I love that the model is called Le Chat
Mistral AI is pretty cool. French startup, they released quite a few open-weight models that are all available on HuggingFace (although I’ve heard that they don’t perform too well… but anyways)
I love that the model is called Le Chat
Mistral AI is pretty cool. French startup, they released quite a few open-weight models that are all available on HuggingFace (although I’ve heard that they don’t perform too well… but anyways)
The tech might be the same, but the models can certainly differ, and something that is trained on US-centric data gives US-centric results, which may not always be desirable.
I can’t imagine EU-centric results would any better though, likely much worse.
How so? For pretty much any application in the EU, EU-centric results are obviously much better?
Because they need to be honest and adhere to the EU regulations and not be asshole thieves like Muricans and Chinese?
Being “nice” severely limits your options.
I don’t think that’s what we’re talking about? As an example, if I ask an LLM for good spots to get Bratwurst and it recommends places in New York, it’s fairly useless to me as a European.