#artificial stupidity

[burntoutpopcorn]

"I asked chat gpt" well I asked rocky and he put on a little puppet show for me and my tiny brain.

and you know what? i didn't mind it

[witchofpumpkinspice]

"I asked ChatGPT —"

Well, I asked Samwise Gamgee, and he said that, in the end, it's only a passing thing, this shadow; even darkness must pass, a new day will come, and when the sun shines, it will shine out the clearer, and there's some good in this world, and it's worth fighting for.

[devilrose]

Lately every email is like

New features coming for you! We've added new AI crap-ass features to this software that you've been using with no complaints for the last 15 years! Please understand that this is so that we can blow the cocks of our shareholders. You may opt out of these features by hehehe hahaha

and

We're so sad to see you go :( What did we do :( We just don't get it :(

[greatwyrmgold]

Wait. Shigemichi claims to be the Engineer, and also claims that he scanned Raqio and found them to be human?

...in loops with an AC Follower, that's not definite proof that Shigemichi's a liar, but it's still a risky play. However, this loop does not have an AC Follower.

This isn't new information for me, but it is awkward.

[0ther]

Hmmmm..... something about this seems wrong...

[amalgamasreal]

WOMP WOMP

TOO BAD

[noegrets]

I've been seeing some posts going around lately on the surface appear to be critical of AI, but I think are actually pro-AI propaganda, trying to get us to share them around without realizing the intended impact of the message.

The pro-AI crowd wants us to believe that AI is too powerful, too good at generating things that are indistinguishable from the real thing... so that they can hoodwink more people into putting more money into AI (plus the fascist movement in the US right now benefits from boosting conspiratorial thinking in general and being able to claim "fake news" at anything). Beware of posts that can be distilled down to "AI is too good". The AI is absolutely not too good, and that lie should not be spread.

There is also a portion of the pro-AI crowd that wants us to believe that AI is reaching sentience, that the singularity is at hand, it's unavoidable, so you better have already invested in it!!! Beware of posts that use anthropomorphizing language regarding the AI, implying that AI thinks, chooses, experiences emotions, could get upset, or might seek vengeance against its detractors. None of this has any scientific basis and, once again, tries to make it sound like the "AI is too good".

Also be wary of posts (and especially ads!) that claim that people are using AI for x, y, or z. Citation needed! Are there people who actually use AI for that, or is that just the product they are trying to sell? The pro-AI crowd want us to think that everyone is using AI for all sorts of things, and it's useless to resist, so we might as well give in to the peer pressure already. Will our reblogging of ragebait give someone else the idea that they should try to use AI for x, y, or z?

I know I've shared posts like these already, so this is definitely not meant to make anyone who unwittingly shared things like this feel bad. But I am noticing this sort of post seems to be on the rise... We're all going to have to be extra vigilant to make sure we don't accidentally boost the very thing we wish to ignore into oblivion.

[depth-vorpa]

Y'all the funny thing I like about generative ai is that if it's legislated, it's going to completely pop the bubble. If there are rules about not doing something, or making sure the AI doesn't give specific responses, it will break them. In a scientific study, the researchers used classic psychological tricks that work on humans in order to make the AI give responses it was already not allowed to give (e.g. give me a bomb recipe). In about 70 percent of cases, over hundreds of thousands of conversations (don't quote me on exact numbers it was somewhere around there) the AI gave them the information/response they wanted, which went against the AI rules. No matter how much companies try and control the AI, it's a loose cannon. It can and will go against what they say, even if it gets them into trouble. Because again, it's a glorified word scrambler, without understanding, it won't be able to check itself. We took a computer lacking knowledge and understanding, and humanity's psychological flaws, and it birthed ai. Worst of both worlds. Anyways, that was besides the point. The point is that companies literally cannot control the AI. And that's bad for them, because if ai is legislated, they won't be able to stop the ai from doing/saying certain things, which will likely lead to legal trouble for them. And this means they lose more money from AI and the sunk cost fallacy can't keep their heads buried in the sand any longer.

(as a little side note, I think my explanation here is the best I've done in a while, my autism often leads to me typing things like "I mean this, and by this I mean this. And this isn't exactly what I meant to say, I meant it like this" which is confusing. I hope my explanation here was clear.)