AlOwain on Nostr: Deepseek R1 is genuinely terrible. If you think it isn’t you should try it out ...

Deepseek R1 is genuinely terrible. If you think it isn’t you should try it out before ceding to the hype.

Llama is better.

It could be that the way I use those models biases me, as I primarily want a model that gets me what I want fast, rather than accurately.