I hate compilers
xeiaso.net
I hate compilers
1–10 of 182 posts
Re: I hate compilers
#2As long as the program is equivalent there isn't an actual problem here. Requiring the output to always be the same is an arbitrary restriction.
If you want to have users trust that someone else hasn't modified it, then sign it with your identity.
Re: I hate compilers
#3LLMs should be trained on and directly output binary.
Re: I hate compilers
#4[flagged]
Re: I hate compilers
#5LLMs should be trained on and directly output binary.
On the off chance that you’re serious, that would result in disastrously bad output. The difference between “jmp $+15” and “jmp $+16” is inscrutable and the LLM would not be able to pick the right one without tooling.
That tooling is a compiler. The higher level, the better chance the LLM can be steered to good output. Machine code is hopeless, don’t bother.
Re: I hate compilers
#6[flagged]
Tell me the flags I'm missing then: https://github.com/TecharoHQ/anubis/blob/Xe/wasm4/utils/wasm...
Re: I hate compilers
#7LLMs should be trained on and directly output binary.
On the off chance that you’re serious, that would result in disastrously bad output. The difference between “jmp $+15” and “jmp $+16” is inscrutable and the LLM would not be able to pick the right one without tooling. That tooling is a compiler. The higher level, the better chance the LLM can be steered to good output. Machine code is hopeless, don’t bother.
What about AOT optimization? whuch brings aot closer to JITs performance? Isn't that something LLM + Harness can easily do?
Re: I hate compilers
#8LLMs should be trained on and directly output binary.
It should not. Abstraction in software engineering brings intelligence. (compression correlates to intelligence)
Re: I hate compilers
#9LLMs should be trained on and directly output binary.
I think you forgot the "/s"
Re: I hate compilers
#10LLMs should be trained on and directly output binary.
On the off chance that you’re serious, that would result in disastrously bad output. The difference between “jmp $+15” and “jmp $+16” is inscrutable and the LLM would not be able to pick the right one without tooling. That tooling is a compiler. The higher level, the better chance the LLM can be steered to good output. Machine code is hopeless, don’t bother.
> The difference between “jmp $+15” and “jmp $+16” is inscrutable
I don't see why that's the case. LLM trained on binary would totally see it, not?
Also the tool can also be running the test and a debugger.