GNU, Linux, the free BSDs, etc. have done a lot, but in the end they have only copied the work that came before them. You can see it from the surface down-- the GUIs are copies of Windows or Mac, and the OS kernel is a copy of AT&T Bell Labs Unix with ideas from other newer OSes.
I've come around to the view that the basement/garage hacker who changes the world is largely a myth. It's a very popular one and a very American individualist one, but I see few if any real world examples.
Most people around here have probably seen this:
https://www.youtube.com/watch?v=yJDv-zdhzMY
This wasn't the product of basement hackers. It was the product of enormous amounts of DARPA money used to pay a good number of very smart people to focus on the problems of human-computer interaction full time for many years.
I want to believe. Please show me the basement/indy hacker equivalent of that demo.
Sure, basement hackers can make innovations... a little here, a little there. They can also do an excellent job making copies of commercial/military stuff, and in some cases the copies are a bit better than the original. I'm not saying this work has no value. But it doesn't innovate much and it certainly doesn't invent. Invention is a hundred times harder than innovation.
IP sucks. It's a bad system. But consider what it is attempting to accomplish. It's trying to create a mechanism whereby really innovative work can really pay. It's attempting to reward investments in really pushing the line forward. In its present form it does a poor job of that, but that's the intent.
Edit: I think the same logic also applies to startups, but to a somewhat lesser extent. Startups do sometimes have money, and can therefore focus on problems and employ people. But they must get revenue fast, which hugely limits their ability to play and experiment. As a result startups can innovate -- improve existing things -- but they generally cannot invent. Most startups that attempt to invent run out of runway and crash in the weeds. Invention seems to require a financier who is willing to throw money away, such as DARPA, NSF, etc... traditionally either nation-states or really freaking huge companies with money to burn. Invention is also hard for startups because there is little to no way to monetize it. That's the problem that IP attempts but largely fails to solve.