Design will be set long before you can know if RDNA4 "Did it's job".
RDNA4 is in the lab, has been for some time considering expected Computex announcement and availability soon after.
Design was set before H2 2023. The timing of GFX12 commits lines up. No clue how it performs until OEM's slip up.
The job AMD needs to do is software more than hardware. DLSS was a major deciding factor for me. AMD needs competitive neural network scaling. It's embarrassing, that even Intel XeSS is better than FSR scaling.
There is more than one way to skin a cat, AMD will never win by simply copying NV verbatim.
FSR2 can work nearly as well, it just requires careful integration from the ground up, fully replacing the native AA.
DLSS is idiot proof, it just overwrites crappy TAA without a care in the world.
Right now AMD just needs to rework Anti-Lag+ and rework FSR2 in the overall FSR pipeline to work better with the other techs, becoming a full plug and play that just works.
They are getting closer to software parity, and mind you SW engineers are all in on DC software right now so the pace of gaming IHV software improvements will likely stagnate for a while.
At the end of the day, the ones who desperately need to catch up are Microsoft and Khronos, the best scenario for consumers will always be hardware agnostic software that lets the best hardware win, no gimmicks, no vendor locking to newer parts to entice FOMO for otherwise bad value purchases.
GPU is a virtual monopoly right now, consumer purchasing power sucks and yet, it is the most important hardware out there.
75% GM's means everyone is being taken on a ride, swallow your pride, ignore the shiny propaganda and buy an alternative if you can. Even if it makes things harder for you, it is better in the long run.
Seriously, if RDNA4 is good enough, NV might not bother making a price competitive part for the sub-$500 market and cede some share. Reward it and just buy the better value part for once. If you are a chronic high end buyer, you might stick with your 4090, NV will probably value target Ampere buyers with high end Blackwell, so wait and see what RDNA5 high end could offer, assuming it doesn't explode.
There is Intel too but that is more of an experiment than anything.
Just don't reward a company with obscene greed and dominance for doing the bare minimum when they don't care the slightest in the market right now.