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

Notes from the Mistral AI Now Summit

koenvangilst.nl

Post reply on HN
Print view View original

101–110 of 230 posts

Re: Notes from the Mistral AI Now Summit

#101

Post by pembrook »

ar0 wrote: post #40
trouve_search wrote: post #9

OK, I'm 100% rooting for both Mistral and task focused small models. But Mistral has fall really far behind since 2025Q3. It seems they can't get good reasoning models working at even medium context sizes, which is necessary to be at the table right now. Gemma4 and Qwen3.6 are currently best in the small size; Mistral's "small" model has ~4x the parameter count at 120B and isn't even competing with models a quarter i…

I agree. I am a paying Le Chat Pro user, really rooting for a European alternative. But the quality difference between Mistral and the frontier labs is growing too big to ignore. It’s worrying to me that they didn’t talk much about new models at the conference, because that is really where their focus should be IMHO. I am wondering what is keeping them back, though: Money? Compute? Skills? Training data? My fear is t…

> what is keeping them back, though: Money? Compute? Skills? Training data?

All of the above and more. Everything holding Mistral back is the same thing that has held Europe back from competing in the entire digital revolution. See this 1991 article lamenting the loss of any viable European PC manufacturer: https://www.nytimes.com/1991/04/22/business/europe-stumbles-...

Mistral being in Europe is disadvantaged with:

1. Money: less diverse private pension fund environment = less LPs to invest in VC funds = less VC dollars to invest in new ventures. European money is vacuumed out of the private sector into state pension funds and dumped into low yielding government bonds. This starves the private sector of capital while inflating the % of GDP driven by government spending every year (government pension funds buying government bonds in circular fashion enable runaway deficit spending...just like circular AI infrastructure spending).

2. Talent & compute: due to #1, Silicon Valley can outbid Europe for the best talent and hardware. Watch an OpenAI launch video and listen to all the European accents.

3. Local market fragmentation: Europe is a collection of countries that pretend to work together while not even having a unified capital market. The average EU citizen can barely communicate with their neighbor in a common language beyond the level of a toddler (english fluency is massively overstated by Americans who only experience tourist capitals).

4. Regulatory disadvantages: In everything from company regs, employee regs, unions, privacy regs, data portability regs, etc.

It's not "culture" or Europeans being "lazy" as most people would claim. There's currently thousands of young french people working 80 hour weeks creating dumb consulting powerpoints or legacy investment banking deal memos as we speak. Ambitious people exist everywhere in equal proportion, they're just working on the wrong things.

Europe can't compete in the digital revolution the same way they could compete in the industrial revolution due to various system design choices. Culture is simply the aesthetically observed byproducts of system design.

Re: Notes from the Mistral AI Now Summit

#104

Post by mediaman »

greyskull wrote: post #50
pavpanchekha wrote: post #42

Earlier quoted context omitted.

OpenAI used to make Codex-specific models, but they stopped. What I've gathered from interviews and similar is that training two models isn't worth the (small) lift from having a coding-specific model. You're pre-training on everything anyway, and coding RL is reasonably useful for general-purpose models too.

Interesting. I'd have guessed there would be meaningful opex benefits to serving smaller models.

What I've heard is that much of the model "intelligence" is a commingled bucket: although you can specialize specific knowledge somewhat, it's hard to specialize advanced reasoning to specific domains because so much of reasoning is a generalized capability that is not unique to, say, coding.

It turns out coding has to do with a lot of the same reasoning needed in math or in legal analysis, even if the grammatical expression is different.

This is less true of lower intelligence tasks. Classification requires a lot less reasoning capacity and so can be much smaller and more specialized.

Re: Notes from the Mistral AI Now Summit

#105

Post by xienze »

gregorygoc wrote: post #93
k__ wrote: post #88

Earlier quoted context omitted.

Why doesn't Mistral distill?

Good question, given that American companies basically threw copyright law into the trash, I think they should.

Well if they did wouldn't be able to feel superior to Americans about that particular thing. Perish the thought!

Re: Notes from the Mistral AI Now Summit

#106

Post by flawn »

crimsoneer wrote: post #54
irusensei wrote: post #46

Earlier quoted context omitted.

Please don't run Chinese models for KYC operations.

Based on what? Is there any evidence of risk at all?

The issue is that you wouldn't be able to even transparently get to any evidence, as these models are blackboxes.

They might start scheming behind employees backs as soon as they realize they are being used in critical infrastructure of adversaries. And nobody would know until it's too late.

Re: Notes from the Mistral AI Now Summit

#107

Post by neonstatic »

antirez wrote: post #83
neonstatic wrote: post #76

Earlier quoted context omitted.

> the EU funding nations are getting marginalized by the countries they welcomed Thank you for reminding us that all animals are equal, but some are more equal

It's not a matter of importance, but that bad actors as they tried to do in Italy, corrupting EU parliament, may be doing the same with counties are have less visibility. A weak EU is not also in the best interest of countries that wanted the AI act, and surely not in the interest of their citizens, but there could be pressures.

I understand your point, I just object to the language and dividing the EU into more and less important blocks. If a voting mechanism is broken, that's where the issue is.

Re: Notes from the Mistral AI Now Summit

#108

Post by sbinnee »

I believe that Mistral team is doing the best they can do. I like the directions they push; open models for various tasks, on-prem has a lot of potential. Sure, I use Claude code mostly for coding. But there are so many tasks other than just coding. Even for coding, eventually, I am certain they will catch up and Vibe becomes tolerable soon.

Re: Notes from the Mistral AI Now Summit

#109

Post by sbinnee »

antirez wrote: post #27

I really want Europe to be part of the AI development and research. And I strongly cheered for Mistral. But they are accumulating too much technological delay. This needs to be fixed, otherwise it will turn into yet another proof we are not able to run large tech with good results. Basically any Chinese lab is doing much better. It's not Mistral that created I don't want to say DeepSeek, but MiMo 2.5, Minimax 2.7, an…

When it comes to MoE, to me, I remember Mixtral model that showed the viability of MoE for the first time. I was impressed by their technical report. To be clear, MoE idea was already out there, if I am not mistaken. If they have pushed Mixtral model family further, who knows they might have achieved the reputation of what the current Qwen family has. A missed opportunity.
Post reply on HN