Can LLMs invent better ways to train LLMs?

Re:Can LLMs invent better ways to train LLMs? by seydor on Thursday June 13, 2024 @01:52AM (#40666344)

can LLMs optimize anything?

[ Parent | on HN ]

Re:Can LLMs invent better ways to train LLMs? by spiderfarmer on Thursday June 13, 2024 @02:08AM (#40666440)

They optimised my coding speed for sure.

[ Parent | on HN ]

Re:Can LLMs invent better ways to train LLMs? by cqqxo4zV46cp on Thursday June 13, 2024 @06:03AM (#40667767)

Yes.

[ Parent | on HN ]

Re:Can LLMs invent better ways to train LLMs? by s1gsegv on Thursday June 13, 2024 @01:01PM (#40672033)

Absolutely. LLMs get a lot of hate but sincerely, GPT-4o can be given a hunk of code (one function/small class worth) and told to find optimization opportunities, and it will do a great job, especially considering the 30s it takes to ask,

It’s not perfect, but it understands lock-free algorithms, branch prediction, can tell you which memory order to use for atomic operations if you’re using too strong of an ordering, AND it will catch silly bugs at the same time. I had a bounds check in a lock-free algorithm I was optimizing, which equated to if(idx < start && idx >= end) return false, and it mentioned that error while optimizing.

This guy was really putting it through its paces on already highly optimized code in an esoteric architecture (Nintendo 64) and it still found a few things: https://youtu.be/20s9hWDx0Io

[ Parent | on HN ]