It would've really made this generation of consoles make a big leap in graphics (by freeing up space for other stuff, instead of trying to push native 4K).
Hopefully something will come up anyway, and we won't have to wait for next next gen.
At least the idea of reconstruction is already a standard with checkerboarding and such, Nvidia seems to be neck deep into machine learning though, i'm not sure how other companies are doing, and how easy it is for them to catch up on that front.
MS should be also into that stuff pretty deeply as well.
Nvidia spend a decade feeding their supercomputer images so we can play CONTROL at 60fps
Even if you catch up on creating the AI you still have to go through the process of feeding it enough content.
Those Tensor cores basically run algorithms their supercomputer crunched on 24/7.
If AMD started this year I would say it would take 3 - 5 years to come up with something that could match DLSS in a similar fashion.
Unless you find a way to achieve the same result with a different technique or find another way to make significant performance gains that rival DLSS.
I can also imagine that if it really takes that long Nvidia will license out DLSS tech or release a stand-alone Tensor Core Unit (TCU?) that can interface with another GPU.
Essentially adding DLSS with an embedded chip to motherboards, smart TV's, consoles etc. . They are currently working on such a solution for the new Nintendo Switch.
This is the closest thing we've ever gotten to real 'Secret Sauce' in computer hardware.
