Mike Gorse

I see a lot of posts on my feed expressing various ethical and pragmatic concerns about AI. And I see quite a few posts from other disabled people along the lines of "stop telling us not to use AI when it's helping me fix accessibility issues that you've never done anything to fix." I worry about the ecological effects of AI's need for resources in the aggregate, but a few disabled people making use of it isn't going to make a meaningful difference, and I get the argument that we need all the help we can get. So I am not going to tell someone else whether to use it or not. Whether it is helpful is a more complicated and nuanced question, and I think that it will depend on the context. I've been avoiding it mostly for reasons specific to me--I just vaguely feel like I don't know how, and it normally wouldn't occur to me to use it. I feel like I wouldn't know what to give it for a prompt. So, admittedly, I'm not the best person to be giving opinions. I wrote about it at the time, but, a few months ago, someone gave me an AI-assisted patch for a bug in my module. At first I was impressed, and I still am to some extent, but the patch was later found to have problems, and the person came back with a revised patch. That, too, caused problems. I eventually removed all of the complicated code that the AI added and solved the bug with just a few lines of code. To be fair, the AI patch helped me to figure out how to think about the problem. So that was my first impression.

I see "PDA" in someone's handle, and I wonder what it means. Personal data assistant? Probably not; it isn't 2005 anymore, after all. Progressive Democrats of America? Probably not that either. Pathological demand avoidance? Probably not. Public displays of affection? Maybe. Or something else?

Oh hi It's another day. I check my work email this morning, and we have 11 new bugs open for FreeRDP CVEs and two new bugs for Gimp CVEs. It isn't clear to me if I'm going to work on them or if someone in Beijing is going to work on them, but someone needs to work on them. Meanwhile, I'm trying to squeeze some API changes into at-spi2-core, and I'm already past the freeze, so I'll need to request a freeze break, so trying to get that done and feeling I'm neglecting the day job stuff that I should be working on. Oh, and apparently there's a crash that I need to look at, too. Yay me.

Just saw that WebKitGTK 2.50.5 is out.
In general, I've noticed that WebKit's Linux a11y code has issues (I get a crash if I try to authenticate a gmail account in evolution, for instance), and, alas, I have too many things on my to-do list right now and don't have time to try to work on it. But at least the role mapping regression is fixed.

I'm noticing that, when I hear "AI," I think of a large language model, but AI can mean many things that can be very different from each other.