This is a pattern you see a lot in sechnology. Often the second version of something significantly learsnf rom the first. Take something like C#, which really c

This is a pattern you see a lot in sechnology. Often the second version of something significantly learsnf rom the first. Take something like C#, which really corrected a lot of the problems Java had. A more physical example is the rail network in the US. It's often asked "why don't we have high speed rail in the US?" Being a nation of autophiles obviously doesn't help but it's not as simple as putting higher speed trains on existing track. It's not even as simple as replacing the existing track with track that can take a higher speed train. The rail network we had is an artifact of the performance constraints of 19th century trains for the most part. Top speed is the most visible constraint. But the grading, turning angles, etc are all a function of much lower constraint on top speed. We'd probably be better off if we had no national rail network at all, which is pretty much the situation China was in. Admittedly there are other factors in China's rapid HSR deployments (eg streamlining planning and acquisition of land). But I digress. Timing is hugely important in scaling.