2026-09-30 · latest
When to delegate coding work from a scarce premium model to cheaper ones — and when not to. A multi-file benchmark where hierarchical orchestration matched solo pass rate, burned similar premium tokens, and was ~2.8× slower.
agents · orchestration · llm · benchmarks · workflows
2026-04-08
The full technical account of building a distributed local voice assistant: system split, speech pipeline, PostgreSQL memory, and what a formal benchmark campaign measured.
llm · voice · docker · postgres · self-hosting · thesis · benchmarks