The case for local LLMs

If you can run inference locally, you should — not despite the limitations of smaller models, but because of them.

AI-infused software is eating up the world

An opinion piece on how LLMs are becoming ubiquitous, the hidden costs of AI inference compared to traditional compute, and where the industry is heading as AI companies approach historic valuations.

A year of early morning dives

A year of early morning swims: building discipline, finding rhythm, and what cold water teaches you about showing up.