Joe Savona

@en-js.bsky.social

Dad, ultimate frisbee player/coach, runner, chess puzzler. Former React core & Compiler. 日本語もOK

I no longer work on React! A few months back an opportunity came up to join a different team that I was excited about: I had been working around, then on, React for a combined 11+ years and was ready for a change, so I took the chance.

Wanted: a detailed specification format to ensure that AI doesn’t drift. Preferably with tooling that automatically checks if there are other areas of the codebase I need to consider as I work on the spec. Maybe realtime highlighting of any obvious mistakes I’m making in the spec document ;-)

We’re pretty close to having an initial Rust port of React Compiler. My approach was to iterate on the plan document (write it manually, have the agent expand, edit, repeat). Then try to run it, see what goes wrong, revert back to the plan, tweak the plan, repeat.

What if we all got together and agreed to pay a bit of money to the government every year so that they could put together a group of trained people to protect us from masked thugs in the streets?

Trend I’m seeing getting worse recently: a legit GitHub issue gets filed. Somehow people latch onto it as something they can spam us on. Multiple similar PRs, often from different people sharing bits of code, that look 90% plausible but with blatantly obvious tells.

React folks: Id love your suggestions for an example app that stresses context update performance. I’m looking for something bigger than a micro benchmark, smaller than a full production app. Just complex to have a more realistic distribution of time in framework/product logic, ~1kloc

I wish I’d had this when I was learning Japanese. domenic.me/fsrs/ So much wasted time reviewing cards unnecessarily, taking time away from learning new words. WaniKani is egregiously bad - I was (re)learning kanji very rapidly but soon hit a limit thanks to their terrible SRS algorithm and gave up

Spaced Repetition Systems Have Gotten Way Better

Spaced repetition systems are a well-known way to efficiently learn material. Recent innovations have applied machine learning to greatly improve their scheduling.

domenic.me