Also unlike a video game, the tech-tree dependencies are not obvious. We don't know in advance whether a tech will be a productive avenue of R&D, or if it depends on other unrelated advances in theory, material science, or computer modeling, or was actually completely unfeasible in the first place. I find wind power [1] as a case study representative, and we see similar stories for e.g. fusion, cancer research, AI
[1] https://constructionphysics.substack.com/p/why-did-we-wait-s...