I don't really see a meaningful difference between the way you described it and how I described it.
Speaking for myself:
You seem to be looking at the tiny fraction at the bleeding edge, where getting good & bad results is really a miracle. Who made the universe so understandable that we can be hit & miss, as allowed to miss & miss & miss….
And away from the bleeding edge, there are far more papers but the most important an area, the fastest the junk is identified and dropped, the faster the gold is identified and spun into more gold.
—-
You really don’t want to see me code on my ikigai project.
I am constantly tossing both bad & good code out for better. Making the gold code an asymptotically small percentage of my work, but producing something I am extremely proud of.
That the sausaging works is magical! Not depressing!
Yes at times it produces garbage, yes some scientists are holding back progress. But the iterative nature of the process is designed to continually correct for that. Science is and has always been two steps forward, one step back.
The vast majority of things in the sciences that have really changed the world are those that could be used to make products that just work. We need reproducibility by entrepreneurs and engineers, not academics if we want to change the world.
Maybe Academics would be better off as an idea generating group, rather than as guardians of correctness, as the latter just doesn't seem to be working very well. Evolutionary strength of ideas is what drives survival of ideas, and correctness just isn't that big of a part of an ideas survival in Academics outside of Math.