Replying to @archeologistdev @patrickgwsmith
overall it was pretty smooth although I did see a couple of errors on qip.dev in the examples
had some fun implementing Conway's game of life with @patrickgwsmith's qip today. hand-coded it in zig, compiled to wasm, works in the browser just as well as in the terminal, and 64k of memory is more than enough. Each frame takes ~137µs to render. Nice!
We deleted 800k lines of unit tests from the @sazabi monorepo earlier this week.
Three hypotheses:
1- Unit tests "lock-in" slop code by making it hard to change or remove
2- Unit tests slow down dev cycles by making your agent do more work locally and by increasing CI times
3- Unit tests waste money by making your agent generate more tokens and by adding useless CI runs
Will report back soon with the results!
Replying to @archeologistdev @bcantrill
My colleague Paul Thompson (I don’t think he’s on x?) shared another interesting perspective. Letting LLMs do the writing erases the author’s “accent” -- their own unique voice.
paulthompson.dev/posts/accen…
Replying to @bcantrill
Yeah, it’s hard to keep reading when a piece reeks of LLM.
I found it difficult to finish reading oxide.computer/blog/performa… for this reason. While it was technically very interesting, the rampant use Claude-isms throughout was a bit of a bummer.









