Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Untitled topic

news.ycombinator.com

Post reply on HN
Print view View original

51–60 of 61 posts

Re: undefined

#52

Post by accheng »

SafeDusk wrote: post #51

Great to see progress being made here! I had tons of fun using AlphaEvolve to optimize Perlin Noise[0] [0]: https://blog.toolkami.com/alphaevolve-toolkami-style/

Thanks for sharing your blog! Very interesting work, 100% agree with your 3 criteria on the sweet spot for AI. Most systems performance problems fit right in

Re: undefined

#53

Post by accheng »

Noumenon72 wrote: post #34

There's nowhere on the page to find out what "ADRS" stands for since the upper left is cut off and isn't a link to your home page.

ADRS = AI Driven Research for Systems. See our previous blog post (https://www.sigops.org/2025/barbarians-at-the-gate-how-ai-is...) and our paper (https://arxiv.org/pdf/2510.06189) for more details!

Re: undefined

#54

Post by accheng »

kristjansson wrote: post #44

This is quite cool, but I must note that the 5x reported in the headline is the _runtime_ of the load balancing algorithm itself, not the load factor or throughput of the system or what have you. > On average, it takes about 540 ms to re-balance the experts and achieves a load balance factor of 0.66 (calculated as the ratio of average to maximum tokens generated per GPU). > ... > We also consider a non-public referen…

That's a good point! The load balancing of the original algorithm was already quite good so our goal was to try to get something that could achieve similar results but could run faster since runtime was also a concern.

Re: undefined

#55

Post by accheng »

mavt6 wrote: post #23

i'm skeptical this generalizes beyond problems that can be expressed as "rearrange tensors faster". it feels like a solution that only works for a very narrow and convenient class of problems.

We've found that these frameworks do well for systems performance problems and expect that the range of problems for which they apply will increase as they models and frameworks improve. See our paper (https://arxiv.org/pdf/2510.06189) for more discussion about this

Re: undefined

#59

Post by coliveira »

selcuka wrote: post #50
coliveira wrote: post #39

Earlier quoted context omitted.

As you say, the odds of this happening are very close to zero. But suppose for a minute that this was possible. Did you learn anything? Do you really have a discovery? Was this done using a novel method or applying something that already exists? If you give this to somebody else, should they believe it works? Is this result even understandable by human beings? You'd need to answer so many questions that in the end ev…

Scientists discover things that I don't understand every day. A sufficiently advanced discovery in, say, mathematics can only be understood by other mathematicians. Does that make it less of a discovery? So what's wrong if a machine discovers something that can only be analysed and proved by other machines?

The mathematicians understand the discovery, so what's the problem? With "AI", nobody understands it. If you ask, the AI may say it understands the results, but it is probably lying.

Re: undefined

#60

Post by PostOnce »

malfist wrote: post #38
PostOnce wrote: post #37

Earlier quoted context omitted.

Theoretically if I were to type into an LLM " Write a novel compression algorithm for images that is at least 25% smaller at the same speed and quality as ___ " and it did, and I ran the code (which I didn't understand) and it worked, wouldn't that count? The odds of that working, though, are of course pretty near 0. But theoretically, it could happen.

You might find that if it did produce something, it might not be _novel_

I might, but by the same token, I might also find that it was novel?

It's a remote possibility, but it is a possibility, isn't it?

Post reply on HN