Someone asked me the other day if #accessibility has been getting better over the last years.
I feel that's a difficult question to answer. Relatively speaking, maybe, in some aspects, yes it has.
But if I'm brutally honest, better does by no means mean good.
Right now, I'd say accessibility as a whole can be described thusly:

People have, on a whole, felt largely ignored and essentially akin to a dog begging for scraps at the dinner table. Maybe every so often someone somewhere decides to do something, but in general, accessibility is ignored, deprioritized, devalued and at times derided by those who should often know better.
Accessibility is in a state where people would rather use a planet-destroying, ethically questionable slot machine to fix things, because the human slot machine of getting people to care has far worse odds and requires far more effort with often nothing to show for it.
And yet, that probably is still better than a decade ago.

One can argue that using those questionably ethical technologies is just that, ethically questionable. I would very much agree.
And yet... We've had the internet for a good 35 years now, things haven't exactly improved, see also the webAIM Million. See also #Foss project after #floss project with absolutely zero regard for #accessibility until it's way too late.

Ethically questionable, sure, but I can not in good conscience blame anyone who's waited half their life to finally see things improve only to be disappointed taking things into their own hands.
Switch away from big tech? People can't. Pick up any old alternative when a tool of choice funds the far right? People can't.
Stop funding the billionaires? People very well might not be able to.
And that lovely bit of entrapment was all human-wrought. No AI, or i'm sorry, AL, required.
This is the world we live in, and these are the hands we're given. Don't like it? Send a PR and fix it.

There's definitely been some encouraging developments over the last 5 or so years. Orca is finally starting to shape up into a proper #screenreader after existing what... 18 years or so? Perhaps we'll see a viable #linux distribution that's #accessible enough to be someone's pleasant productivity machine. Heck, @elementary
seems to be fighting that fight almost all by themselves at this point and they're actually getting somewhere, although the work that must have taken honestly scares me. You're braver than i! :)
We're seeing Wayland accessibility improve, bit by bit, phenomenal job being done over in those parts but again, probably swimming against the current to a scary degree. Whoever's working on that now, you're appreciated!

@zersiax @elementary Orca is, as far as I know, mostly Joanmarie diggs' work. If she had a team of at least ten (I'm not saying fifty or three hundred!) permanently available developers, it might have happened earlier. I mean, becoming a viable screen reader. You know I'm super skeptical about Linux desktop, but Joanmarie deserves to be mentioned every time we talk about it.

@menelion @zersiax @elementary Yes for sure, she did it for years but I thought that she stepped down. Can't find a source of that now though. Besides, the mix of window managers, desktop environments and relatively old accessibility APIs seem like a hard environment to build a screen reader for to me