I think more reliable in Silico experimentation will yield much better results in the long run but is probably akin to a spacex or Tesla type of investment and

I think more reliable in Silico experimentation will yield much better results in the long run but is probably akin to a spacex or Tesla type of investment and 1-2 oom more compute intensive.