>Adding even more intellectual property nonsense isn't going to work.
[citation needed]
Where does your confidence come from?
GPL itself was precisely the "intellectual property nonsense" adding which made FOSS (free as in freedom) software possible.
The copyright law was awfully broken in the 1980s too. Adding "nonsense" then was the only solution that proved viable.
Historically, nothing but adding "more IP nonsense" has ever worked.
>The real solution is to force AI companies to open up their models to all.
Sure. Pray tell how you would do that without some "intellectual property nonsense".
We don't exactly get to hold Sam Altman at gunpoint to dictate our terms.
>We need free as in freedom LLMs that we can run locally on our own computers
Oh, on that note.
LLMs take a fuckton of compute to train and to even run.
Even if all models were open, we're not at the point where it would create an equal playing field.
My home computer and my dev machine at work have the same specs. But I don't have a compute farm to run a ChatGPT on.
> Where does your confidence come from?
From the fact that copyright infringement is trivial and done at massive scales by pretty much everyone on a daily basis without people even realizing it. You infringe copyright every time you download a picture off of a website. You infringe copyright every time you share it with a friend. Everybody does stuff like this every single day. Nobody cares. It is natural.
> GPL itself was precisely the "intellectual property nonsense"
Yes. In response to copyright protection being extended towards software. It's a legal hack, nothing more. The ideal situation would have been to have no copyright to begin with. The corporation can copy your code but you can copy theirs too. Fair.
> Pray tell how you would do that without some "intellectual property nonsense".
Intellectual property is irrelevant to AI companies.
Intellectual property is built on top of a fundamental delusion: the idea that you can publish information and simultaneously control what people do with it. It's quite simply delusional to believe you can control what people do with information once it's out there and circulating. The tyranny required to implement this amounts to totalitarian dictatorships.
If you want to control information, then your only hope is to not publish it. Like cryptographic keys, the ideal situation is the one where only a single copy of the information exists in the entire universe.
AI companies are not publishing any information. They are keeping their models secret, under lock and key. They need exactly zero intellectual property protection. In fact such protections have negative value to them since it restricts the training of their models.
> We don't exactly get to hold Sam Altman at gunpoint to dictate our terms.
Sure you do. The whole point of government is to do just that. Literally pass some kind of law that forces the corporations to publish the model weights. And if the government refuses to do it, people can always rise up.
> Even if all models were open, we're not at the point where it would create an equal playing field.
Hopefully we will be, in the future.
So... People are going to rise up? What makes you think most of them have enough slack in their finances to pack up and haul off to D.C.? Only the Elites do, and they pay full time lobbyists to do exactly that to make sure laws like you mention never pass. Not saying it can't work. Just saying it the game is rigged against the very people you want to rise up and in favor of the ones who'd rather you stayed in bed.
If people don't rise up they will become soylent green. Over the long term, AI threatens to replace all human labor. It cannot remain locked away in corporate servers. This is an existential issue. The ultimate logic of capitalism is that unproductive people need not be kept alive since they add nothing but cost. So either we free AI, collapse the very idea of having an economy and transcend capitalism into a post-scarcity society, or we will be enslaved and genocided by those who control the AIs.