128
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 29 Aug 2024
128 points (99.2% liked)
chapotraphouse
13511 readers
1166 users here now
Banned? DM Wmill to appeal.
No anti-nautilism posts. See: Eco-fascism Primer
Vaush posts go in the_dunk_tank
Dunk posts in general go in the_dunk_tank, not here
Don't post low-hanging fruit here after it gets removed from the_dunk_tank
founded 3 years ago
MODERATORS
People are being super weird about this paper. People who's income depends on AI hype are saying that "in the future all games will run on neural networks", which is retarded. And people who hate AI hype are complaining about the energy use, and saying it's pointless and unimpressive. And like... I guess we should all stop doing anything with computers that doesn't directly contribute to curing cancer or something.
We should have budgeted and project planning for energy use, yes. And it shouldn’t include AI deepfakes of women, or play a game that are 30 years old.
Why?
Because they use inordinately more resources than traditional methods. Running the actual complete doom game uses about as much electricity as having an AI generate just a handful of frames.
Yeah, thats fair, as a means of playing doom this is maybe as bad as it gets. But it was a research paper. It's not really being proposed as a game engine, it's just a dumb demo of what you can do by prompting a finetuned diffusion model with user inputs. As it stands the paper is reasonably interesting. Especially the need for noise injections. I feel like people are giving the researchers a lot of hate just because Elon retweeted it, and he's lame.
I think the issue is that really, this is research for research's sake, and has no truly useful destination. In of itself it's incredibly wasteful, able to only predict 60 frames of content from of 40 year old game, but the potential use of predicting frames for a game without using an engine is.... What, exactly? Making games that are far more process intensive for the same results? Simulating real environments based off already existing images, which we can already do with those same images?
We have limited resources and manpower to pour into research, but we're using it to have computers do what people already do for fun, instead of concentrating it on applications that could free up humans to do those same things with much less effort. Each of those frames used more electricity than most people use in their entire lives, but there's no potential in it to actually improve those lives. In a world with limitless clean energy and no major geopolitical issues it could be nice to have these kinds of treats, but we don't live in that world, so it's just a massive waste of resources that could be used to meaningfully improve the lives of billions of people around the world.
Counterpoint: You can run doom on your fridge or your mixer and it eats next to no energy. That's fun and impressive as far as uses of technology go.
There are many uses of technology and computing that are good. There are also many uses of computing that would be super nice to have, that we are not doing because no one can easily profit off of them. Heck, if you ask me, computer games are great, they are a new medium of entertainment that we haven't had through most of our history and that has unlimited potential for expression and fun! That's so very worthwile to pursue!
This though, this is just pointless. This is people desperately trying to keep investors hype about a technology that fundamentally won't do what it was advertised as, that won't ever reach the heights that people associate with the buzzwords used to sell it and that, while people keep pushing it, is actively burning untold resources in a big, pointless fire that produces nothing but climate change and samey, boring titty pictures.
It also involves theft from working people in nearly all cases, with this specific one being one of the very few where that's not the case. So good on them for that I guess.
but that still doesn't mean this is good. Not just because of the energy use, but because the 'game' they are simulating has a 3 second memory. That's not a trivial problem to solve, as each frame of memory they add also adds tons of computing complexity to the problem that needs to be solved to generate the next frame. The levels also make no sense and would still make no sense if you add this long-term memory. The AI doesn't know what a blue key is, that it should place it, that it should place a blue door with it and that you should not be able to reach the part of the level past the blue door without going through that door. There is more but I think the point is made. The only way to solve these issues would be to basically program the AI to run doom, which defeats the whole purpose of this experiment.
And then even if they do that, there is no AI. At no point did the program doing this understand what it is doing. Heck, you could have a better result then this by having the AI generate doom map maker files. At least then it would not need all the graphical input and output in real time! But they didn't do that because this so-called AI has one very specific thing it can do and that is making pictures and so every problem has to be solved with that one and only tool.
Basically, as soon as generative AI enters the picture, it's bad. There are some few applications that it has uses for, but right now we are in a giant stupid bubble that is killing our planet while it gets more and more inflated and that....is bad.
it's just easier to say it uses a lot of energy then to say all this and engage with this crap. People are tired of this AI shit. Saying it kills the planet really should be enough of a reason to stop doing it, but we live in a society where people with a big digital number on their bank account get to do whatever the fuck they want and we can't stop them. Yet.
Yeah, we should. In this dog shit world it will inevitably be a grift anyway, but at the very least it’ll attract people who actually care about doing good things to maybe develop something good for the world instead of new investment vehicles.