makefunstuff

We have different IR backends to make it portable btw. There are llvm, qbe, etc In the repo I mentioned in the link am goofing around exactly with llvm, to actually figure it out.

I think you, got the point. I am under impression that we are going to nowhere with these approach of that token generation is a solution for everything. For me for some reason it feels like bruteforce but stakes right now are too high, so step back is not an option for bigger players.

Maaaybe, just mayyybe, in training weights there are not enough examples of producing valid assembly? And spitting something that was being fed by scraping oss repos is easy to impress for glorified autocomplete machinery?

If we assume and believe in statement that code produced in industry is not for reading but a product of llm, why do we need human readable code in the end then?

I have decided to play around with clankers to produce the most stupid PoC to find out where I can go with it to “vibe” out some “context compact” unreadable programming language. So my motivation of this post to discuss, why do we need to produce human readable code at all? If modern programming languages semantics optimised for humans, and society (at least in all these hype posts) are being forced to generated code no one reads, why do we need to waste compute on smth that is made for human to read? What the point of this compute waste?