I remember when it bothered me that .docx is a proprietary format that requires a proprietary software to open. Both OpenOffice and LibreOffice suck big time back then. I have annoyed a lot people asking to use other formats. And then web-applications overrun the world.
Web-applications obsoleted most of free software notion. Four freedoms were still important to developers and system administrators, but users no longer had them. You can setup your git(1), but you are going to use gitlab the way somebody else decided for you. The same "eat it and like it" power dynamics as with proprietary software.
I remember when it bothered me that software was written using huge, bloated graphical frameworks that re-implement basic data types and include CSS parser. And then Electron appeared, bundling full browser engine into every messenger and calendar application.
I remember when it bothered me that people took performance penalty of virtual methods (aka "jmp eax") lightly, and then, thanks to my $dayjob, I learned about notion of low-code, which means, buzzwords aside, adding numbers using HTTP REST request.
I remember the init system controversy1 in Debian. It overshadowed almost every other issue I ever cared about in computer world, and yet now it feels small and insignificant with advent of AI. There is multitude of problems with AI, but I begrudgingly admit -- it is occasionally useful development tool, like gdb or valgrind, but unlike other tools, it costs non-trivial amount of money.
https://andrewkelley.me/post/renting-is-for-suckers.html
https://drewdevault.com/2025/03/17/2025-03-17-Stop-externalizing-your-costs-on-me.html
I have always been attracted to the "by programmers, for programmers" motto. The beautiful idea of free society building something for themself; society where foundation of power is skill, not money. Independence and meritocracy2. Sure, it just an idea, and it has been twisted many, many times before. You can't buy decisions directly, but you can always hire people with skill do your bidding. Full time employee has a lot of momentum in any volunteer community.
But still, there was an indirection layer. Now, there is money factor on previously even field of skill3 -- there are developers who can/agree to pay for Google/Antropic/OpenAI subscription, and those who don't. And from where I stand, prospects of both groups are grim.
It was important for me that my Free Software work didn't cost me money. Spending time writing code and having a discussion on a mailing list is a pastime no worse than any other, but if I spend money, I expect a payoff more substantial than a warm fuzzy feeling. I think many people feel the same.
On other hand, community built around "we don't do AI here" principle will explicitly be at disadvantage4, and that disadvantage will grow over time. Some people will definitely embrace the challenge, but over time these projects and communities will be more and more like Kolibri OS.
From these observations, I can draw only one conclusion:
No matter what you do about a problem, it will be rendered irrelevant by a
bigger problem, so why bother?
That sounds awfully like fate, and it sucks.