Raw LLM Responses

It's funny, because everyone that have tested and used the local models for a while mostly agrees that they are really far behind for example ChatGPT. [A bit of discussion around it](https://www.reddit.com/r/LocalLLaMA/comments/13gpnhq/addressing_the_elephant_in_the_room/) Part of this misunderstanding comes from various factors, from tests being very one dimensional to using very un-scientific means to compare *(like letting chatgpt rate each model)*, and some trying to find the comparison that shows their pet model in the best light. So that google paper and the resulting discussions and panics are largely built on a false premise. Still, it's pretty interesting to see how the industry is reacting to this perceived threat. That said, I really hope open models keep on evolving and becoming better and better, and some day surpass OpenAI's models while still being able to run on normal hardware.