Open-source should have speed up the evolution of new ideas and yet more often than not it's just keeping old proprietary software barely alive or duplicating it.
You have to take into account that the open source software constitute a very small fraction the overall software both in term of quantity of products and in term of developpers involved.
Also there is a lot of innovation in open source software, the thing is those are often not seen nor directed to end users. Because many open source projects are initially started by developers and not by product design and marketing teams, the most successful open source projects that make products that solve developers problems or meet their needs.
Taking blender as example of a successful open source project, with a strong incentive for developers to create a great tool, you still have the problem that proprietary software is probably better, otherwise major 3d studios would use blender.
What are better proprietary alternatives to Kubernetes, Nomad, Docker, Podman? Aren't the most widely adopted web servers and frameworks open source? I am only naming a few areas.
One could argue that since these projects are often developed by contributors who have proprietary software in mind, it is in fact proprietary code wich creates the collaborative open-source effort in the background.
Money may make the motivation much of the time, but that doesn't mean that OSS isn't where much of the innovation happens.
The same is true of games. Counter Strike started out as a free Half-Life mod. DOTA started as a Warcraft 3 mod. Both spawned entire subsections of e-sports that are still wildly popular today (tactical FPS arena and MOBA respectively).
I worked at Tangent Animation, now defunct - we used Blender as our primary render software, and produced several feature projects, two of them still running on Netflix last I checked. We were not major, but it has been done.
The reasoning was that we had paid a very large sum to use Maya on our first project, and had received basically no support whatsoever for that money, aside from the software itself. So we worked with the Blender Foundation instead, and got much better support, collaboration on new features, etc.
I don't have any personal exposure to any of Blender's commercial competitors, but I learned to make cartoons as a side effect of working there, and Blender basically fills me with awe at what they have accomplished. Perhaps Maya is even better than that, but that's for people with the money to pay for it to decide, I suppose. From what I heard while working in the industry, Autodesk is a husk of a company with no interest in anything but swallowing smaller companies and increasing profits with as minimal investment as possible in new features.
Meanwhile, the whole world has access to Blender, and it works just fine, if you want to make cartoons and have ideas and ability but no money. "Better" is extremely subjective here.
Proprietary software is software that keeps the lights on for the people that work on it.
Sun Tzu has figured out a while ago that it's a bad idea to back your enemies into a corner. Like it or not, fighting with your back against a wall makes you lethal.
Now, I'm not saying open source is unable to match this. They can and have. I'm just saying that most proprietary software has that advantage naturally (in a capitalist society).
Proprietary code appears to create all of the innovation because you have to market and sell it. I don't think you can meaningfully say "which creates more innovation" in either direction, though if pressed I'd say, "open source in the long run" -- e.g. Linux runs a much greater variety of things than Windows.
In general, I don't think this is true at all and much of the most innovative, boundary-pushing software is open source.
It might be more true for open source games specifically, due to the economics of creating games. But as someone who is a casual/nostalgic gamer at most, that's absolutely fine by me.
