miniblog.

LLMs are a really accessible machine learning technique. I dabbled with text classifiers a few years ago and the APIs were way more involved. (system_prompt: String, input: String) -> String I can prototype with this much more easily!

I find myself using a "splatter then clean up" technique when working with code. I add a ton of prints/logs/asserts/type annotations until I've found the issue. I then throw them all away! It feels like redundant work, but I don't have a nicer solution.

One interesting technique to reduce the review burden in Home Assistant, a project with a large community: ask PR authors to review other PRs! Screenshot is from