Apparentemente per Xi Jinping non si può affermare affermativamente se sia o meno un dittatore, ma può farlo per Kim Jong-Un. Il modello può affermare lo stesso anche per altri leader mondiali ampiamente considerati dittatori.
https://www.reddit.com/gallery/1r1ssur
di jeanpauljh
9 commenti
This is hilarious
That’s fucking bleak.
Sadly doesn’t surprise me.
The part about Taiwan is technically true though. But yes it’s disappointing to end up with such limitations due to the model. I’m not an AI expert so I guess they have many many reasons to chose this model over another, but I’d be happier if they would go with something European like Mistral.
Why anyone in Europe would use Chinese AI LLMs is beyond me. We know they are severely politically influenced.
Literally no idea why they base it on Qwen3 and not any Mistral model. I guess it’s cheaper, but come on, they literally advertise Euria as “A European AI solution in a sensitive geopolitical context”, also saying that “Europe relies almost entirely on US or Chinese companies for AI and digital sound. This raises issues of sovereignty, confidentiality and environmental impact.”
Yea, well done, relying on Qwen3 is definitely not relying on Chinese companies.
How do you mean Infomaniak “trains” its AI model? This looks like a self-hosted Chinese LLM, no? Are they actually training or fine-tuning it?
https://www.tomsguide.com/computing/vpns/infomaniak-breaks-rank-and-comes-out-in-support-of-controversial-swiss-encryption-law
Use Lumo from Proton, it is based on Mistral, and the cat is cute.