignOperations -
15 Jul 2026
Systems
→Constraints shape system behavior long before they break it
Limits change feedback and stability well before anything crosses a threshold and trips.
ThermalControlFeedback -
08 Jul 2026
Systems
→Why local optimizations quietly break systems
An improvement made in one place can remove the slack another part of the system was relying on, and the bill arrives later.
PerformanceSecond-order effectsResilience -
01 Jul 2026
Systems
→Why most system failures aren't isolated failures
Failures in complex systems usually emerge from interactions between healthy components, not from a single defective one.
ReliabilityRoot causeFeedback loops