Sure, but we've been building houses for a long time. We know a lot more about building houses today (or even 40 years ago) than they did when the practice of house-building was still relatively new. And most of this knowledge didn't come from building houses over and over, but from developing better theories of how buildings stay up, using better materials, etc....
I think we'd both be surprised on how much actually does come from "building houses over and over." My assertion being that much of material science advancement came from required advancements from previous failed houses.
That's my same experience with programming. That's kind of what I was getting at in this article actually: that you get better by doing your best and making mistakes. Nothing beats hard-earned experience.
