Progress in science as I understand is only on the shoulders of giants (previous research), not starting over from first principles with every generation. If we did that we'd get nowhere, each generation would get to the wheel or so and then the next one would start over.

I don't think gp is talking about continuous refinement of ideas, which I get the sense is more what you meant. Spending time and energy reinventing cubicles again and again every 10 or so years seems best to avoid, in my opinion.