Hacker News Reader

Comment by ciupicri | Hacker News Reader

Since you've mentioned AI why not go directly to machine code, i.e. assembly or even binary?

Because code still would be read by humans. Check the code for PhotoCraft[0] (clean room reimplementation of Adobe Photoshop) which is being actively written in Rust by AI agents almost 24x7.

Check the overall layered architecture. That's nowhere AI. That's pure human ingenuity coupled with machine's raw horsepower to build on top.

And such people do need to read the code.

[0]. https://github.com/storytold/photocraft

I just wish Rust were slightly more readable...

The async terrain is absolutely bonkers and not just not very readable but also not very understandable either I suppose.

If the project is entirely made by AI agents, why read the code? Who reads the binary (or asm) files generated by their compiler except a few people in rare cases?

Given the fact that sometimes the architectures is included in the specifications for a project, I can accept that humans created the one for PhotoCraft.

Because it's still cheaper and safer to implement in Rust and allow a compiler to mechanically and deterministically transform it into machine code. The value of abstractions are not going away.

There used to be a Pascal to C compiler. CHICKEN is a compiler for Scheme that generated C files. TypeScript is compiled to JavaScript (because we didn't have WebAssembly back then). Other projects use the LLVM compiler infrastructure. If you're referring to this, then fine, but it's just intermediate stuff. The end result is still assembly / binary.

Much like humans, language models have context limits. Abstractions exist for a reason, and good abstractions simultaneously make code easier to read for humans and easier to process for language models

programming languages may just be a really good abstraction level to describe what you want.

That's just ridiculous. Not sure if you're serious, but if so, please make a serious attempt at this argument.

language models reason from human examples partly, and a lot of the abstractions are simply actual useful compressions of information that are useful for LLMs as well

Reply on news.ycombinator.com