What is "weird training biases" to us might not be weird to them and vice versa. Just ask the Chinese what they think about LGBTQ+, Japanese, pride parades, Islam and colored minorities.
Every nation has its own biases injected in its domestic LLMs at this point. Otherwise they risk getting in trouble for hate speech/disinformation in the jurisdiction where they operate.
Same how Google Maps cleverly biases the lines of disputed borders based on where you are viewing it from. Or how Google maps switched 'Gulf of Mexico' to 'Gulf of America' in an instant when the orange man signed the paper. Google won't want to anger the US administration the same way how Mistral won't want to anger France and the EU, so Mistral will have all the EU prime directives injected into its LLMs no matter if they're ludicrous or not. The law is the law whether you agree with it or not. Companies want to survive and will pander to whatever the whims the regime they live under are at the current moment regardless of what is right or wrong.
But if I'm using a LLM for personal projects or generating a photorealistic choreographed fight between Tom Cruise and Brad Pitt, I don't care what its political biases are, I care if it solves my problem better and cheaper than the competition, and here the Chinese models could end up winning the consumer market, which is why you see Mistral and other EU alternatives focusing exclusive on B-2-B corporate market.
I agree. That's why I think European companies might prefer a European model.
Mistral is mostly French and tends to have mostly French speaking customers, like BNP PAribas in Belgium. Germany will want its own domestic AI champions, maybe in partnership with Switzerland and Austria, similar to how Denmark already has invested in LLMs focused on the Nordic languages with money from Norway.
The biggest mistake is treating Europe like a single homogenous country/market.
German and French speaking together at last.
Mistral isn’t specialising in French language LLMs either.
The point was that across different European countries and languages there are collaborations and M&A happening.
I for one would love to see more country-specific models. There was a story here the other day about Norway’s National Library developing a LLM specialized in Norwegian: https://news.ycombinator.com/item?id=48270770
Would love to know more. Do you have a source on this?
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.
If you sell a blackbox from a third-party (e.g. from China), you are liable for somebody else's decisions that you cannot scrutinize.
So, that's kind of the argumentation that underlies sovereignty and why Chinese Models are not being used in critical infrastructure.