What even makes it worse is that they always make the models worse a few days after release, this is a wide problem across all the companies
do you know of anyone that's put out quantified studies on this? I'd be super curious to see. Like benchmarks at release vs 1wk vs 1mo?
I would 100% be down to research it, I just know it but I cant prove it because I can feel I have agents running 24/7 almost and almost every time at the beginning of the launch they are great and they gradually decrease everyday...