Oh, especially in games i actually don’t expect much. Except another kind of “single-player”-games that totally need to be always online due to AI. As our machines can’t handle the AI on top of the game itself. Also it would be a real challenge to give it “ai” and also kinda restrict it to a given ruleset/theme/boundaries/etc.
Like an NPC using ai. What does he know? Also it would be pretty immersion-killing if you’d ask him something and then it would take 1-3s for an answer. Like those incredibly stupid alexas.
Guess we’re maybe just way too far away from an actual and usable integration
Plenty of games has that problem without IA. And i’m guessing we aren’t gonna restrict ourselves to hardware, the majority of games already doesn’t care to take a lot of space and graphic processing.
That said NPCs in a game doing various actions using the simplest language model has been done quite some time ago. But it wasn’t actually used more than as a novelty. Maybe later idk.
For dialogue i agree that it takes to much time to respond, at least for now. But i think it doesn’t really has to be done in real time, it can just be created beforehand.
And that’s true in general, we probably are restricted to use it in the creation process now.
What’s probably the most useful would be on asset generation, since that’s one of the most time consuming part in game dev.
I’d guess especially for all the random generated games that would be amazing too at some point.
Recently for example we’ve seen ai be able to create very good looking videos. And the technics they use for image consistency includes thinking about the object as 3d ones.
For now though it’s mostly used for interpolation, creating images between two existing ones to get more fps and fluid movement, pretty effective.
Another one i’ve also seen it used is for lighting, the process being pretty calculation heavy it can really help. That’s mostly useful for the prebaked light that’s created once and then stored. But i think it will be very useful for dynamic lighting one day, i feel it’s not that far either.
Oh right, i was totally fixated on the npc and dialogue-thing. For everything else in production(!) it’s surely helpful. Texture-Generation. Seamless. Audio. Music. Even level generation and whatnot, that’ll be helpful a lot.
I let ai design my watchface-background as i would’ve never found one i like and I’m creative as a doorknob.
But what it lacks is theme. You can only fill gaps, but you can’t let it generate everything. You still need an artist with a vision and a given theme.to have a consistent art-style.
Oh, especially in games i actually don’t expect much. Except another kind of “single-player”-games that totally need to be always online due to AI. As our machines can’t handle the AI on top of the game itself. Also it would be a real challenge to give it “ai” and also kinda restrict it to a given ruleset/theme/boundaries/etc. Like an NPC using ai. What does he know? Also it would be pretty immersion-killing if you’d ask him something and then it would take 1-3s for an answer. Like those incredibly stupid alexas. Guess we’re maybe just way too far away from an actual and usable integration
Plenty of games has that problem without IA. And i’m guessing we aren’t gonna restrict ourselves to hardware, the majority of games already doesn’t care to take a lot of space and graphic processing.
That said NPCs in a game doing various actions using the simplest language model has been done quite some time ago. But it wasn’t actually used more than as a novelty. Maybe later idk.
For dialogue i agree that it takes to much time to respond, at least for now. But i think it doesn’t really has to be done in real time, it can just be created beforehand.
And that’s true in general, we probably are restricted to use it in the creation process now.
What’s probably the most useful would be on asset generation, since that’s one of the most time consuming part in game dev.
I’d guess especially for all the random generated games that would be amazing too at some point.
Recently for example we’ve seen ai be able to create very good looking videos. And the technics they use for image consistency includes thinking about the object as 3d ones.
For now though it’s mostly used for interpolation, creating images between two existing ones to get more fps and fluid movement, pretty effective.
Another one i’ve also seen it used is for lighting, the process being pretty calculation heavy it can really help. That’s mostly useful for the prebaked light that’s created once and then stored. But i think it will be very useful for dynamic lighting one day, i feel it’s not that far either.
Oh right, i was totally fixated on the npc and dialogue-thing. For everything else in production(!) it’s surely helpful. Texture-Generation. Seamless. Audio. Music. Even level generation and whatnot, that’ll be helpful a lot. I let ai design my watchface-background as i would’ve never found one i like and I’m creative as a doorknob.
But what it lacks is theme. You can only fill gaps, but you can’t let it generate everything. You still need an artist with a vision and a given theme.to have a consistent art-style.