I realize now that what Nvidia is doing with this generation is priming consumers to conflate hardware performance improvements with software performance improvement. What Nvidia is trying to do is to sell us 4050 silicon at 4060 prices and have the software magic and fake frames to convince the user that rhat there's enough of a generational improvement. I'm going to predict that it'll be even worse with the next generation. So what we'll see is that say a gen over gen performance increase used to 50%.....now it's 20% hardware and 30% software, but next gen will be 10% hardware and 40% software. I feel that this is why Jensen was so vocal about declaring "Moores law dead", to get us to believe that and to expect less out of the hardware....he's basically going to try and turn Nvidia into Apple in the respect that apples software and close integration with hardware is what makes them perform (i.e. how a Mac mini with 8GB of RAM is still very usable despite the same amount of memory on a PC being inadequate). And, just like apple, despite getting less on the hardware side (which would indicate a lower coat on the BoM), those saving will not be passed on to the consumer.
Some of you may say, "Hey. I don't care if the performance comes from the software or not", but you should, because if Nvidia keeps going down this path, it may necessitate further proprietary software and a walled off ecosystem which would only make it even more difficult for competition. Also, prior to DLSS and frame generation, a consumer could be confident that all frames are equal, whether across different reviews or across different hardware, a frame was a frame and it was a known quantity, but now that's not the case. We're entering into territory where all frames are NOT equal, and I think this will only make things more difficult for consumers.