My brother in Lemmy, this is what stopping Big Tech looks like.
Europe made laws that say that Google and others need to pay if they want to link to EU publishers. Well, maybe the price they are asking is not worth it.
You’re right about the firewall energy, but that’s simply how these laws work. The point of copyright, as well as age verification and other such laws, is to control who may access certain information.
Because that’s one key feature in the “2019 European directive adopted into French law”. It’s also what the Google fine was about.
Also, X isn’t really suitable for copy/pasting entire articles, like is done on lemmy. So that’s probably not it.
Misunderstanding. The news media is suing X. I pointed out what news media does without paying.
You think people should pay X to link to tweets? Or generally for quotes?
More like: They want to sell the cake and be paid when you recommend it to others.
Mind that news media don’t pay when they link to social media, quote people, or even report what other media has reported. The real question is, if this law has any beneficial effect for society. I don’t see how.
Oh, he’s saying that snippet view lets us have sites like lemmy. I didn’t get how cracking down on that would help lemmy.
It should only show the title and the link imo.
That’s infringement in Europe, which makes it effectively a link tax.
Huh? How you mean?
Not quite. You can’t turn movies into books or games, or vice versa, for example. Sometimes such projects get stuck in limbo. Or think about how everyone hated the final season of Game of Thrones. Can’t do anything about that in our life times.
The “battle” is the result of copyright people trying to use open source people for their ends.
In the past, for software, the focus was completely on the terms of the license. If you look at OSI’s new definition, you will find no mention of that, despite the fact that common licenses in the AI world are not in line with traditional standards. The big focus is data, because that is what copyright people care about. AI trainers are supposed to provide extensive documentation on training data. That’s exactly the same demand that the copyright lobby managed to get into the european AI Act. They will use that to sue people for piracy.
Of course, what the copyright people really want is free money. They’re spreading the myth that training data is like source code and training like compiling. That may seem like a harmless, flawed analogy. But the implication is that the people who work and pay to do open source AI have actually done nothing except piracy. If they can convince judges or politicians who don’t understand the implications then this may cause a lot of damage.
Other way around. The NNs are written in, mostly, Python. The frameworks, mainly Pytorch now, handle the heavy-duty math.
Musk also has his hands on a fair amount of data through X and Tesla. But yeah… Copyright expansion seems like an odd place to start breaking the constitution.
You’re right about the regulation but I’m not so sure about the copyright exemptions. All in all, you’d think he’s more likely to side with property owners - especially the heirs of media empires - over progress.
After a quick skim, seems like the article has lots of errors. Molmo is trained on top of Qwen. The smallest ones are trained on something by the same company as Molmo.
The FTC under Biden has begun to push back against tech monopolies.
Maybe you could call it recoupment but it doesn’t have quite the same ring. It’s not quite the same thing, either.
You could also talk about coercive monopolies but that doesn’t mean exactly the same thing.
If the same user can generate the same input, it will result in the same hash.
Yes, if. I don’t know if you can guarantee that. It’s all fun and games as long as you’re doing English. In other languages, you get characters that can be encoded in more than 1 way. User at home has a localized keyboard with a dedicated key for such a character. User travels across the border and has a different language keyboard and uses a different way to create the character. Euro problems.
https://en.wikipedia.org/wiki/Unicode_equivalence
Byte length of the character is irrelevant as long as you’re not doing something ridiculous like intentionally parsing your input in binary and blithely assuming that every character must be 8 bits in length.
There is always some son-of-a-bitch who doesn’t get the word.
You should accept Unicode; if doing so, you must count each code as one char.
Hmm. I wonder about this one. Different ways to encode the same character. Different ways to calculate the length. No obvious max byte size.
I think, for a lot of people, technology has come to mean a few websites, or companies.
There are a few lemmy communities dedicated to AI, but they are very inactive. Basically, I’d have to send you to Reddit.
Hmm. It was a big issue at the time. In truth, I’m really not sure how it works in France. Anyway, the big fight going on is really about minimal previews. Unfortunately, there is no disinterested reporting on the issue. The media is very much profit-maximizing.
The recitals aren’t part of the law, but should only guide the interpretation. Also, this is a directive. That means it directs the member states to make law, but has no direct effect, as such.