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 that you are really only getting really good models by training on very dubious data (outputs from the frontier models etc) and that Mistral is too European and too enterprisey to take those risks.
I think the big difference is that Google is free: everyone is using Google because it doesn’t cost anything and for a long time was the best search engine out there. I am sure that if Google would suddenly charge a few dollars per month for access, Bing market share would explode overnight, because it would become “good enough but cheaper”.
With the AI models, using a model that is “good enough but cheaper” is already an option.
Yes, this statement is so old and wrong that it gets boring real quick…
You can just look at the roads of any country that has the same tariffs for U.S. and European cars and you will still not see all that many American cars there.
This is a very shady website and thus not a good source for legal advice of any kind… they call themselves “Deutsche Gesellschaft für Datenschutz” (German society for data protection) but are actually located in Bulgaria. They are not any kind of “official” data protection organisation.
The German BSI (Federal Office for IT Security) quotes the advisory from CrowdStrike (which is behind their customer login portal) as saying you need to roll-back to snapshots prior to 04:09 UTC:
I don’t know… I think a very big reason why people don’t take taxis is because they are very expensive especially for longer rides. This seems like a thing robo taxis might change. If the driver goes away, they shouldn’t be much more expensive than e.g. car rentals.
I do not read this court decision like that at all: the point of contention there seems to be that the customer was just sent a link to a webpage (where the contractual terms can be changed from under him at will by the company, thus this not being durable). The court makes it pretty clear in my (non-lawyer) opinion that attaching a PDF to the email would have been fine.
I don’t know… the overall tone seems to be a bit too negative for me here.
I have used Nextcloud at home for years now without issues and we also used it at a large university where it worked just fine (from a user perspective; I don’t know if it gave the administrators nightmares). I do agree that they should invest more time in polish and stability and less in swanky new features that many won’t need, but that would not lead me to discourage anyone from using (or at least trying) Nextcloud.
I am wondering what is keeping them back, though: Money? Compute? Skills? Training data? My fear is that you are really only getting really good models by training on very dubious data (outputs from the frontier models etc) and that Mistral is too European and too enterprisey to take those risks.