Hacker News Reader

Comment by v3ss0n | Hacker News Reader

No amount of $ would improve mistral if they cant fix fundamental flaws. They aren't even on par with chinese models a year ago.

View on HN

I actually believe the amount of $ is what was missing for them to improve their fundamental flaws. They're a very competent team, but were working with a tiny fraction of the budget of US/China teams.

You sure? Kimi K2 Thinking was apparently build by just 80 people and not more funding.

> Kimi K2 Thinking was apparently build by just 80 people and not more funding.

But they're not competing (offically at least) for GPUs and they have much, much cheaper electricity (which makes pre-training more affordable).

If Mistral aren't going to distill other people's large models they obviously need to train their own large models. This obviously requires money for optimization, tuning and training hardware.

They've started hosting GLM-5.2, that should be very telling of their capabilities at the moment. Hopefully the investment will allow them to hire the right people to become competitive.

can you elaborate what are their fundamental flaws that are not fixable by funds?

> can you elaborate what are their fundamental flaws that are not fixable by funds?

Culture.

In terms of output, EU working culture is inferior to American/Chinese working culture.

Work life balance is irrelevant for frontier/sovereign AI.

Holy oversimplification. Do you get all your information from the internet?

There is logic to it. The ones who work 12h/day for 6d/week does more than the ones who work 8h/day for 5d/week does more than... It's simple scaling that pays off over time even if the general quality of work by the former isn't as high as that of the latter.

>inferior

Way to reduce a multi dimensional concept into a single scalar value. Care to explain what exactly make it inferior? Try using more than one word if you can.

In reality US and China are just not efficient enough.

Brand new account, only 1 comment.

Checks out

Thats one of the valid point.

They've never made a near-frontier model. Which is a valid criticism for a model training company

"Frontier" is a relative, ever-changing target.

It's not really: it means those out in front.

OpenAI, Anthropic, China

Those in front can and will change over time, and regions determine what is measured in the first place. So if they want, the EU can for example define "frontier" as "models made by a European nation that score highest on Math benchmarks".

They have done that in the past. I clearly remember the times when they dropped releases as a simple tweet and it made the front page of HN - just like the recent Chinese models releases.

Chinese models are trained on dubiously collected model traces from Claude/OpenAI that are purchased from model routers. All the major Chinese models use this data. That's one reason they've been able to catch up with Anthropic/OpenAI so quickly, they have so much data.

Mistral can't train on that data, because this data would be illegal to purchase & train on in the EU.

It's dubiously collected data the whole way down. This is the AI industry.

Honestly being only 1 year behind makes me an optimist. You’re telling me Europe can be slightly behind with 1000x less capex and way more sustainable economics? Awesome. The world moves slower than AI progresses, I can see a scenario where 1 year isn’t a problem.

Chinese models are open, available to distill, and they also publish papers about their research. Being one year behind is a skill issue.

I think the most of the money would go to purchase hardware, But I hope they can start making adequate compensation for AI engineers and researchers to move forward.

My feeling is that its a difference in how funding works in different places. The USA will go all in with the populations pensions on a gamble, the Chinese subsidize. This way of operating is typical for the EU.

> Being one year behind is a skill issue

E.g. if you are an AI researcher in Europe you can just go to USA and make generational wealth. This is not a criticism of the EU model, but rather insane American capex effectively monopolizing.

> I think the most of the money would go to purchase hardware,

So it's not a skill issue?

> > I think the most of the money would go to purchase hardware,

> So it's not a skill issue?

I meant by offering sovereign cloud/inference, not for training. But even if it was for training, Chinese labs have limited supply of GPUs, look what they've done. So it is a skill issue.

Also to clarify, I didn't mean European engineers' skills, I meant "you get what you pay for" as a company, that's why I hope they start offering better compensation to retain talent.

Understood! Thats reasonable

> Chinese models are open, available to distill

Handing in a photo-copy of a photo-copy of a photo-copy of a photo-copy of a photo-copy of a photo-copy of someone else's homework might work once or twice, but it is no way to run a (non grift) business.

Its not a grift if it works. If a Chinese corp copies at 90% quality at 50% price within a year it means your American business model was never a good one for competing in international markets. The real grift arguably is pretending that it was.

Where is that skill coming from in the first place? What are they doing to attract, nurture and keep it?

There is nothing exciting about mistral but they're the only european ai lab i've even heard off (jeppa doesn't count)

there was "H" at paris at the time, they raised 200M or so, but never got so much visibility. don't know at which stage they are now or if they accomplished something

All original scientist co-founders of H have left.

From what I've seen in the past few months, Mistral is the unique European lab still trying to compete (I wouldn't count Poolside as EU).

There is Apertus from Switzerland, a real "lab" project, not a company one.

Reply on news.ycombinator.com