4 ms·
> he was extremely dismissive of non-traditional approaches to graphics research (especially ML/AI-based approaches) Not necessarily a poor decision.
by antegamisou 4y ago
> he was extremely dismissive of non-traditional approaches to graphics research (especially ML/AI-based approaches)
Not necessarily a poor decision.
- ml-anon 4y agoBy necessity a huge part of the rendering pipeline is augmented and assisted by ML based approaches. AI models touch >50% of the frames you see in almost every film released and that number is only growing. So yeah, a pretty damn poor decision that would have relegated pixar to the past, if anyone actually still listened to that dinosaur.
- antegamisou 4y agoSo you're talking about the rendering part, which has obviously seen tremendous improvement from AI-based techniques e.g. neural material rendering, ReSTIR etc. Since Catmull is known for contributions computer animation research, where the application of ML hasn't exactly thrived, at least not yet, to the same extent as rendering has, I assumed you were talking about geometry processing.