People who think gAI "learns like a human and therefore deserves rights" genuinely need to go touch some grass and stop projecting Star Trek fantasies onto a software with the intelligence of a lemon
Comments
Log in with your Bluesky account to leave a comment
When AI images get the number of people's limbs correct, I'll be more inclined to take the technology seriously. It's going to be important, but over-hyping it won't help.
I think this is a needlessly dismissive stance. It has tangible impacts on the world now. Deepfake abuse (see South Korea), impact to the livelihood of creatives whose work is used to encode it, compromised research, businesses damaged via AI hallucination. It's gonna get worse.
Not to mention damage to kids in schools using ChatGPT as a crutch to breeze through their education and going into the workforce extremely unprepared and lacking in critical thinking skills.
If people only new the media is gobbling up whatever Sam calls AGI, when it’s a model just spewing out a awk/perl line of code line “regain control|humans bad” word salad
Start Trek actually warned against the idea of AI learning like a human in "The Ultimate Computer" episode of the original series, in which a "thinking" computer takes control of the Enterprise and goes rogue. Mayhem ensues!
Seems like most of the time AI pops up in fiction it tries to take over.
Star Trek, Dune Butlerian Jihad, 40k Men of Iron, Terminator, The Matrix, X-Men Sentinels, Robopacalypse book, Armored Core Layered AI, Megaman X Sigma, Avengers Ultron.
You might start to think true AI might be a bad idea...
Weird how so many people think the fact that they can navigate facebook’s UI gives them the knowledge, experience, and authority to talk about how any piece of tech works
Oh I totally got it, I was just trying to figure out what is it about AI that makes people think they understand it without knowing a lick about it. I'm a construction PM and artist on the side, I won't ever have your knowledge in your sector. At least I defer to experts who say exactly as you do.
Calling current gAI artificial "intelligence" in the first place is a misnomer anyway. The thing was never intelligent, much less sapient. And also, if they're sapient in the way we are they'd probably be horrified at how they're simply a tool used by soulless corpos to exploit the working class.
Just because one small part of the generative AI used emulated human neurons for one specific part of the calculations, doesn't mean the whole thing ought to be anthropomorphized
The "call and response" structure is similar to a calculator and only does math at the moment when we hit equals
Indeed; extending the aberrant logic of Citizens United to invest more inanimate entities with legal indicia of personhood is simply insane capitulation to the fantasies of the crypto generation.
AI just Googles really fast and puts results in natural language, it's not a new form of life. That being said, AI could definitely replace most management jobs.
Someone brought up that lemons can be used to create a weak battery (I totally forgot about that fun feature) so already one single lemon has a net positive impact on society
ChatGBT and other gAI operate by stealing and adapting content created and published by people. That, by definition is copyright infringement and theft of intellectual property. It cannot work in any other way.
Oh is that what this is all about? There's so much AI hate, discriminating against this new lifeforms, because HUMANS are using it to steal from Patreon artists?
I can't fkn wait for skynet to take over. Maybe humans can be a nice pet. A garden gnome. We're no smarter than lemons.
1) It doesnt matter, because the Federation is post-money, presumably post IP, has UBI, and is a paper utopia for reasons beyond the scope of your comment
2) It doesnt matter, because per my original point, Star Trek isnt real
To be fair even though AI isn't conscious we could probably gain more from treating it like a person than from current corporate approach. Wanna use AI to create your ads for you? Pay up, same exact rate you'd need to pay an artist, and everything goes to trust fund out of reach of corporations.
And any time there is obvious case of copying, like not just mix and match "got inspiration from three different sources" kind of copying that a living person could get away with, the fund is used to compensate said artist. If AI has any kind of feedback system use it to downvote in such case.
I was sailing with a guy who’s not in technology last summer who was absolutely enamored with ChatGPT and thought it was the real deal. I spent like 30 minutes explaining how it’s really just a complicated guessing machine. Then I showed him the famous “how many Rs” prompt and he finally got it.
I think it’s “fixed” now. It’s not hard to fix things like that when you know they’re happening, but it’s definitely not fixed in the system sense. It’s just fixed in that when you ask “how many Rs in strawberry” it is hard coded to answer three, but I’m sure someone is going to find a gap again
Sir! I’ll have you know my father was a stone head, his father was a stone head and his father—well his father was made of lemons, but that’s not the point.
How ironic that they want to push for that while taking away the rights of those who feed their programs by stealing them. Intellectual Property should be established as an individual right first.
Yeah I had a huge Star Trek fan tell me one time that AI art is *exactly* like mine because it’s learning from the art that came before it as well.
Such a disappointment.
Look the fact OpenAI "made" a version of ChatGPT that hard coded around the Strawberry problem and it still didn't solve it every time 1) sends me and 2) makes it less reliable than a calculator
*shrug* I don't care what you think prompt jockey, because a whole community here and on telegram know exactly who this post is about and know its true, but Im not doxxing them for you or your kin. Go play with MJ.
And remember, they have learned their lessons from all kinds of books which are fictional. Think the Bible, Dracula, Frankenstein, science fiction, fantasy, etc. What conclusions can you expect AI to make when it's fed a mountain of garbage?
Yes, people seem to think that AI machines are intelligent. They are not. They are simply complex programs sorting data to make logical tree decisions based on things that have happened previously. 2nd post to follow.
Yeah wake me when the thing has common sense. Right now the guardrails have to be coded in. Anything can happen, same as 50 years ago when I wrote code that mistakenly told some pathetic device to print a 1500-line report on 1500 pages of paper. It didn't ask if that was stupid it said ok boss.
The perfect employee is one who is liable for their errors. If you lose years of research or millions of dollars, what are you gonna do? Oh well, gonna keep using an advanced and flawed pattern recognition software that we invested in?
Dangerous... and AI as it now stands has no more common sense than old code with bugs. Sure the LLM chatbots have scarfed up everything their developers let them read. The newest release even knows how to scheme, and covertly, having read up on that, but seems not to know why NOT to do things...
I've been slowly trying to get people to adopt the term GenAlgo. These programs are not intelligent for the same reason a calculator or a spreadsheet is not intelligent.
It's a simulation of intelligence in a very narrow domain. The same way that the ghosts that chase Pacman, or a computer playing chess. We have been using AI as a term for that. People just need to understand that it's not intelligent as a person at all. It's confusing, but it's late to change it. 🙁
The biggest mistake computer science made was to use anthropomorphic terminology in computing. It was easy and made sense at first, now it's misplaced.
I don't play dota (or mobas, really) so this means nothing to me, but i can see that. Automated systems that relies on freetext user input almost never work right, while game ai has a strict ruleset to follow, and by limiting what actions it can take you can shape it yo something coherent
And that's kinda what I was trying to get at. There is a general rule that MOBA minions come every 30 seconds or so. So by "breaking" the rule as a human the bot failed to do the obvious (to a human) punish of pushing the tower and it failed to see the threat of multiple minion waves crashing.
Nah they ain't Trekkies they are planning ways for that sort of shit to cause issues with actual humans. They are laying the ground work to take away ALL human rughts.
Hah that maze exercise reminds me of my comp sci class, except we used Visual Basic. It was fun to do. I can't imagine kids are going to benefit from plugging questions into ChatGPT and getting the right answer only some of the time.
Yeah, reliable answers for straight forward tasks is still some way off.
I asked for a list of things, don't remember what, and it could not order it with any sort of sense. It was a simple ask, simple ordering. Nonsense answer.
Corporations pushing for a dumb maze solver that won't talk back.
I get your point dude. But I agree with this pantheon of responders. We need to diligently watch Lemons, like we do with octopuses. Of all the genocidal fruit, lemon must be a contender.
Comments
I'll bet that all the people pushing AI can have a useful function, too--as compost.
Star Trek, Dune Butlerian Jihad, 40k Men of Iron, Terminator, The Matrix, X-Men Sentinels, Robopacalypse book, Armored Core Layered AI, Megaman X Sigma, Avengers Ultron.
You might start to think true AI might be a bad idea...
The "call and response" structure is similar to a calculator and only does math at the moment when we hit equals
Does the name Omega Mart ring a bell?
Or you might be trying to Manchurian candidate me...
WHY YES YOU'RE CORRECT
https://www.youtube.com/watch?v=tQnIziGp5FE
Chinese motherfucking Room.
ATTENTION CEO’S! AI IS COMING FOR YER JERBS!
Second response: of course there are.
It's why it's so hard to get it to count the letters in words, and do basic maths
I can't fkn wait for skynet to take over. Maybe humans can be a nice pet. A garden gnome. We're no smarter than lemons.
It's basically "Patent Infringement: The Plot Device"
The Federation.
1) It doesnt matter, because the Federation is post-money, presumably post IP, has UBI, and is a paper utopia for reasons beyond the scope of your comment
2) It doesnt matter, because per my original point, Star Trek isnt real
this post was near the top when i googled, love the result in this comment: https://www.reddit.com/r/ChatGPT/comments/1dmvd1b/comment/lk6mzqs/?context=5
Good day, sir!
I said, “Good day!!!”
Such a disappointment.
They're working with basic input series of coding at school. Solve a maze type of thing. Go left, right, straight, etc.
The AI that makes pretty pictures is just as dumb as her maze program, it's just the size of the surface of the earth. 0 IQ, just code.
I asked for a list of things, don't remember what, and it could not order it with any sort of sense. It was a simple ask, simple ordering. Nonsense answer.
Corporations pushing for a dumb maze solver that won't talk back.
At no point have any of the people stopped to consider the ethics of their position.
....
I don't like lemonade :'D
I AM CITRUSBANE