[ants_everywhere]

It's the one LLM we know of that is specifically trained to be opposed to democracy. Can you say more about why that doesn't concern you?

[djeastm]

I would throw Grok in there, too, but I suppose that might get us into a political discussion...

[ants_everywhere]

I don't know anything about Grok, but AFAIK they have different problems. Grok lacks alignment to push back against users asking for bad things, like advocating for hurting people.

But Deepseek is trained to manipulate the user into wanting bad things like hurting people.

[spencerflem]

I disagree that Grok, manipulated by Elon Musk, does not push ideas that hurt people