The work is very different. The work to make a new product like this is completely different from the work to improve an existing, large, established behemoth of a project that has a large number of users. It's a software lifecycle thing, analogous to the difference between spinning up a mini startup, and working for a settled big company with tens of thousands of co-workers.
Perhaps the big product would be better if everyone worked on it, but you're talking about unpaid authors, doing what they enjoy in their free time, often for fun.
What's the reward to the unpaid author of refining the existing, established, behemoth of a solution?
They aren't going to be able to try out their own big new ideas or do any significant architectural design that way.
They aren't going to be able to take it in a new direction, because the existing tool has a lot of users so can't change that much.
They aren't guaranteed that what they want to work on, or what they want to see as features (maybe to use for themselves) would be accepted upstream.
They aren't going to be able to change anything that existing users depend on really. For example if they wanted to try a different syntax for formulas, or a different editing paradigm.
They aren't likely to end up with something they can point at when asked what they spent all their free time working on. Contributing to such a large project, if they enjoy doing it that's great, but nobody outside a small group of people will associate them personally with their bit of work - unlike an independent project, if that gets a reputation.
They also have no chance of monetising their work later, if that's what they choose to try to do in future. Or more likely, to make a commercial branch.
For some people, that kind of work on an established project like that suits them. But it's completely unsatisfying kind of work for other people. It's not reasonable to pressure them, through oblique criticism, into doing it unpaid if they wouldn't enjoy it.
For example, there are a number of fledgling open-source nonlinear video editors (NLEs) -- to name a few in no particular order Kdenlive, Openshot, Olive, OpenMovieEditor, Blender Video Sequence Editor, Shotcut, Cinelerra, and Flowblade. I say this with full respect to the authors and contributors to those projects, we just need ONE open-source NLE that is on par with Final Cut or Premiere Pro.
But we really can't blame the authors - there's no money in open source user apps and polishing isn't fun to most people. I'd immediately jump on the opportunity to build an NLE from scratch if given the time and financial freedom, but I can't see myself digging though someone else's code and sanding all the rough edges for fun.