[Zababa]

I think the case of Fabrice Bellard is one where working on building the tools instead of the product gives you a way bigger reach. For example, if you're doing webshit react in a company, if you spend time automating the testing infrastructure instead of producing yet another feature, you'll have a disproportionated impact compared to the person that worked on the feature. Of course Fabrice Bellard is still way better than the average developer, by a very large margin. But you can improve your "output" without getting better, just by picking the right problems.