zack (in SF)
@zack_overflow
24 eng @bunjavascript | i like systems programming/compilers/graphics/Zig/Rust/FP/Wasm/WebGPU
Revamped the gratuitous explosions and particle effects to the code editor I'm building And added sound effects! (Sound on!)
Whenever I meet someone I try to be as high energy and excited as possible because there's nothing more boring than meeting someone unexcited and low energy
Zig made a really good design decision to make the pointer dereferencing operator postfix so it could be easily chained But for some reason it made `try` prefix (unlike Rusts `?`) which makes not chainable and ugly

hey @github, why did you delete zackradisic’s (@zack_overflow’s) account and can you un-delete it?
Reading the React source is harder than reading the webkit codebase or the linux kernel
I wonder why JS bundlers decided to invent a new name and call it "tree-shaking" I feel like the traditional term, dead-code elimination, is way more descriptive
Zig will very annoyingly emit compile errors for unused variables, parameters But it won't for unused imports, which IMO is far worse than unused variables
zig is such an annoying brat about unused variables, there is a fork to solely add a flag to disable them seriously wtf is wrong with you andrew
I spent the whole day trying to fix a bug in Bun that @jarredsumner helped me fix in 5 minutes of explaining the problem to him
Sometimes I miss Rust's iterators so much I feel like a caveman when I have to write a for-loop by hand
TIL C++ lets you put return types in both directions (why?)
What do you think about trailing return types?
This is 100% a skill and tooling issue The real speed and productivity boost comes from running and orchestrating multiple coding agents at once Claude Code enables this (Cursor doesn't) and I don't think most developers really know yet how to micro >1 concurrent agents
We ran a randomized controlled trial to see how much AI coding tools speed up experienced open-source developers. The results surprised us: Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't.