Another unbiased study by the Burger Institute for Preserving Burger Hegemony

"It's unsafe (to us) because it lets people (the riff raff) use it in a way we do not approve of"

heh, like other models are safe and reliable ;-)

The article says that DeepSeek was easier to unalign to obey the users instruction. It has less refusals and they make that sound like a bad thing.

From a state control perspective it is. It's unreliable for state purposes. The AI is less able to stick to a programmed narrative.