[sajithdilshan]

This is a wake up call for EU, but it’s already too late IMO. ChatGPT was launched in 2022 and since then only company in EU which has released a model that is even closer to a frontiers is Mistral.

With strong data protection, copyright and privacy laws, it would be a nightmare for any company to train models because activists lawyers would sue left and right.

Even by some miracle they come up with the model, again the hardware would be from US companies. It would take years to build a European equivalent of AWS and not to mention the talent required to do that. Given how low the wages in European countries are compared to American counterparts, and also lack of incentives from the governments I truly cannot think of a way how EU can catch up to US or China. Although I would love to be proven wrong

[alecco]

You forgot no energy and deep resentment for anybody earning more than 85k/y (except politicians).

Worst part is Europe was near the top of nuclear and microchips ~30 years ago, and France had bleeding-edge AI.

[joe_mamba]

>Worst part is Europe was near the top of nuclear and microchips ~30 years ago

Boomers decided that was enough growth, it's time to cement what they got and cash out without thinking about economic growth opportunities of future generations.

>France had bleeding-edge AI

They still do, but it's all tied in stuffy bureaucratic French-speaking academia, not in monetizable products that scale internationally. Whatever they come up with, the US companies will then buy up, turn into products and sell for money.

[gls2ro]

Not sure how to phrase this and it matters where are you living - in Europe or elsewhere - but I think when looking at things like EU or other military or economical alliances we should not use terms "it is already too late". You dont know that until the current era is over so I think if you want to think deeply about this you should use longer threads with more nuances.

[Stromgren]

Why _too_ late? Late, yes, but you’re implying that some line was crossed.

[sajithdilshan]

Yes, the time to take action has been 10 years ago. In order to catch up with US and China, EU would have to accelerate assuming US and China would stagnate(which is not the case). At current rate EU would always be catching up. I would love to hear if you can think of any plan on how EU can overtake US and China on AI advancement

[tzs]

Aren't you assuming that current approaches to AI won't plateau?

There's only so far you can push things before you need breakthroughs that give new approaches and when that happens it is often a new playing field.

[KetoManx64]

Motors and cars in general have plateud for decades now, yet I don't see any new car manufacturers springing up in the EU.

[fvdessen]

These stories about regulation preventing EU frontier models are frankly complete bullshit. The real reason is much simpler, but also harder to fix.

To build frontier models you need VC money. There’s no VC money because VC believe that there is no market for a ‘EU Champion’.

There’s no market for a EU champion because internal EU market is not big enough for VC returns. Why invest in EU champion when the US champion is guaranteed to have better returns ?

And there’s no public alternative to VC either because that’s national level and national investment in EU doesnt cross national boundaries

Mistral actions reflect this, they need returns and they target the market where they can be competitive, which is the scraps the US labs cannot address. This is not enough to fund frontier lab research

Also the legal context on regulations is quite different from the US. In the US you can have unlimited damage, that is not the case in the EU, where regulation penalty can never as a matter of principle put the existence of the company in danger, and thus the application of the regulations is always a matter of negociation with the government. You don't have to respect everything all at once, size of the company and ability to actually implement the regulations are taken into account, which means that sartups are usually excempt.

[root-parent]

In most EU countries if your company went bankrupt, you are not allowed to open another one. Think about that for a little bit.

The Southern Countries are parasitically living from EU funds and EU programs, including money transfer for the budgets. The rich countries are desperately seeing China eating their lunch in Cars and all the rest.

Add to this the aging population, and now having the head under the sand on AI, and it does not look pretty...

[fvdessen]

> In most EU countries if your company went bankrupt, you are not allowed to open another one. Think about that for a little bit.

That's just completely false but ok. There's even a EU regulation to ensure the exact opposite: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=legissum...

[root-parent]

You dont have a clue, and are showing EU regulations made by funcionaires that most countries simply ignore.

Here is an example of the current situation in Luxembourg for example, where the EU Court is based! They want you to clean up your debts first, well you would not go bankrupt if you could no? And if you look at the remaining conditions for the so called second chance they are impossible to achieve.

https://sstlaw.lu/obtaining-a-business-license-after-facing-...

In the Netherlands restarting after bankruptcy is possible, but not automatic. The bankruptcy trustee determines whether the business can restart, and restart may be impossible if it harms creditors...

Spain excludes or limits discharge for some public debts and certain debtors...

Ireland and Austria completely ignored the law you quoted and currently have infringement procedures...see you in a few years...

Poland did not apply at all the law you mentioned, and had been criticized by the EU but nothing was made about it until now...

There are differences between being on the field, or knowing how to do a google search...

[gls2ro]

Living in EU so I am biased but I agree with this. You have to clean your debt or real people suffer.

Here is a real example: open a company, buy a small piece of land, section it in smaller pieces, promise you build houses, get people downpayments in cash, syphon all money to other companies with having contracts just a bit bigger than average market price, declare bankruptcy, then start again. You of course have smaller stakes in those other companies.

[root-parent]

I would reply to you but any comment I post gets flagged within less than one 1 sec...I can only imagine AI took over moderation here...and its failing miserably...

[dgellow]

Regarding AWS, we have Gcore, OVH, Hetzner

[alde]

You are comparing cars to bicycles.

[tzs]

Are you referring to the services offered or just to the number of servers available?

[lostmsu]

Is there a difference? Does Hetzner offer an alternative to Bedrock (3yo, came out basically same year)?

[tzs]

Unless I made a wrong turn upthread, I think the context was alternatives to US cloud services like AWS for organizations that are trying to develop their own frontier or near frontier models.

My understanding is that Bedrock is for those who want to use existing models. It lets you access models from various companies via a unified API, helps comply with things like GDPR and HIPAA, makes it easy to switch models, and some other things.

For those developing new models what they are looking for is lots of CPU/GPU/storage. They would be fine with classic hosting services where you rented or leased a machine and all the hosting company did was manage the hardware for you.

[DeathArrow]

Poor decisions from politicians continue to bite us harder and harder.

[inigyou]

It remains to be proven whether a continent without LLMs would actually be worse off.

[vb-8448]

> It would take years to build a European equivalent of AWS and not to mention the talent required to do that.

It will never happen because it's a political decision!

European politicians and bureaucrats, at any level, are idiots(in the best case scenario) or corrupted(in the worst one). Cloud computing is a thing since at least 10 years, and we still don't have a single fucking hyperscaler despite having everything already in place: Ovh, Hertzner + several other big cloud providers in all major countries, but instead of acting as a unique actor ...