So now instead the bottleneck is intellectual, is finding people who can learn the most from a limited number of runs.
It'll be interesting to see how tbis gets metered but probably wont really achieve the ibjective of equalizing. Some big company's going to have a vast serverfarm with latest greatest gpus and interconnects & run simulations for the same hours but at 20x the detail, because they had the serverfarms & well tuned aoftware on hand. Trying to democratize this intellectual pursuit is an interesting challenge; what an experiment F1, good luck.