TsumegoBench: How Good Are LLMs at Go?
I gave current LLMs simple life-and-death problems. More interesting than the ranking was how quickly I could build a benchmark of my own.
Read moreThings I build and think through
AI, personal tools, cost reality, small experiments, and the life around them.
Latest notes
Remote Computing
How I leave my Mac at home and still access it from my iPad and work with coding agents.
Read moreHow Much AI Can I Afford?
After the Anthropic cut, I compare my Claude Code usage against API prices and look at which models still fit into my 100-dollar budget.
Read moreGoodbye Anthropic
Why I'm cancelling my Anthropic subscription, what made Claude Pro so important for my agents, and why we should get ready for real API pricing.
Read moreMaking More Memories with the ABC Challenge
Why I decided to collect one new experience for every letter in 2026, and how that turned into my ABC Challenge Tracker.
Read moreHello Internet
Why I started another website
Read more