Serpent OS has a lot of focus on performance - will this extend to GPU compute and related development i.e. include the ROCm (most important for me personally), CUDA and implicitly the OpenCl/SYCL development stacks?
I use a lot of scientific and engineering packages which could greatly benefit from GPU acceleration. Of course there's also AI everywhere so support for GPU accelerated PyTorch, TensorFlow... is probably a must anyway.