Artificial Intelligence thread

caudaceus

Senior Member
Registered Member

Kimi Code scores better and uses fewer token. That's good stuff

from my own testing, it looks fine, but I was already happy with 2.6, so this doesn't do all that much for me. I did tell it to make some scripts better and it did that.
I'm just a bit disappointed with Kimi still nvidia-pilled. Having said that indeed Kimi as a model is excellent, I can confidently put it above Gemini pro
 

dripblackcoffee

New Member
Registered Member
View attachment 176573


View attachment 176574

Dario called for regulation of models, well he got what he wanted?

Please, Log in or Register to view URLs content!
I mean yes unironically this is what he wanted, they seem to think that RSI is soon and that their lead is soon to be permanent, that and mythos is the first step to this victory, that the IPO is simply a means to an end. Money doesnt really matter if you win, while I have alot of problems with this argument, unfortunately mythos is genuinely extremely good and an actual step change in ai akin to when ai models first actually got good at coding.
 
Top