I get the disdain for GenAI, but are AI chips really the problem? Maybe they’re more expensive and price people out, but it’s not like they’re built on plagiarism like most generative AI models.
As far as I’m aware, they’re just capable of running highly complex multivariable calculi in parallel, making them more efficient for AI applications, but wouldn’t the same features make them better for more realistic physics and other game mechanics like procedural generation, NPC pathfinding and behaviors, etc.?
I guess it would suck for anyone who doesn’t have the hardware to play a game, but there could always be options to configure in the settings to make it playable, like “don’t use tensor calculus in game physics” or whatever
I get the disdain for GenAI, but are AI chips really the problem? Maybe they’re more expensive and price people out, but it’s not like they’re built on plagiarism like most generative AI models.
As far as I’m aware, they’re just capable of running highly complex multivariable calculi in parallel, making them more efficient for AI applications, but wouldn’t the same features make them better for more realistic physics and other game mechanics like procedural generation, NPC pathfinding and behaviors, etc.?
I guess it would suck for anyone who doesn’t have the hardware to play a game, but there could always be options to configure in the settings to make it playable, like “don’t use tensor calculus in game physics” or whatever
Far as i know, GPUs are more specialized on vector calculations. Some upscaling/frame generation techniques use AI hardware but that’s it.