Oh man, AI generated content all over the internet wasn’t enough, now they’re going to also generate our video games? It really feels like all our entertainment will be soon randomly generated.
Well shit, 20 years and it looks like we may have replaced our entire workforce with robots that dont need to sleep, and can produce everything we need using low environmental impact devices. You know. Maybe the Jetsons weren’t wrong, just early.
How do we not become WALL-E though.
Obviously 20 years is a vast overestimate when humans are involved
You think the world we live in supports this future? That’s cute.
So… unlike Stable Diffusion or LLMs, the point of this research isn’t actually to generate a direct analog to the input, in this case video games. It’s testing to see if a generative model can encode the concepts of an interactive environment.
Games in general have long been used in AI research because they are models of some aspect of reality. In this case, the researchers want to see if a generative AI can learn to predict the environment just by watching things happen. You know, like real brains do.
E.g. can we train something that learns the rules of reality just by watching video combined with “input signals”. If so, it opens up whole new methods for training robots to interact with the real world.
That’s why this is newsworthy beyond just “AI Buzz” cycle.
I have the same feeling about this as I have about text written by A.I.: why would I want to read something no one could be bothered to write? Art isn’t interesting because it’s technically adequate. It’s impressive because a human (or humans) made it and it meant something to them.
I’m all for A.I. as a tool for creative people and doing rote tasks but I don’t get the point of A.I. generated “art.”
People are different. I just like to look at cool pictures. I couldn’t care less about the person that made it. I expect anyone to do the same with anything I create.
People go to casinos and play slots for a hit of brain chemicals. They’re not there for art. Nor are whales in most mobile games.
I’ll bet some exec thinks an AI-enhanced Skinner box would interesting to see.
Fuck this hits hard.
I never understood slots, nor gambling in general.
So I never saw how fucking HUGE it is. Oh my god people sink THOUSANDS of dollars of their hard earned money every month on the stupidest slots imaginable. The more stupid the slot is, the more popular it gets.
Online, offline, doesn’t matter. People are at it EVERY SECOND of the time. It does not stop. Ever.
I worked in a bar and I can’t remember exactly but the money taken on a busy Saturday the pokies were 5x the money taken on the tills for drinks. Obviously drinks cost money also so that’s not profit. But I was amazed the country I come from people don’t really gamble like that.
It’s a game not art. What’s most important import is player agency and mechanics.
Shit people have been clamouring for interactive stories where you can explore things on the fly for ages. This is a step in that direction, one day we could have game/movies where you get to explore a world without following someone else’s pre-scripted reality.
I’ll gladly take that over some bougie notion of what art is.
I agree. And AI is a tool, like a brush or a modeling software. It’ll accelerate artistic processes, but won’t completely replace people.
Innovation happens at the intersection of people and technology. Each is more capable with the other than on its own.
I don’t think you could say this is a video game. It’s ai image generation based on the previous image and a user command like “go left”. A video game would also need a plot of some sort.
This tech is going to get pretty wild.
Years ago Nvidia were playing around with things like this as the far future of DLSS.
Even imagine something like a remake - you could literally just pump the gameplay from GoldenEye 64 into a model that redoes the graphics with CGI levels of detail when generative AI like this can consistently pump out frames in realtime.
Particularly when the models can also predict inputs based on input so far, there wouldn’t even be perceptible lag (GeForce Now does something like this actually).
What’s the plot of tetris
You’re failures pile up while your accomplishments disappear
Break blocks, get laid.
You’re not wrong, but you’re wrong to be dismissive.
What you’re seeing is about how good text-to-image was in early 2022. Comparatively for complexity.
Hold onto your butts.
