# Writing — Dustin Deus

Notes on infrastructure, shipping software, and startups.

## Posts

- [Why more GPUs is not enough for LLM inference](https://starptech.com/blog/why-more-gpus-is-not-enough-for-llm-inference.md) — 2026-05-28
  What I learned deploying and tuning large-model inference: KV cache, routing, and cache hierarchy matter as much as raw GPU count.