Much of this depends on your time scale, of course. If you stretch out the time granularity just a bit, we really did make a pretty rapid jump akin to a step function.

Indeed, I wonder if historians will look back and drop all references to Moore's law, and just recognize that From 1950-2050, we developed the whatever-the-end-game-is transistor, and pretty much ignore all the interesting (to us) developments that got us there.