HOW TO : Build a validated dataset with a decision model
I used Jev, a decision model, to sort 43,202 Texas food permits by cuisine in six minutes for about $1.30. Fast, cheap, no made-up answers, but a confident answer still isn’t a correct one.
I used Jev, a decision model, to sort 43,202 Texas food permits by cuisine in six minutes for about $1.30. Fast, cheap, no made-up answers, but a confident answer still isn’t a correct one.
Data backup wisdom meets manga. I used Meta’s Muse to turn the master-and-student lessons from Dr. Ross Williams’s taobackup.com into a comic.
I wanted to know whether lab.kudithipudi.org resolves the same way in São Paulo as it does in Ashburn. Instead of managing VMs around the world, I used eight region-pinned Cloudflare Durable Objects to run DNS-over-HTTPS lookups from different locations and compare the answers.
Every coding agent has hooks. None of them agree on what to call them. Instead of memorizing eight APIs, I think in three buckets: before an action, after an action, and at the session boundaries.
Bookmark is a self-hosted, AI-powered bookmarking service built with Python and SQLite. It automatically extracts metadata, generates tags and descriptions, creates local embeddings, and combines tag matching with semantic search to help you find saved links later.
A simple setup to save AI tokens when using coding agents