I'm currently writing a compiler (forever garage project) the speed at which Claude is able to debug with gdb what was wrongly generated is insane. I had not much prior expertise so maybe a professional in the field would be able to do it faster, but now people new to the field are able to debug at a very fast pace as well. It looks at ask code generated, understands llvm, can read and instrument gdb, all for 15$ a month
I recently just referenced the selective applicative functors paper and let it write me an implementation in scala. There is one already available in github, so I can't judge if it really just read the paper and implemented it, but the result was so minimal quick and amazing.
No, I don't think OP's comment is useless. It's not just a blanket statement, "Wow, LLM's quick", but also critically reflective, like a brief limitations section. This opens the discussion, by indirectly posing questions like, "Is it really this fast or "cheating"? How could we measure this experimentally? Etc.
It's useless because this is not novel at this point, its literally everyone posting on HN - I did a thing with an LLM but idk if its good! That doesn't open up the discussion, it retreads the exact same discussion that both sides are not listening to each other on.
If they succeed the software will be more reliable with less memory issues that are very likely significant security issues at least some of the time.
When we've seen linux having a new significant exploit every other day now thanks to LLMs being better at weaponizing memory bugs this seems significant.
> False, it creates consumer demand for inference chips, which will be badly utilised.
There are so many CPUs, GPUs, RAM and SSDs which are underutilized. I have some in my closet doing 5% load at peek times. Why would inference chips be special once they become commodity hardware?
