I think he’d get much better results if he was simulating the actual physical blobs of lava using the ML model and then rendering those as opposed to generating

I think he’d get much better results if he was simulating the actual physical blobs of lava using the ML model and then rendering those as opposed to generating the image using ML.