Tim Soret

@timsoret

Oh and I’d like to correct my estimate. With all the wasteful code that could be native, multi-threaded, optimized, or run in the GPU, we’re probably barely reaching 5-10% of efficiency, maybe even less.