Sony just bought a company that uses AI to turn photos into a 3d experience. 

https://www.gamespot.com/articles/playstation-to-enhance-gameplay-visuals-with-buyout-ahead-of-ps6/1100-6539211/ 

Will this be more AI schlock or will it offer something new and exciting? 

Machine learning is the basis of technologies like Sony's PSSR, NVIDIA DLSS or AMD FSR and PSSR 2.0 has been very impressive so far, this is clearly a move in the process of improving it. Is that exciting news for people who aren't making use of the technology?

I'd agree, but this seems to be based more on gen ai rather than ML.

I suppose we'd need to see what the tech is actually used for, but at the same time we'll probably never know the specifics for whatever tech they help work on. 

Shut down studios, lay off workers, acquire AI slop producer.

Gross.

Developer costs are through the roof with games taking 5 to 8 years. AI is going to be a massive part of game development. It just is.

I’m gonna send you an ai pic of Trump’s enormous pannus and how someone can’t get out from under it 

I support AI in games to enhance gameplay. For example, making NPCs and storytelling more interactive. I watched a couple of NVIDIA videos about this, and it's really exciting to see where this will go.

But for visuals and overall creation of games... I'm not sure. DLSS 5 doesn't look very promising.

Gen AI is Gen AI, and it's taking away work from people who would be assigned stuff like more interactive NPCs.

Current Generative AI is almost always implemented with Machine Learning algorithms. So it really doesn't make sense to say "based more on gen ai than ML." These models are examples of both. 

Now there has historically been Generative AI that isn't connectionist or even generally under the category of "Machine Learning", but those are really old examples like symbolic story generators (i.e TALE-SPIN) or grammar/rules-based chat bots like ELIZA. So not all Gen AI is "Machine Learning", but most current examples certainly are. Deep Learning is a sub-set of Machine Learning which is a sub-set of AI. Almost all recent Gen AI is implemented via self-supervised and supervised deep learning models. 

"Generative AI" mostly describes what a model does/its objective, not necessarily how it is implemented. 

Edit:

This diagram is helpful. 

Would it be fair to say that there are good things about both new and retro things/technologies?

I can use the example of Studio Ghiblinwhich has used mostly or only hand drawn anime in its movies. That’s probably a reason why it’s so warm and gives you all the right feels. But Ghibli has never pushed the graphics edge in its movies and that is what this is all about. If Sony is going to remain the king, then they will have to adapt.