It's still in its relative infancy, but AI integrated upscaling and frame generation.
The fact is that hardware improvements aren't having nearly as noticeable of an effect on the image as we've seen in the past. A lot of people assume this is because the performance improvements aren't there anymore, but that is not quite accurate. A large part of what we are seeing is a function of our perception. Take a look at a square and then an octagon. Its simple to see which is more circle like with only four additional sides. Now look up a hectogon (100 sides) and a chiliagon (1000 sides). The actual perceived difference, ESPECIALLY considering the object we would look at would be in motion, is negligible. And that's with 10 times the sides.
For us to see the leaps that we saw in the past we need larger hardware upgrades than ever before. All the while people are asking for increasingly higher resolutions and framerates. But with techniques like DLSS and frame generation developers can satiate the consumer when it comes to resolutions and framerate while still having a large portion of hardware resources focusing on improvements to textures, character models, lighting, and anything else.