50 comments

[ 3.5 ms ] story [ 161 ms ] thread
That bluff deserves to be called.
In both directions
I'm perfectly fine with that. It's not as if OpenAI is going to be the sole supplier even if they have a head start today.
These models are useful enough that people are fine accessing them through VPNs. It doesn’t seem like a big deal to geoblock the EU.
The problem is when third-party (to OpenAI, and I include Microsoft here, for the purposes if this comment) companies try to build on top of this product - they will not be able to launch products on top of it in the EU either (unless they do a weird user-initiated transfer of data to the US, process it, and return the result.

Even then, the EU has been experimenting with trying to claim sovereignty over income from outside the EU for non-EU-based multi-nationals, so they could concert ivably target this "workaround". It remains to be seen whether the USA will attempt to persuade them to remember how Westphalian sovereignty works, or I'd the US government decides to match them and start targeting foreign companies' foreign turnover in applying remedies.

All I want is for the US to vassalize EU, not because I'm against the EU (I am), but because it'll be funny. If by economic growth rates from 2000 onwards, I think that could be achievable a few decades after I'm dead.

I wonder if anyone would like to make a long term bet with me

Currently they're the sole supplier of GPT-4 quality models, and the only competitors on the horizon are also US startups to whom the same regulations would apply. It's perfectly possible that given a choice between spending mental bandwidth on competing with OpenAI in the USA or dicking around with trying to please un-pleasable Eurocrats, they'll all decide to punt the EU to later.

I've been warning about this on various places on the net since the start of GDPR. It's never been a popular message, but the idea you can just repeatedly and indefinitely levy huge fines on US corps without eventual blowback was never sensible. Whilst the class of the 2000-2010 era, the Googles and Facebooks, might not leave due to large employee bases there, the class of the 2020s are a different matter entirely. It's not even just about access: do you think OpenAI will be in any hurry to hire people in any EU country now?

I just hope if/when that happens they remember that not every country in Europe is in the EU. Hopefully the UK, Switzerland, Norway etc are kept properly separated.

If the wheel was invented today we'd ban it before it put sedan chair workers at risk.
I’m wondering if OpenAI will even be still relevant once this legislation is in place.

I’m not convinced that they won’t. But I think it’s also not a certainty anymore, with all these amazing breakthroughs coming out every week or so.

Sounds like his talking backfired.

"We need to regulate AI! No, not that much!"

There's no way OpenAI will quit EU and leave it for competitors to take, easy bluff to call. The whole thing makes it seem like the desire for AI-related regulation isn't ethics motivated and is instead only trying to make it harder for competition to start.

Precisely, it's a very transparent attempt at regulatory capture. Utterly disgusting.
Bard isn’t available in the EU or Canada, so there’s one competitor they don’t need to worry about. That doesn’t seem to be a bluff?

Perhaps people will learn to use VPN’s like some US folks did for cryptocurrency exchanges? Or maybe it will be a boost for an open source project that gets a reputation for reliability. (Well, reliable for an AI chatbot, which is not very reliable by most standards.)

For all of Microsoft's flaws, they typically are good at working with regulators and ticking the boxes needed to get into a market like EU. It's unsurprising that Bard isn't there yet, as the whole field has moved at "faster than large corp/government" speeds but I don't imagine MS will stay out of those markets for long.
What exactly does "quit EU" mean. They make an app that connects over the internet. Anyone including businesses, are free to use their money how they see fit by paying for a service that runs in a another country.

All the EU can do here is get rid of the business and infrastructure presence of this company in the EU. Getting rid of jobs and tax revenue. They can also do a China and prosecute their own citizens for trying to access things they want on the internet.

There doesn't need to be Any "AI-related regulation." There's nothing specific or magical or new about "AI" that's not covered by current law. What it does look like to me, is the government couldn't press some authoritarian laws, so they slapped "AI" on it to mislead and misdirect their voters.

Nothing backfired, it wasn't a shot or a bluff. It was a statement that if they pass draconian laws, their citizens will have to buy the product online with their visa card without paying VAT.

Funnily enough, the EU guy who is leading this and doubling down, is Romanian. I've had some friends from that country and been there a few times. It's not a place I would think of when I think honest regulatory body, ethics, or bleeding edge technology. It definitely is a place I think of when I imagine a peacock puffing its feathers, a tiny monkey hitting it's chest, and a drunk idiot with low IQ doubling down till the grave.

> Funnily enough, the EU guy who is leading this and doubling down, is Romanian.

What is the relevance of your personal (and poorly informed) opinions about Romanians? It’s a needlessly low blow and it’s tiring to hear the same myths perpetuated by people who have zero context on Romanian culture and life. If you do not associate Romania with honesty, morality, and innovation, then it’s because you have been misinformed.

> Anyone including businesses, are free to use their money how they see fit by paying for a service that runs in a another country.

I don’t think that has ever been true in any legal jurisdiction. Every country has restrictions on trade and payments that can be made when acquiring services domestically or abroad.

What is the relevance of my personal experience based opinion on the thing I am talking about? That's how opinions work. We have experiences, we get information, we give our opinions based on those.

Now what we could do, instead of giving our own opinions, is repeat statistics that anyone could look up in a minute. Since I am a person, not a database, I gave opinion based on experience.

Here's the statistic though for people with smooth brains. Look for Romania - it's towards the bottom there, near Hungary. You know Hungary, the country with the dictator, who supports that other country with the dictator. https://www.statista.com/statistics/873736/corruption-percep...

>Every country has restrictions on trade and payments

Right. As I literally stated. Every country has the right to prosecute their own citizens in their country. I'm sure that's what the EU voters want - to be prosecuted under the law proposed by this Romanian lawmaker.

There's nothing wrong with every person in Romania. But when a country is known for corruption, perhaps they shouldn't be leading the charge in EU-wide laws.

Let me rephrase that, since you're going to double down and do the "purposefully dense thing." I propose Poland lead EU-wide legislation on Abortion.

You got a problem with that? What, you got something against Poland?

In EU it's not possible to develop a viable competitor due to overregulation. Where's EU's Google? Where's EU's Starlink? Where's EU's Microsoft? Where's EU's Samsung?
Some would view the development of megacorps and monopolies a negative, rather than a positive...
For those, a world without a desktop computer, smartphone, Starlink constellation and internet search is better. The downvoting illustrates that it's a prevailing viewpoint on HN now, indeed.

"What the honorable member is saying is that he would rather that the poor were poorer, provided that the rich were less rich" - Margaret Thatcher

Trillion dollar corporations that pay taxes and provide jobs are negative things?
In many cases, evidently, yes
(comment deleted)
Google has threatened and left jurisdictions. It's not an easy bluff to call. If the regulation makes it practically impossible to deliver a good product, why wouldn't they leave?
If EU rules are too draconian, neither OpenAI nor their competitors has anything to gain from the EU. The only bluff to call would be if OpenAI were pretending they're too draconian when they're really not. I'll believe that bluff when EU compliant competitors popup who are on par with openAI. Thus far it hasn't happened.
Google Bard doesn't operate in the EU so your theory is unfounded.
Yet. If OpenAI quits the EU, the EU market would be for Google and others to take.
It's not a bluff. Silicon Valley CEO's hate european policies and are willing to remove their stuff because the Euro's homegrown tech startups are atrocious and everyone knows it.
One can only hope. What was he even doing talking to so many government leaders? Did he imagine they’d all fall for cheap fud and embrace copyright theft overnight?
Slow down progress for 5.8 billion people so that 400 million feel protected.
It's people that goes bankrupt for some medical procedure, that makes their kids get shot at school for some freedom fetish, and with a insane wealth hunting culture, as a european I would hardly care about the needs of americans or the rest of child employing wastelands
EU regulations are basically a full employment program for VPN engineers.
VPNs will be illegal in the US soon
> VPNs will be illegal in the US soon

No, they won't.

What makes you say that ?
Does OpenAI have all the technology to implement the EU AI regulations?
Altman seems inexperienced for this type of role. I guess he didn't really expect OpenAI to be the leader of the AI race.

Hopefully strong competition comes along soon.

and hopefully unlike openAI it really is open this time.
It's easy to say that, but who would make a better candidate? It's a very weird and unique situation he/they are in. He just looks young, he's been around the block a couple times.
Agreed. I hope there is competition so there's not just one person, alone.
Do you understand that NN is not like rows in database. The machine can't unlearn what it has learned. You can't make AI unlearn something. They are going to ingest large amounts of data and trying to stop it is going to be nothing more but layering endless filter in front of it. Unfortunately or fortunately, it can be prompt broken. So if they try to abide and still fail, they will be fined enormously. See Meta & Google for example. This is about leverage. OpenAI has leverage and I'm happy to see a leader have balls. Let EU go back into the stone age if that's their wish.
So lets just change laws to accommodate an unethical data model from a corporation pushing techno feudalism?
(comment deleted)
OpenAI could roll back to a previous version of their model.

Anyway, I'm not sure if you intended that as a reply to me because it's not relevant.

How? Do you think they have a previous model that's missing one single data point for anything that needs to be forgotten? The only way this would work would be to filter out whatever the EU demands then retrain the model without that data. So every time they get a request, they will have to retrain the model. Scale that.
I don't care how they scale it. They obviously have versions of the model and can go back to a previous version.
> The machine can't unlearn what it has learned. You can't make AI unlearn something

This is not even remotely true: From https://en.m.wikipedia.org/wiki/Catastrophic_interference:

> Catastrophic interference, also known as catastrophic forgetting, is the tendency of an artificial neural network to abruptly and drastically forget previously learned information upon learning new information.

South Korea regulates porn sites, but the regulation shrinks the alternatives, and traffic using VPNs is maintained. Sounds similar