I actually disagree with that, which is why I'm a bit annoyed at the state of keyboard enthusiast communities. I think that there's plenty of room for improving keyboards, not necessarily for better performance but mainly for more comfort.
Over the past decade or so I've been using more and more "esotheric" layouts and mappings. I'm typing this comment on an ergodox mapped with a pretty heavily customized dvorak layout. And you get get much deeper in this rabbit hole with custom 3D printed keyboards like the Dactyl for instance. I've been working on a custom model in OpenSCAD and I'm looking at buying a 3D printer to experiment, I'm looking forward to using a custom Frankenstein'd keyboard in the future.
So clearly, I don't think keyboards are a solved problem. There's plenty of room for experimentation. What about foot pedals? Palm keys? Integration of touch pads or trackballs in the keyboard? What about weird modifier keys, for instance using shift keys to input ( and ) when not used as a modifier (I've been using that for about a year and I love it). What about adding cameras and some processing so that you can use hand gestures on top of typing on the keys? A compromise between "minority report" style interfaces and a good old contacts, with the strengths of both?
But I agree with your overall point: it seems like the bigger hobbyist communities around keyboards these days are about collecting and making fashion statements, not hacking keyboards. Just look at https://www.reddit.com/r/mechanicalkeyboards , if you like pictures of keyboards and keycaps (and ridiculously overcomplicated cables) it's amazing but beyond that it's mostly boring old typewriter layouts and qwerty. That's a bit disappointing IMO, but I guess it's just not for me.