I find it hilarious that the Youtube spokespersons go out of their way to clarify that this is "not the bad GenAI shit that we know everyone hates but the good kind, you know, machine learning and stuff, you know, trust us"

The Youtube liason statement is truly something to behold. It's not "upscaling" it's "unblurring". There's no genai, just "traditional machine learning". Oh, phew! I wouldn't want to have misidentified post-processing as unblurring when it was really upscaling.

I think we may see more of this which, to be blunt, I would say are stupid equivocations that are orthogonal to the actual concern, namely that it violates a fundamental trust that images represent something that really happened. We saw this already with Samsung's headspinning justification of their post-processing a fake moon.

Strap in for more JV debate team level "gosh what is reality anyway" equivocations, which I suspect will become increasingly prevalent.