'We can't do computer graphics anymore without artificial intelligence. We compute one pixel, we infer the other 32': Jensen thinks AI is integral to next-gen graphics tech
At this point I just want to have quality games. The graphics are beyond good enough. I don't need to custom render a photo realistic action movie from my PC.
As an old fart, I actively dislike photorealistic graphics in most cases. I'm playing a game, and I kind of want it to look like a game, which generally means more surrealistic - exaggerated contrast, high saturation, low texture - than realistic. I'd rather play where the characters look like caricatures than my next door neighbor. And that doesn't even go into great games with sprite-like graphics.
Enough is enough. You've saturated the art budget, it's time to pay writers more.
You’ve saturated the art budget, it’s time to pay writers more.
I wish writing got more focus in general. There is a lot of theory to good writing that is often just completely ignored while the latest theoretical papers are taken into account for photorealistic rendering and such things that are much less important.
Yup, I honestly avoid the hyper-realistic games anyway. The closest I have gotten recently is the Yakuza series, and even that is very clearly a game, even in their high-quality renders. Gameplay is far more important than graphics quality. I don't even care at all about RTX, just give me a fun game with an interesting story, and give the art team a lot of leeway on how to represent that.
I almost never buy games day 1 because they're full of bugs (though they do look pretty), but do you know what game I'm excited to buy day 1? Zelda: Echoes of Wisdom. It's basically the opposite of the big-budget, hyper-realistic games, and I'm all for it. I expect great gameplay and minimal bugs, and I'm willing to pay a premium for that.
I wish the big studios went back to putting fun first, instead of trying to compete on who can run my PC temps the highest.
I'll be honest, I have trouble with low graphic games. But mediumish is fine. I still enjoy plonking monsters on Diablo 2. But I also enjoy Satisfactory and Snow Runner. I can't conceive of a good reason graphics need to go further. Hell, Balder's Gate 3 was beautiful and these guys want us to buy more graphics capacity? Why? It's ridiculous.
It's an underappreciated fact that art direction trumps graphical fidelity, and it's not close. Grim Fandango for example is old as dirt but still holds up remarkably well thanks to its unique look and strong art direction, despite being challenged polygonally and resolution-wise etc. There are many other examples.
I've been repeatedly disappointed with most modern games, so I've taken to emulating my old games and playing them on my laptop. It's honestly a pretty good time, would recommend.
I'm getting to play games I loved that I haven't touched in over a decade (MotorStorm, all the Ratchet and Clank games, the good Need for Speeds, etc). Plus if there were any games I wanted as a kid but didn't have the money, I can buy most of them off eBay for cheap.
It entirely depends on the genre. I probably would want a bit more detail on the faces than that if it was an emotional story line, say the kind of quality that To The Moon had.
We certainly can. NVIDIA’s CEO realizes that the next buzzword that sells their cards (8K, 240hz, RTX++) isn’t going to run at good framerates without it.
That’s not to say AI doesn’t have its place in graphics, but it’s definitely a crutch for extremely high-end rendering performance (see RT) and a nice performance and quality gain for weaker (hopefully cheaper) graphics cards which support it.
As a gamer and developer I sort of fear AI taking the charm away from rendered games as DLSS/FSR embeds itself in games. I don’t want to see a race to the bottom in terms of internal, pre-DLSS resolution.
With you there. The workload on developers is reduced with these features, to a degree. But, instead of saved effort then getting directed to working on gameplay mechanics and such, to me it feels like many devs just see it as time/money saved, producing a game that looks and plays like one from 10 years ago, but runs like it's cutting edge.
For instance, Abiotic Factor. That game on my RX 6800 XT runs at 40-50fps when at 100% resolution scaling at 1440p. Why? It's got the fidelity of Half Life 1, why does it need temporal upscaling to run better? (I adore that game btw, Abiotic Factor is so much fun and worth getting even if playing alone!)
Not saying that's how every dev is, I know there are plenty of games coming out nowadays that look and run great with creators that care. Just feels like there are too many games that rely on these machine learning based features too heavily, resulting in blurriness, smearing, shimmering, on top of poorer performance.
Just hoping the expectation that something like an RTX 4090 does not become the default cost-of-entry in order to play PC games because of this. It would be unfortunate for the ability of game developers to create and tune by-hand to become a lost art.
As a (non-game) developer, AI isn't even that great at reducing my burden.
The organization is enthusiastic about AI, so we set up the Gitlab Copilot plugin for our development tools.
Even as "spicy autocomplete" only about one time in 4 or so it makes a useful suggestion.
There's so much hallucination, trying to guess the next thing I want and usually deciding on something that came out of its shiny metal ass. It actually undermines the tool's non-AI features, which pre-index the code to reliably complete fields and function names that actually exist.
This was my fear when they announced they got an AI to generate doom in real time with no code (well code for the AI but they got the game via a prompt) yeh its amazing but they were running oldschool doom at 20fps on top end hardware with additional AI hallucinations.
Next thing you know we'll be simulating an entire universe just to play bee simulator.
If you are inferring 32 pixels from 1 pixel that is because the model has been trained on billions of computed pixels. You cannot infer data in vacuum. The statement is bullshit.
I would go so far as to say if you get rid of the graphics completely and have text descriptions (think Dwarf Fortress which has many things that are not represented in its graphics at all, just in the textual descriptions) you fully free the imagination of the player.
Some things are just not representable graphically at all, my go to example is "the most beautiful woman he had ever seen", easy to write in text, impossible to portray on screen in a way that every viewer will feel the same.
Idk, I really suck at imagining things like in DF, but I really enjoy the gameplay. I don't need a good representation of what I'm interacting with, but I do need something.
Perhaps, we should be more concerned with maintaining and keeping relevant current hardware, over constant production of more powerful hardware just for the sake of doing it. We've hit a point of diminishing returns in terms of the value we're actually being given by each new generation. Even the PS4 and Xbox One were able to produce gorgeous graphics.
Even the PS4 and Xbox One were able to produce gorgeous graphics.
But not always (or even usually IMO) at decent frame rates. I get major motion sickness and nauseous if a game doesn't run near 60fps constantly, it's really unfortunate.
I mean we could do things like Arkham Knight, Flight Simulator, The Last of Us 2 and so on. Do we really need to do everything realtime or could we continue baking GI?
I would assume that they are saying in a bigger scope and just happen to divide down to a ratio of 1 to 32.
Like rendering in 480p (307k pixels) and then generating 4k (8.3M pixels). Which results in like 1:27, sorta close enough to what he's saying. The AI upscale like dlss and fsr are doing just that at less extreme upscale.