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

Open source AI is the path forward

about.fb.com

Post reply on HN
Print view View original

681–690 of 936 posts

Re: Open source AI is the path forward

#681

Post by manux »

CardenB wrote: post #532
photonthug wrote: post #519

Earlier quoted context omitted.

> Not sure why a publicly accessible GPU cluster would be a better solution than the current system of research grants. You mean a better solution than different teams paying AWS over and over, potentially spending 10x on rent rather than using all that cash as a down payment on actually owning hardware? I can't really speak for the total costs of depreciation/hardware maintenance but renting forever isn't usually a…

Do you have some information to share to support your bias against leasing especially with a depreciating asset?

In Canada, all three major AI research centers use clusters created with public money. These clusters receive regular additional hardware as new generations of GPUs become available. Considering how these institutions work, I'm pretty confident they've considered the alternatives (renting, AWS, etc). So that's one data point.

Re: Open source AI is the path forward

#682

Post by fnordpiglet »

msnkarthik wrote: post #680

Interesting discussion! While I agree with Zuckerberg's vision, the comments raise valid concerns. The point about GPU accessibility and cost is crucial. Public clusters are great, but sustainable funding and equitable access are essential to avoid exacerbating existing inequalities. I also resonate with the call for CUDA alternatives. Breaking the dependence on proprietary technology is key for a truly open AI ecosy…

I suspect we are still early in the optimization evolution. The weights are what matter. The ability to run them anywhere might come.

Re: Open source AI is the path forward

#683

Post by ohthehugemanate »

Has anyone taken apart the llama community license and compared it to validated open source licenses? Red Hat is making a big deal about releasing the Granite LLM released under Apache. Is there a real difference between that and what Llama does?

https://www.redhat.com/en/topics/ai/open-source-llm

Re: Open source AI is the path forward

#684

Post by troupo »

fnordpiglet wrote: post #682
msnkarthik wrote: post #680

Interesting discussion! While I agree with Zuckerberg's vision, the comments raise valid concerns. The point about GPU accessibility and cost is crucial. Public clusters are great, but sustainable funding and equitable access are essential to avoid exacerbating existing inequalities. I also resonate with the call for CUDA alternatives. Breaking the dependence on proprietary technology is key for a truly open AI ecosy…

I suspect we are still early in the optimization evolution. The weights are what matter. The ability to run them anywhere might come.

The training datasets and methodology are what matters. None of that is disclosed by anyone

Re: Open source AI is the path forward

#685

Post by Aeolun »

stavros wrote: post #425
causal wrote: post #41

Earlier quoted context omitted.

"Open weights" is a more appropriate term but I'll point out that these weights are also largely inscrutable to the people with the code that trained it. And for licensing reasons, the datasets may not be possible to share. There is still a lot of modifying you can do with a set of weights, and they make great foundations for new stuff, but yeah we may never see a competitive model that's 100% buildable at home. Edit…

"Open weights" means you can use the weights for free (as in beer). "Open source" means you get the training dataset and the methodology. ~Nobody does open source LLMs.

I suspect that even if you allowed people to take the data, nobody but a FAANG like organisation could even store it?

Re: Open source AI is the path forward

#686

Post by miffy900 »

ab5tract wrote: post #569
cheema33 wrote: post #502

Earlier quoted context omitted.

> True open source advocates will find little to refute in what I’ve said. No true Scotsman https://en.wikipedia.org/wiki/No_true_Scotsman OSI helped popularize the open source movement. They not only make it palatable to businesses, but got them excited about it. I think that FSF/Stallman alone would not have been very successful on this front with GPL/AGPL.

Like I said, honest open source advocates won’t take issue to how I framed their position. Here’s a more important point: how far would the open source people have gotten without GCC and glibc? Much less far than they will ever admit, in my experience.

> Most of us who were there remember it differently. True open source advocates will find little to refute in what I’ve said.

> Like I said, honest open source advocates won’t take issue to how I framed their position.

Yet you've failed to provide even a single point of evidence to back up your claim.

> "honest open source advocates"

You've literally just made this term up. It's meaningless.

Re: Open source AI is the path forward

#688

Post by WithinReason »

stavros wrote: post #425
causal wrote: post #41

Earlier quoted context omitted.

"Open weights" is a more appropriate term but I'll point out that these weights are also largely inscrutable to the people with the code that trained it. And for licensing reasons, the datasets may not be possible to share. There is still a lot of modifying you can do with a set of weights, and they make great foundations for new stuff, but yeah we may never see a competitive model that's 100% buildable at home. Edit…

"Open weights" means you can use the weights for free (as in beer). "Open source" means you get the training dataset and the methodology. ~Nobody does open source LLMs.

If weights are not the source, then if they gave you the training data and scripts but not the weights, would that be "open source"?

Re: Open source AI is the path forward

#689

Post by Phelinofist »

llm_trw wrote: post #601
saurik wrote: post #598

Earlier quoted context omitted.

Then what is the "source"? If we are to use the term "source" then what does that mean here, as distinct from it merely being free?

It means nothing because LLMs aren't software.

Do they not run on a computer?

Re: Open source AI is the path forward

#690

Post by exe34 »

jononor wrote: post #484
JimDabell wrote: post #205

Earlier quoted context omitted.

I don’t think it’s that simple. The source is “the preferred form of the work for making modifications to it” (to use the GPL’s wording). For an LLM, that’s not the training data. That’s the model itself. You don’t make changes to an LLM by going back to the training data and making changes to it, then re-running the training. You update the model itself with more training data. You can’t even use the training code a…

No the preferred way to make modifications is using the the training code. One may also input a snapshot weighs to start from, but the training code is definitely what you would modify to make a change.

how do you train it in a different language by changing the training code?
Post reply on HN