All the early opinions on fsr 3 are to be quite frank, pointless. They were based of what amd "showed" (which was very limited/short footage), no one has had a proper hands on yet (and latency will be key here given nvidia are relying on hardware to overcome this and amd aren't from my understanding). The initial PR titles for FSR 3 should also be taken with a huge pinch of salt too as look at what happened with fsr 2, it was pretty good in deathloop then it got out in the wild outside of PR sponsorship deals and ended up being pretty awful 90% of the time. I am fully expecting the "dlss 3 killer!!!!" videos etc. to start popping up after the first game with fsr 3 is reviewed
The thing I find quite hilarious with fsr is that the game with the best fsr implementation is cp 2077......
DLSS 1 was indeed dog ****, the worst upscaling to date but that doesn't excuse amd from being basically 3 years behind on this area now, difference is dlss is continually improving where as fsr doesn't seem to be much, if at all and amd don't seem to be doing anything to address the constant points that are highlighted in every comparison but again, this all comes down to how amd have picked this path i.e. hands of approach and let the developers fine tune to get the best from it, which as we all know, ain't going to happen, they need to adopt a similar approach to dlss with t-shirt size fit so at least devs don't have to spend too much time to get good results.