

File format standards certainly, and OSS generally embraces those (at least if they’re non-proprietary), but UI doesn’t have to be standardized. On the other hand though not everything needs to be a unique snowflake. UIs should take the things that work well and experiment with what doesn’t.
Lets also not pretend that proprietary apps don’t screw around with UI design just as much. I can’t count how many times now Microsoft has redesigned the UI of something that was perfectly fine and didn’t need redesigning only to end up making it worse.
I think part of the problem is that “good” UX isn’t a single thing but a continuum. It’s very dependent on the skill level of the user. Often what makes a good UX for a newbie is a bad UX for a power user and vice versa. OSS tends to attract power users and particularly the ones working on some software in a particular area tend to be domain experts. That in turn can lead to designs optimized for very advanced use cases that end up being frustratingly opaque to an “average” user or even worse a newbie.
Blender is an excellent example of this. It’s regarded as one of the best 3D programs out there but it’s far from a simple piece of software to pick up. What saves it is that all the commercial alternatives are just as obtuse as it is and so the ground level expectation is that it’s going to be complicated.
Likewise many OSS and Linux tools expect or even require CLI usage which while great for power users putting together scripts and pipelines are often opaque and unintuitive to someone who is still learning the domain.
This focus on power users leads to turning newbies away and funneling them towards the commercial offerings where they then get used to their quirks and limitations of those apps so that when they do eventually become power users the quirks and limitations of the OSS alternatives feel strange and off-putting to them.