Hacker News Reader

Comment by ensocode | Hacker News Reader

What are your first impressions using it? (Not available in Europe currently). Is it a game-changer?

>Not available in Europe currently

I hate how its the same story for every new AI technology. If someone can tell me who to vote for or where to protest to change this awful EU law, that would be great.

It has nothing to do with the EU laws, or at least not in the sense they want you to think about it: no laws prevents AI players to release their AI models here, but they are all also big tech players who are affected by GDPR, DSA and DMA which harm their business by protecting the consumers.

That's why they use their AI products as a leverage to turn European people against the laws that protect them from big tech. It's just blackmail.

No, it has nothing to do with the GDPR or DMA. It is due to the AI Act.

https://artificialintelligenceact.eu/wp-content/uploads/2021...

https://en.wikipedia.org/wiki/Artificial_Intelligence_Act

The AI Act doesn't prevent Grok from releasing their model in the EU! (And ChatGPT early issues were all linked to GDPR)

For the record, Facebook has put a restriction on the use of Llama models in the UE even before the AI Act was passed (and the AI Act doesn't even apply to Llama anyway, except Llama 3.1 405b)

It will come to EU countries eventually, but it takes a long time to go through "conformity assessments". Notebook LM, for example, was geoblocked for the EU for a full year before it became available in June. Grok 1 was released everywhere else in the world in November 2023, and in the EU in May 2024. About a 6 month delay.

> Grok 1 was released everywhere else in the world in November 2023, and in the EU in May 2024. About a 6 month delay.

And here you should see that it has nothing to do with the AI act, as it wasn't enacted before last August!

Furthermore, neither Grok 1 nor Notebook LM would have been subject to the AI act even if it had existed at the time.

As I said before, all of these companies have vested interests against EU's legislation as a whole, and they've tried to blackmail the EU from the beginning. They didn't wait for an actual AI legislation to exist to use AI as just another blackmailing tool.

Apparently EU regulators were blocking it for unspecified reasons until an agreement was made in May. And even then, they blocked news summaries until after the EU elections. If you can find more info, feel free to cite it. Info about these behind-the-scenes dealings are hard to find online.

https://www.socialmediatoday.com/news/xs-formally-twitter-gr...

The regulation that the author of this article think about is DSA, even if not named directly. See this quote:

> Well, probably because Grok has already spread various false stories

(The European regulation that deals with disinformation is DSA).

And again it couldn't be the AI Act, because it wasn't in place at that time!

I think you're misapplying the term "blackmail" here and thus poisoning the well. The EU is applying pressure to companies and companies are applying pressure back--that's not blackmail. They each have their own means of leverage, and they both use them.

First of all both sides don't have the same level of legitimacy and then one side is blatantly lying about its intent by claiming that they are blocked by regulations instead of admitting that they are putting pressure (because they know they have no legitimacy to “apply pressure” on democratic institutions).

totally agree. And this is one of the reasons the EU is falling more and more behind, all the silly regulations.

The EU regulations are there to protect the average citizen, not to help the 1% run wild with whatever business idea they have. You personally might not like it, but the non-entrepreneurs, which is most people, are pretty satisfied that the laws in the EU are more focused on the citizen and their rights and not on boosting the shareholders' profits.

How exactly am I being protected by not having access to the latest models that the rest of the world has?

Nothing is stopping X from complying with EU regulations to make it available to you. I'd wager that they most likely lack compute capacity to make it available everywhere, not legal compliance.

> Nothing is stopping X from complying with EU regulations to make it available to you.

Given the personalities involved, I'd wager he's doing it out of spite than for any actual legal justification.

...though if there was an actual legal risk then I'll agree the economics probably don't bear-out the risks - as someone who identifies as European I'll admit that Europeans generally pay far less money for tech/software/apps/services than Americans do[1]; salaries in Europe are also significantly below their US equivalents: paying $200/mo for OpenAI's service is a much harder sell to a London or Berlin-based SWE on €90k/yr than when you're a Bay Area type on $300k/yr.

[1] e.g. If you can take Apple at their word, the EU accounts for 7% of their App Store revenue: https://techcrunch.com/2024/02/01/apple-says-eu-represents-7...

Besides, anyone in the EU who really wanted to use it can just use a VPN service.

Libel, from all the models hallucinating things done by whatever your real name is.

I mean, at least I get the advantage of being overshadowed by a famous film director with the same name as me, so nobody's going to assume anything associated with my name is actually about me…

…hopefully…

> not to help the 1% run wild with whatever business idea they have.

And IMHO regulating 1% doesn't hinder strategic advantages much. Otherwise China would not have came up with DeepSeek models. Regulations are fine, they just have to be "based".

> Regulations are fine, they just have to be "based".

What? I know what "based" means, but I can't quite grok what you're saying.

I'm increasingly happy we have these regulations that prevent us from being ruled by the likes of Musk.

Why? Europe is getting extremely expensive on all levels. Time for a Musk there.

I rather spend life in a poor democracy than a rich technocracy which supports dictators around the globe.

> for every new AI technology.

Well no. Mistral.ai

Companies need to adhere to GDPR in order to enter the European market, poeple have the right to request to delete their PII. It's a good law, actually this should be applied everywhere. As an European, though, I'm scared: what if companies are actually testing if excluding us is so much of a problem for their business?

They are doing it. At some point we have to agree that Brussels are idiots and rarely savants.

I would be very surprised if they aren't monitoring the cost-benefit curve of delaying EU launches. Why wouldn't they? It costs extra money, time, and legal risk to launch in the EU. It's especially bad for XAI due to Musk's involvement.

(Note that it's not just GDPR, there's also the EU AI Act which has a whole extra set of requirements for compliance.)

Have you tried asking the AI people to ship an AI that complies with EU law?

EU regulation are hit and miss. USB C and Opening up apple store are hits. AI regulation and cookie banners and idiotic bottle caps are miss.

I don't get the bottle cap hate meme. Is it useful? Probably no, but the amount of hate it gets is unexplainable by lack of concrete usefulness alone…

Also, the problem with GDPR is that it wanted to leave too much room for business to still collect an obscene amount of data, hence it allows the cookie banner. Please note that I emphasized “allow” because that's all GDPR does: it allows companies to use a cookie banner to extract the consent to collect data. It doesn't mandate it in any way.

None of my multiple websites have a cookie banner on them because I'm not trying to extract consent from my users to abuse their data, I just don't collect it and I'm effortlessly GDPR-compliant in the least obnoxious way.

Cookie banners are just malicious compliance.

You could move to America and avoid the fake delays blaming the regulations tech companies don't like.

It's not an awful law.

The Digital Market Act is a bit of an overreach but the AI law is not.

It classifies AI into risk categories, so that it doesn't kill anyone, carelessly handle sensitive information, etc.

A chatbot can easily comply with it.

No, it was underwhelming, failing basic coding tasks, OCR/Image recognitions that none of the other existing models screw up.

Reply on news.ycombinator.com