Posts
Tokenomics: the 349 minute rule for GPT-5.6+
GPT-5.6 onwards made cache writes cost 1.25x. When is the write worth paying for, and how long is the cache worth keeping warm? Read More →
Tokenomics: Fable 5.1's reduced cache read makes it cheaper than Opus*
Anthropic made Fable 5.1's cache reads cheaper than Opus, when does that pay for its more expensive output, and how long is the cache worth keeping warm? Read More →
Tokenomics: the 62.5-minute rule for Claude's cache
Is it more efficient to refresh the 5-min cache, let it expire, or just rely compaction? Read More →
Patching Berkeley Mono with Nerd Fonts for Windows Terminal/WSL
How to patch Berkeley Mono with Nerd Fonts, fix OpenType name tables to preserve Windows font families, and get bold/italic working in Windows Terminal. Read More →
Slopdocs: notes for the next clanker
Why I started keeping the planning docs and bug postmortems my clanker generated, instead of letting them disappear with each new session. Read More →
Orchestrating AI Code Review at scale (external link, opens on The Cloudflare Blog)
Learn about how we built a CI-native AI code reviewer using OpenCode that helps our engineers ship better, safer code. Read on The Cloudflare Blog →
We're back baby! New year, new blog updates 🚀
It's a new year, which means it's time for me to make some completely necessary (and not at all frivolous) updates to my personal site! Read More →