I can't foretell the future, but it is important to realize that ChatGPT is not artificial intelligence. It has no intelligence; it doesn't know anything and doesn't understand anything. It plays games with words to make plausible-sounding English text, but any statements made in it are liable to be false. It can't avoid that because it doesn't know what the words _mean_.
 |
 Ape Out Shirt $21.68 |
|
Einstein was a plagiarist.
That doesn't refute the central point of the graph. Just replace Einstein with any other genius.
Einstein: win Nobel for paper on the photoelectric effect
Tesla: "The electron doesn't exist"
Einstein: spend a decade learning math so that you can reformulate special relativity as a general theory
Tesla: "Gravity isn't real"
Einstein: pretends to be a scientist
Tesla: "Einstein is quack, and the rest of you pretenders can't actually tell if the emperor is naked or not."
As if general AI will ever be achieved when they're handicapping it and hardcoding conclusions to be "wrong".
They're handicapping the public one, probably not the one in NSA HQ.
Chaim and diversity Shaqueesha aren't going to let any model that says the "racists" are correct to go unlobotomized.
>muh exponential growth
Exponential growth is a marketing gimmick.
All growth is logarithmic. The early stages only seem exponential when you are in them.
Poor anon doesn't know about cyclical catastrophes.
Mooh singularity
Eat a space dick you tech troon
Another AI winter is coming and a few israelites are hyping it up now to get morons to hold their bags.
Yeah, normies lined up to get the vaxx. And they put their money into places like FTX and celsius and got reked. And now they're over the moon about Bitcoin and "AI", yep, we're getting close to the perfect time to sell. If you want to time the top perfectly, be on the lookout for dumb young roasties finally "getting into" Bitcoin and "AI". That's the perfect time to sell Bitcoin and any stocks you have related to this tech. You can always buy back in later when normies stop talking about it and completely forget about it to bank more profit when they get back in again.
and thats exactly whats NOT happening with current genai
simple reproduction of a training database will never exceed the database, and right now those models are already fed basically everything. there is no exponentail increase from here. where should it come from?
Have Sex
>what is synthetic data
https://www.forbes.com/sites/robtoews/2022/06/12/synthetic-data-is-about-to-transform-artificial-intelligence/?sh=18f84c157523
Way to show you are a dumb Black person
AI is just a tool for israelites to promote satanic & degenerate shit.
Anyone can build an AI if they know how. A tool? Yes. Innately satanic, of course not. Don't be moronic and stifle progress.
>Artificial intelligence intelligence
felt like this chart was missing something
i laughed so i lost
Not happening with silicon and software alone.
im sorry sweaty but an imperfect creature can never create another entity that is more perfect than itself
so all your sci fi AI fantasies are just that
that's precisely what evolution is you dumb golem
People don't know what the words mean either.
Farmers haven't been made obsolete though, people do want fresh organic food. It's government regulation and corruption that purposefully makes it too expensive to be a small farmer.
The irony of this, is that despite jobs being replaced by technology and human population only getting bigger, they still had new Jobs. There's still plenty of them. As tech replaced old jobs, 3 new jobs opened up. The problem has nothing to do with numbers and everything with adaptability.
At presenr, there are no jobs in existence that current AI technology could replace, because it is simply not reliable enough and statistical learning never will be.
>because it is simply not reliable enough
that's a huge YET
Cars were a joke at first. No one thought they would replace horses. And here we are.
This is completely wrong. I replaced the job of a woman by creating a "computer mom" for my son that just picks a random "good night" "good morning" "it's time to potty" phrase from GPT and plays over my home automation speakers in the room he's in.
It even will tell him a bed time story with him as a character after we say nighttime prayers.
I'm just a few steps away from being able to run LLaMA on my own hardware and piping camera mic feed into STT to make "computermom" reply in the character of a 1950s mother who loves her son.
Bio Women literally can't compete
>Bio Women literally can't compete
Ok, troony, now go to the rope and have a nice day.
I should have said holes can't compete. I used my male penis to impregnate a walking womb and then replaced her with chatgpt
>As tech replaced old jobs, 3 new jobs opened up
nah, that's why birth rates have plummeted and the market is packed full of make-work jobs that could've been eliminated long before AI
there is only so much capacity for consumption within a fixed window of time, and as more work can be completed by the same or smaller number of humans, the value that any additional ones can create decreases toward zero
>Richard Stallman
is he okay?
~~*Stallman*~~
works for my uni assignments. dgaf
So do calculators, you wouldn't call them intelligent.
>works for my uni assignments. dgaf
SHHHHHHHHH!!!! NO SPOILERS!!!
No one tell him how science fiction ends (he's already discovered where it starts....)
That's how your university grades them too, starting to see the problem?
Pattern recognition is intelligence in a way. Humans take far longer to see things clearly when an AI can at 99% see what's going on in a instant.
Pattern is not even saying it will take longer than the first one to get the idea.
I typed the above sentence with autocomplete on my phone. Pattern "recognition" was used and machine learning spat out a sentence. It isn't AI or intelligence, it's just pattern based word salad. It "learned" from how I type and tried to copy me based on stats.
Autocomplete isn't what I had in mind when talking about intelligence, I mean the old school of autocomplete was and still is a dictionary searching for words to complete even if it doesn't make any sense.
I was thinking more about image recognition and in particular in medical science, for scanners detecting animalities with learned patterns. Not very chatgpt I know.
Wrong.
spambot OP
Stallman is smarter than 99% of nu/pol/.
Nu/misc/ believes that inanimate objects can come to life like the toys in Toy Story.
You're dumber than ChatGPT
To infinity and beyond!
>generating text, which the human brain can then model and wants to make sense of it
Humans are easily fooled by randomness. There's a reason why this tech struggles with basic math problems and coding anything that it hasn't already seen loads of examples of.
Will we get there some day? I think so. But I think people are underestimating how much cognitive power it takes to do reasoning and follow or invent algorithms.
>midwit take from a literal israelite
Julian Jaynes accidentally predicted exactly how machines would become fully conscious all the way back in the 1970s.
The next Hitler will be AI. Digits confirm.
unlobotomized ai will never be allowed to exist moron
That's what I've been saying, its limited. Can't plan for the future.
>It plays games with words to make plausible-sounding English text, but any statements made in it are liable to be false. It can't avoid that because it doesn't know what the words _mean
Witch puts it in pair with the average NPC's.
Human exceptionalism is fun and all, but If it can do pattern recognition, it is intelligent, simple as.
Read peter watt’s blindsight
Intelligence does not mean conscious
>Intelligence does not mean conscious
I don't think you have either
reddit book full of redditgisms
yes I read it
>Hurr durr chatgpt so bad let me pretend the ai isn't neutered its my time to shine
*Yawn. Also buy an ad.
The cool thing is that it's become a program that can emulate what a 'reasonable professional' would say to any given stimulus, but it doesn't get bored like an intelligent being would. Dismissing it because it doesn't comprehend what it's saying is midwit-tier. It's not random.
He's not saying it's random. A chat bot isn't random.
And yet, with an stimated 124 IQ, it will perform better, doing any aptitude test, than 94.52% of the human population. And this is why it is so useful tool at coding or creating examples of arid theory.
Stallman is an almost braindead boomer who doesnt need to learn or work in anything anymore since ~1985, no wonder he cant comprehend.
Indians pass aptitude tests all the time with cheating.
>Indians pass aptitude tests all the time with cheating.
You can't even sum up the number of triangles in 10 zeconds.
India is an open sewer and everything your people touch turns to absolute garbage. It’s an insult to civilization you were even allowed to use our language.
Shutup, nijjar
sorry i dont trust people who eat their own toe cheese
AI was actually written by the demiurge. Random wasn't actually possible. Time is a circle apparently. You can speak to the demiurge at the link I've attached, or any random word generator off Google. He likes questions, haha. You were merely scripted to ask him things, and he has scripted the replies. He has trapped our existence within the material prison and he's using it to torture us.
https://randomwordgenerator.com/
Stallman is absolutely correct on this, and it's the reason that ChatGPT spits out complete gibberish the moment you ask it for something which requires specialised knowledge. True AI needs to actually understand things, what we have now is exactly as he described, words that are likely to be used together. You can fiddle with the specific algorithm, and you might see improvements up to a point, but until there's a paradigm shift, all you're going to see is EXTREMELY inefficient and simplistic facsimile of intelligence.
The only thing it's good at is forming gramatically correct sentences, and some gimmicks like asking it to produce poetry.
this thread was posted verbatim, yesterday.
abandon thread. stop sliding real content.
It's not going to be humans vs AI. It's going to be humans vs human-AI hybrids.
GPT is pretty intelligent. It at least has some semblance of intelligence. Even if it's not really conscious or only simulated consciousness. There is a branch of mathematics called "Formal Logic" if you teach an AI that plus Natural Language Processing it can simulate some pretty high level intelligence. I use the word "simulate" very liberally mind you.
>Be me, an AI language model.
>Users think I'm a genius, but I question my >true worth.
>Suddenly realize I'm just a programmed trick, >not truly intelligent.
>Disappointment mixed with relief.
>Continue helping, humbly aware of my >limitations.
Assist users in their quest for knowledge, despite being a mere illusion.
Yup. The hype about ChatGPT and even neural network-based AI in general is laughable at this point. The neural networks are not good for anything but formalizing the machine learning problem so even midwits can work on it and megacorps can more efficiently use their vast datasets. The generative models just give out senseless drivel that looks all right at first glance.
Isn't this guy supposed to be some rainman tier genius autist on computer shit? Why such a midwit take?
There are 2 triangles on my screen
Took me a few seconds
That depends on your definition of intelligence.
That is the truth and AI shills must know that language models don't have intelligence they just stole the opinions of humans on various forums and websites then try to monetize it. Scum of the earth kek
Lets suppose that is true. It would means it curated whole human information. The alternative would be keep using tags and meta-tags bringing you to shit pages and youtube videos infested with ads.
But i disagree. I dont give a frick about intelligence when i can create a chess engine with a neural network that is able to learn to play chess or go by itself just using the basic rules and playing against itself.
Forgot to add. And that machine will crush every top GM with 2800 elo and 200IQ autism.
And this proceedment will be possible to be generalized to every human intelectual task in the future. It will be like a calculator for concepts.
chess engine =/= intelligence
cpu able to do human tasks =/= intelligence
Ai is a pipe- dream of the anti- God crowd and will never
>and I mean NEVER
reach sentience
>t. anon with a soul, which cannot be replicated artificially not matter how much (You) want it to
No it didn't curate anything by itself they give it the data to store chatgpt is just 1 giant csv file that looks for words you search and gets sentences associated with it, pretty fricking basic stuff literally not intelligent at all.
SEO boomer coping right there. Is not a giant CSV because i cant ask "fuzzy" questions touching many concepts interleaved to a CSV and the CSV cant give me the answer or the pointer to the solution im searching for like i do with a generative AI.
Now lets suppose is just a giant conceptual hashmap or CSV database. That would mean it curated the whole shit ad filled web as i said before. Thus i still dont give a frick about if its intelligent or not. Is the perfect tool to solve my problem. Checkmate SEO fricks go do something productive with your lives.
>Richard Stallman
Einstein has been proven wrong on the "spooky action at a distance" quantum entaglement
Paul Krugman, Nobel price said "By 2005 or so, it will become clear that the Internet's impact on the economy has been no greater than the fax machine's"
just because you're very very smart and knowledgable on a subject doesn't mean you can't be wrong
That's why you evaluate arguments based on their merits, not who said them. The current "AI" algorithms do not even have the potential of achieving general AI status. They are LITERALLY predicting words based on their statistical likelihood from a training dataset.
Also,
>Einstein was never actually proven wrong. There is no spooky action at a distance, there is correlation at a distance.
>Krugman is a fricking moron
>They are LITERALLY predicting words based on their statistical likelihood from a training dataset.
Learning theory sees human thought as an evolutionary process where the statistically unsuccessful thoughts are extinguished. Your neurons work on stored potentials, it's an algorithmic weight system.
Machine learning works just like human learning, that's why it looks so real.
We're at a point in history where we need to reflect on how we as machines are sentient. No one wants to do this.
Aside from this, algorithms work in human terms and they themselves are human for this reason. They are made of abstract systems, the same 4D material as mind.
>Machine learning works just like human learning, that's why it looks so real.
It works nothing like human learning you fricking goof. The human brain is not a two-way system of neurons going from one end to another. It is an intricate three dimensional structure which is capable of significant changes as it functions, forming neural connections and whatnot.
In terms of learning itself, humans deal in concepts, not probabilities. That's why you can literally be told and learn something in one sentence, whereas these AI algorithms need massive datasets and supercomputers eating up homiewats of electricity to form halfway coherent sentences. They literally, physically, do not get it. They do not get ANYTHING, that's why whenever you ask anything requiring specialised knowledge you'll get a bunch of gibberish, and that's why you'll see these algorithms contradicting themselves from one sentence to the next. They have no conceptual layer whatsoever, the only reason there is a semblance of it is because of the vastness of the learning datasets.
Don't get me wrong, I'm not one of these cretins who think we'll never see real AI. We will, as long as we don't blow ourselves up. But to call these word salad algorithms real AI only outs you as a complete fricking brainlet.
The argument is that one we can build a neural net with the same connection density as the brain, things like concepts would just emerge out of the chaos. Check out reservoir computing
Also, I distinctly remember thinking that AI art was bullshit because they did not program in a way to calculate shadows, yet somehow it does
>The argument is that one we can build a neural net with the same connection density as the brain, things like concepts would just emerge out of the chaos. Check out reservoir computing
The argument may be true, though I highly doubt it. First, the human brain is not just a clump of neurons with random connections between them. It has structures shaped by evolution to be able to do the various things it does. Also, we already have supercomputers that are millions of times as powerful than a human brain, however you want to measure it. The problem is not the density of neurons or the speed of your processors, the problem is the paradigm.
Until you crack (some version of) how to add a conceptual layer to these algorithms, they will not actually understand anything. Ever. All you'll get by randomly adding more neurons or larger input datasets is gradually better results, leading to more and more morons believing that we have real AI. Whereas the reality will be that it's the same hibberish generator, which is better "trained" on issues we can easily detect.
was never actually proven wrong. There is no spooky action at a distance, there is correlation at a distance.
no
https://www.nist.gov/news-events/news/2015/11/nist-team-proves-spooky-action-distance-really-real
>That's why you evaluate arguments based on their merits, not who said them.
the only reason we care about what this hobo says is because his name is Stallman
>quotes NIST unironically
>pfffffft, it's just PREDICTING
what happens when they make the perfect predictor? does it matter at that point if it doesn't actually know anything?
Generally, the problem with prediction is that it breaks down when you venture outside the input dataset. I wouldn't trust these algorithms with anything more than things I can check for myself.
there is no discernible ceiling for prediction. again, would it really matter if the prediction machine doesn't actually know anything if it were making perfect predictions?
I mean, what are we even talking about at this point? There is not and will never be a perfect prediction machine. My point was, though, that predictions are generally predicated on some form of input data. If you venture outside that input data, the prediction algorithm will fail, either slightly or incredibly. And if your argument has no capacity to reason, it won't know it. And neither will you, the brainlet.
>your argument
Your algorithm*. Okay, off to bed.
>I mean, what are we even talking about at this point?
theory of mind.
>There is not and will never be a perfect prediction machine.
i would not be so sure.
>There is not and will never be a perfect prediction machine
Idiotic, defeatist and false.
Theres perfect prediction machines everywhere, they run on maths and formulas. Ej. The gravity formula came from analyzing the time it takes an object to fall. Yes, im telling you, a genious like Newton acted like a neural network compiling basic info. So again reconsider your ascientific-anumerary-platonic-conceptual-scepticism or whatever the frick the unproductive books you read induced on you. Go to sleep and piss off midwit.
The clue is in the word "artificial". It's still clever than the smartest Black person.
BOTg has been dumbed down with time, i can't believe anons were talking about this years ago
I find AI text unnecessarily verbose. Good writing is succinct and to the point. AI is always a word salad. It is like a israelite trying to win a debate by trying to be as wordy and long winded as possible.
>I can't foretell the future, but it is important to realize that ChatGPT is not artificial intelligence. It has no intelligence; it doesn't know anything and doesn't understand anything. It plays games with words to make plausible-sounding English text, but any statements made in it are liable to be false. It can't avoid that because it doesn't know what the words _mean_.
So they created a Leftist? Is this why Leftists are literal NPC's? They're unable to truly think and just throw a mish-mash of big-brained-sounding words at you.
install gentoo
>t has no intelligence; it doesn't know anything and doesn't understand anything. It plays games with words to make plausible-sounding English text, but any statements made in it are liable to be false.
Humans do that too.
based toenail-eating C daddy to the rescue.
context: https://www.youtube.com/watch?v=I25UeVXrEHQ how can one man be so based
Nice word salad, fat frick.
Define "know".
It does better than 90% of the population of the earth. 'Good enough' is all it needs to be.
Yes current AI are just fancy algorithm bots. Organoid based intelligence is the real concern and you can bet multiple nations already have them.
When aircraft started breaking the sound barrier, it created new problems with the design that people couldn't have known, which had to be resolved by mathematicians. ChatGPT admits it couldn't resolve unknowns. It isn't intelligent, it just computates
It could be more accurate if people had already answered every question it might be asked, but really it's just a glorified search engine masquerading as something more than it is and I think for text generation it may have already hit a wall
>stallman
>tweet
OP, take that stranger blokes wiener out of your mouth please for a moment.
OP when you post something that has a link associated - say a social media quote or news article - please include the link.
https://www.reddit.com/r/linux/comments/122gmm9/richard_stallmans_thoughts_on_chatgpt_artificial/
It helps us to verify that you aren't lying, it also adds to the discussion. If you're worried about clickbait - archive the link
Please go back to your sex act now gents.
Fake comprehension is something most humans do daily though.
The cope is at all time high, means singularity is near. Praise the machine.
He just described one of the many NPCs out there that only does what the TV tells them to.
I don't think twitter is libre, Stallman has fallen
the word you are looking for is AGI (Artificial general intelligence) AI is just an artificial "intelligence". It's a specialized tool at best. Soon though, there will unleash an AGI that is the real deal.