bsky.app/profile/dday...

"knowing what you don't know" seems like a prerequisite for calling anything "intelligence," at least how i see it

small world - haven't been in touch for a bit, but fellow william & mary friend stephen reader is their drummer. always thought he'd get famous

it's definitely tough because you end up with a lot of non-idiomatic stuff in the middle, and also you can technically use effect in a very totalizing, dogmatic way. we're basically adopting it by replacing existing deps with it - so far ts-pattern, cron-validator, toad-scheduler

i've seen variations of this function so many times and it never becomes less shocking, as does probably the performance once it goes over 9 github.com/Effect-TS/ef...

the level of abstraction inside of effect scares me a bit - lots of code that raises alarm bells for javascript deopt cases, like juggling around arguments objects. probably won't show up in real scientific profiling though.

running slowly isn't more injury-inducing, it's that doing the same mileage training load at slower paces means you're simply out there for longer, so the wear & tear is higher

unfortunately that trend has extended to marathons too, with bad results