AIs (well, LLMs, at least) aren’t coded, though. The engine is coded, but then they just throw training data at it until it starts parrotting the training data.

Humans can create scripts around the LLMs. Scripts that filter certain stuff out of the training data (though that can involve some pretty tricky natural language processing and can never really account for everything) or scripts that watch responses for certain keywords or whatever and either preempt the response from getting to the user or try to get the LLM to generate a different, more acceptable answer.

I think for poisoning to work well, we’d have to be creative, keep shifting our tactics, and otherwise do things in ways that can sneak past the LLMs’ babysitters. It would be a bit of an arms race, but I don’t think it’s as doomed from the start as you seem to think it is.

He seems to really want to be the president that makes the 51st state.

Puerto Rico is right there.

District of Columbia too.

When I bought my current car, I read the privacy policy and it says that they’ll record anything in the cabin of the car they damned well please and upload it to the mothership(/car manufacturer/Subaru).

For a while, I adopted the practice of repeating disparaging things about Subaru while I drove. I’ve kindof gotten away from the practice lately. What I really ought to do is find and unplug the OnStar MOBO to kill its internet connection. I’ll do that one of these days.

As for what you’re talking about, I don’t think LLMs (typically?) learn by your interaction with them, right? Like, they take a lot of data, churn it through the algorithm, and produce a set of weights that are then used with the ending to produce hallucinations. And it’s very possible (probable, actually) that for the next generation of the LLM, they’ll use the prompts you used in the previous generation as more training data. So, yeah, what you’re getting at would work, but I don’t think it would work until the release of the next major version of the LLM.

I dunno. I could be wrong about some of my assumptions in that last paragraph, though. Definitely open to correction.

So it would be pretty hard for them to hide a backdoor or something??

It happens, though.

I know the pain. I’ve worked at Windows-only places and places where the options were Windows or Mac but you were strongly encouraged to use Mac. Honestly, when I started at the latter place, I hadn’t touched Windows or Mac in like a decade, so as far as what I was familiar with, Windows and Mac were about the same for my purposes. And since most of the team used Mac, I just went with Mac.

The graphical system was terrible (to the point I even looked into what it would take to replace the default Mac graphical system (was it called “Aqua” or something? Don’t remember.) with something X11 based, but that’s like 100% impossible), but the thing that I hated the most was the touch bar. The Siri “button”(/“icon?”) on the touch bar was like one millimeter away from the backspace key (which is called “delete” in Mac for some reason, even though it acts like backspace). I’m sure I wasted so much time just closing Siri dialog boxes.

All that said, I’m not saying Windows would have been better than the Mac I had to use there. I probably would have been just as frustrated with Windows.

I’m lucky enough now to be working for an employer that lets me use Ubuntu. I disabled all the default desktop environment stuff. I unfortunately can’t get away with Sway because I need to use Zoom and desktop sharing doesn’t work with Sway, but I use i3 which acts virtually identically (and does support desktop sharing).

Yeah, Gnome is far to bulky for my taste as well. I use Sway. It’s one single process. And it’s a Wayland compositor, so I don’t have any separate process for the X server. And Sway is currently using less than 90MB of RAM on my computer. With nothing else running but a minimal terminal emulator, htop, SystemD, and various daemons, my whole system is using less than 480MB of RAM.

And that all makes me happy, of course, but just seeing small numbers isn’t really the point either. Aside from resource usage, I spend less time fixing, fighting with, upgrading, configuring, and otherwise maintaining Sway than I would KDE or Gnome or XFCE, and more time using my computer for the stuff I want to do on it. (As an aside, Sway’s tiling model is absolutely baller. I rarely have to think about where I want my windows, and when I do have to think about that, I don’t have to go to the mouse to position/resize them.)

KDE and Gnome are two different varieties of seven(-hundred?)-layer bean dips of dependencies atop dependencies. I like that my entire graphical system is one process with comparatively few dependencies that I can wrap my head around pretty easily. (And, honestly, Sway is a step up in bulkiness/heaviness/complexity from dwm, which is what I used previously.)

I meant more like not taking up shit tons of hard drive space, memory, or CPU, not having a billion dependencies, starting instantaneously, low cognitive load, etc.

It was kindof a sarcastic dig at KDE. I deserve downvotes.

Minimalism.

Yeah, I’m planning to switch from Arch to Gentoo. Systemd isn’t the only reason, but it’s a big one.

(Yes, I know about Artix, but it’s… kindof a Frankenstein’s monster, still mostly depending on the Arch repos and still with certain relics of Systemd. Or at least it was when I last tried it.)

Can you get bird flu by sharing needles with a bluejay?

Just about everyone looks better when they smile.

You’re the one who brought looks into it.