I’ve always wondered what the modern equivalent of the NeoGeo or 3DO would look like, capability-wise.
Would it abandon traditional polygon rendering for gaussian splatting?
Would it be able to do movie-quality path tracing?
Would it have multi-level hardware-based procedural generation (each level getting more precise in terms of scale)?
Would it be a SoC the size of someone’s head?
Would it have a CPU that isn’t measured in normal cycles but instead has sub-PUs coordinated by the CPU which run at their own speeds (with PUs for simpler instructions running faster than PUs that incorporate instructions that’d traditionally take multiple steps)?
Would it have an AI processor on par with an Nvidia A10 (or whatever the current top-of-the-line AI chip is)?
I’ve always wondered what the modern equivalent of the NeoGeo or 3DO would look like, capability-wise.
Would it abandon traditional polygon rendering for gaussian splatting?
Would it be able to do movie-quality path tracing?
Would it have multi-level hardware-based procedural generation (each level getting more precise in terms of scale)?
Would it be a SoC the size of someone’s head?
Would it have a CPU that isn’t measured in normal cycles but instead has sub-PUs coordinated by the CPU which run at their own speeds (with PUs for simpler instructions running faster than PUs that incorporate instructions that’d traditionally take multiple steps)?
Would it have an AI processor on par with an Nvidia A10 (or whatever the current top-of-the-line AI chip is)?