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

Can Europe train a frontier AI model on the compute it owns?

github.com

Post reply on HN
Print view View original

161–170 of 313 posts

Re: Can Europe train a frontier AI model on the compute it owns?

#161

Post by noddybear »

poszlem wrote: post #138
rootlocus wrote: post #38

Earlier quoted context omitted.

We can debate the details and implementation but EU legislature is, at least in spirit, trying to protect human rights. What does it say about progress that the same laws that protect human rights also stifle innovation?

That's a bit like arguing that the USSR was, in spirit, trying to defend workers' rights, and therefore we should not have opposed it. At some point, the gap between what something claims to be "in spirit" and what it actually is in practice becomes too large to ignore.

Right, the purpose of a system is what it does.

Re: Can Europe train a frontier AI model on the compute it owns?

#162

Post by k__ »

michaelt wrote: post #119
aspenmartin wrote: post #54

Earlier quoted context omitted.

You can do what frontier labs do today which is to properly license things that are copyrighted and use open source web crawls for things that don’t have copyright issues. You can then also commission new datasets (volume needed goes down when quality is high). The European regulations are the thing that will kneecap anything meaningful coming out of Europe. Mind blowing to me that this is worth the tradeoff since Eu…

> You can do what frontier labs do today which is to properly license things that are copyrighted and use open source web crawls for things that don’t have copyright issues. You can then also commission new datasets (volume needed goes down when quality is high). It cost Anthropic $1.5 billion for training on libgen's 480k pirated ebooks. Investors will cough up that money if you're already clearly a frontier lab wit…

I thought the joke was that people aren't paying enough money.

Re: Can Europe train a frontier AI model on the compute it owns?

#163

Post by fancyfredbot »

Forget regulations, would it make sense for Europe to train a frontier model of it's own? Would it be sufficiently better than fine tuning a Chinese model? Would it actually be competitive with US frontier models? Would enough people pay to use it even within Europe to pay for the training costs? Do we have enough inference capacity that enough people /could/ use it? Would being "European" allow any governments in Europe to trust it, rather than deciding that actually there needs to be a French, German, Italian, Spanish and UK sovereign AI?

I am guessing that enough of these questions can be answered with "no" that nobody really wants to invest.

For the same reason there isn't really a serious third start up competitor to OpenAI and Anthropic.

Re: Can Europe train a frontier AI model on the compute it owns?

#164

Post by noddybear »

pptr wrote: post #74
LastTrain wrote: post #61

Earlier quoted context omitted.

Yep. I wonder if being at the forefront of AI makes for a better & more free society.

You can only make money if you produce value. If you stop producing value because your old industries decline/collapse and you miss out on new industries, your wealth will trend towards that of a third world country.

Maybe. I would contend that wealth accumulation over the long term is a function of the strength of local institutions e.g educational, judiciary etc. This is effectively the thesis of ‘Why nations fail’. Europes relative decline will only continue if the institutions of the US (and others) continue to be robust and healthy. I am not sure if this can be relied on into the future.

Re: Can Europe train a frontier AI model on the compute it owns?

#165

Post by dofm »

TMWNN wrote: post #156
ben_w wrote: post #132

Earlier quoted context omitted.

> No EU project for its own "Silicon Valley" can succeed, because the French would insist on it being in France, Germans would insist on Germany, Dutch in the Netherlands, etc. France and Germany and the Netherlands may compete with each other to attract VC and tech firms, but they're ultimately no more capable of preventing each other from getting a runaway success loop than New York and Utah were at preventing Cali…

>France and Germany and the Netherlands may compete with each other to attract VC and tech firms, but they're ultimately no more capable of preventing each other from getting a runaway success loop than New York and Utah were at preventing California from getting actual Silicon Valley. Yes, but why hasn't said success loop occurred anywhere else yet? Silicon Valley has been a "thing" for five decades now. > The EU as…

> Yes, but why hasn't said success loop occurred anywhere else yet? Silicon Valley has been a "thing" for five decades now.

So have twenty-three Superfund sites —- land from 1970s businesses that the federal government had to take over because nobody else could or would fix the uncosted externality.

The most in any single county.

An increasingly useful, very vivid metaphor.

Re: Can Europe train a frontier AI model on the compute it owns?

#166

Post by ben_w »

TMWNN wrote: post #156
ben_w wrote: post #132

Earlier quoted context omitted.

> No EU project for its own "Silicon Valley" can succeed, because the French would insist on it being in France, Germans would insist on Germany, Dutch in the Netherlands, etc. France and Germany and the Netherlands may compete with each other to attract VC and tech firms, but they're ultimately no more capable of preventing each other from getting a runaway success loop than New York and Utah were at preventing Cali…

>France and Germany and the Netherlands may compete with each other to attract VC and tech firms, but they're ultimately no more capable of preventing each other from getting a runaway success loop than New York and Utah were at preventing California from getting actual Silicon Valley. Yes, but why hasn't said success loop occurred anywhere else yet? Silicon Valley has been a "thing" for five decades now. > The EU as…

> Yes, but why hasn't said success loop occurred anywhere else yet? Silicon Valley has been a "thing" for five decades now.

Same reasons (plural) it also hasn't happened a second time anywhere else in the USA.

The list is long, and economics is full of anti-inductive loops.

> But that is going to require a level of support that is probably beyond the scale of one national government.

The EU as an institution is tiny in comparison to its member states, total budget only €192.8 billion: https://eur-lex.europa.eu/EN/legal-content/summary/2026-euro...

> The EU can trace its origins back to France and Germany agreeing to combine its coal and steel.

While true, that's like saying the US can trace its origins back to some cold salty tea: it misses quite a lot of both the good and the bad.

Re: Can Europe train a frontier AI model on the compute it owns?

#167

Post by noddybear »

sajithdilshan wrote: post #147
mnewme wrote: post #94

Earlier quoted context omitted.

Bullshit. The AI Act is not the problem, Mistral models are definitely usable and some were/are competitive. European future looks bright compared to the political landscape the US has now

Political landscape of the US right now will change in few years like it has been changing for the last 250 years. USA does not equate to its current political situation. But EU would keep on rotting from inside and I would eat my hat if Mistral doesn’t go out of business in next 5 years

It’s an interesting contrast that you claim that the USA does not equate to its current political situation, whereas conversely the EU is ‘rotting’ from the inside. Do you think this is a balanced and reasonable claim?

Re: Can Europe train a frontier AI model on the compute it owns?

#168

Post by signatoremo »

fxtentacle wrote: post #146
beernet wrote: post #31

Absolutely not, even if Europe was given all of the compute in the world. The issues are much much worse. Starting at "data privacy", continuing with the "EU AI Act" and just an overall mindset of regulation and German angst. Many structures in the EU actively prevent and fight against innovation. Oftentimes more subtle but the consequences stay the same. If you ever had to use one of the "sovereign" AI provides such…

The thing that nobody talks about is how many EU citizens already are working on frontier models, just inside US companies. That’s why the prohibition was so crippling for Anthropic. And, btw, the bang per buck that I got from OVH was better than EC2.

> The thing that nobody talks about is how many EU citizens already are working on frontier models, just inside US companies

Why don't you ask yourself: how come those brilliant European minds couldn't find a job that pays well enough at home? Why could they in the US? There are many more Chinese and Indian, and other internationals working for US companies, outnumbering Europeans. It's not like Europe was intentionally targeted by US companies.

Real GDP has grown 84% in the US since 2000 [0]. EU grown 40-45% in the same span. The two regions were basically had the same economic output in 2000. Europe has been left behind economically.

Living in Europe is nice (I love Europe by the way), but the question is why EU can't compete with the US and increasingly China. Sooner or later that'd affect your living standard as well. Look at how much China has caught up with the West in terms of quality of life.

[0] - https://fred.stlouisfed.org/series/GDPC1

Re: Can Europe train a frontier AI model on the compute it owns?

#169

Post by red75prime »

mopsi wrote: post #160
red75prime wrote: post #143

Earlier quoted context omitted.

Why shouldn't I? Your garden is not of a particular interest to me. It's just one of millions of gardens that I might look at as a part of scenery, or to get my bearings if I happen to be in the area. You'd be better off fighting street views in general, unless you are OK with the Streisand effect.

I prefer not to have views of my home permanently archived and made available to anyone in the world, unless they can present a reasonable need for it. City planners? Go ahead. Local people for navigation purposes? Go ahead. But some random bloke from another continent? That's clearly too far. Services like Street View should have distance-based friction to preserve privacy. The further you are, the less (or at lower…

OK, feel free to ruin experience for millions of people.

Re: Can Europe train a frontier AI model on the compute it owns?

#170

Post by sajithdilshan »

sumeno wrote: post #150
sajithdilshan wrote: post #144

Earlier quoted context omitted.

Exactly, it’s quite funny that everyone equate US and US legal system to Trump. The founding fathers created a constitution that can whit-stand and survive people like Trump and still the Republic would thrive. Trump would be gone in few years but US would still be there like it has been for the past 250 years for the people by the people. On the other hand EU started as an economic union and has rotten into a behemo…

The founding fathers created a document that was already struggling with modern realities prior to Trump. 250 years is not a particularly impressive amount of time for a country to not fall apart.

Care to elaborate on concrete examples on where it struggled? 250 years is quite impressive even if you don't believe it or not because only a handful of countries in the whole world has an older constitution.
Post reply on HN