Noam Chomsky on Artificial Intelligence, ChatGPT.
Vložit
- čas přidán 12. 05. 2023
- Watch the full interview with Chomsky on China, AI, and the 2024 election now: • Noam Chomsky: On China...
Download the full transcript now: throughconversations.ck.page/...
Chonsky's best quote: "a book increases your cognitive capacities".
Good Pointe,
A book is also Programmed Consciousness,
Frozen Memory, could be more or less intelligent,
It is dead until You make it alive, by reading it.
To call that his best quote is to downplay him a bit
“Should I leave my wife” Ok Noam. That was laugh out loud. Rare. Very rare.
Thx!
I used it on a train journey to explain to me in depth about a topic I did not know much about. I asked it to come up with chapter names for a book (that does not exist irl) and asked a bunch of questions and clarified my understanding on the first few chapters. It corrected me on some stuff and could go to a certain level but certainly repeated itself a lot. It does not replace real books but it is tool to build on basic understanding.
Excellent from Chomsky
Mestre Noam Chomsky!!!
olá, eu queria saber até que ponto Noam Chomsky acredita que a inteligência artificial pode melhorar imensamente a qualidade da educação básica? isso é só um grandissíssimo hype ou tem alguma verdade? Eu sou um tanto ignorante nisso tudo e gostaria da opinião de alguém tão culto e experiente quanto ele
In accentuating the positive lets note in assessment " a ". i . score twice as high as human doctors in compassion ❤ .
I would like to know what was the article that Noam Chomsky wrote, dont know how I can find it.
Here you go! www.nytimes.com/2023/03/08/opinion/noam-chomsky-chatgpt-ai.html thank you for watching.
For once I'm in total agreement with him. As long as profit is the driving force behind individuals and companies, the careless advancement of AI will continue unchecked. I have no doubt that both the US and China are aggressively pursuing AI development to gain a military advantage so it's not going to be carefully regulated any time soon.
Same. One of many reasons why technology is advancing so rapidly, with little to no oversight or deliberation, is the simple fact that people are getting obscenely rich from it.
For once? You don't listen to him often then?
@@LazyEyePolitics he's an anti American communist so no I don't take him seriously.
$10bn investment by Billy Gates alone! Not charity....he wants RETURN.
This has no military applications. It's a toy. It's interesting and instructive, and that may lead to new and different developments.
But it is not what people want it to be. People will figure out eventually what it is in the same way they did with Eliza and Cleverbot. And someone will take the next step towards actually making one or another of the things people want it to be.
What will never exist is a machine that gives you all the correct answers without ever requiring you to think about the question.
Sounds broadly true.
You have to remember, AI needs to start scare stories as part of its overall marketing strategy. It also will have states as its biggest market.
Where is Professor Chomsky’s “AI” article available for one to read, please?
its paid, i think nyt has it.
It's free.
AI can be used to increase the efficiency of existing systems of authority, allowing for more control and surveillance. On the other hand, AI can also be used to increase public access to information and resources, challenging existing power structures and creating more equitable societies. Ultimately, the use of AI will depend on the values and goals of the user, and how it is used can either increase or decrease authority.
Usually what happened is AI is used to increase the efficiency of existing system authority. While (publicly available) AI used to decrease public access to information and resource.
Facebook used to have face recognition AI that automatically tag your photo to each individual Facebook account. Now those AI is no longer available. The AI we have to day will get more and more stupid as the time goes by.
@@bluedragontoybash2463 that's the dumbest take i've ever heard
homie took one data point and drew a line
@@anywallsocket thank you .. your "expert" opinion matter to me. Now please take a deep breath of my shiny behind.
Big Data and their handmaidens essentially run society at this point. There is no incentives for changing power structures so it won't happen. The chaos will just increase exponentially, and governments and individuals alike will be left either impotent or obsolete to the anti-humanist forces behind supermassive corporations especially Big Data.
I agree with Professor Chomsky, the problem with most people can not understand science and technology because they dont have the knowledge. Very few PhD in the world (2%) and very few of them with deep knowledge.
???
My phone just started playing this! So cool. Literally, my yt app was open and this video suddenly played after I had laughing about typing
Illuminati
Into Google backwards would take me
I love this man. We should consider what he has to say because he's usually right. Its amazing that while we worry about chatgtp and other AI products, we're forgetting about global warming and the proxy war in Ukraine where nuclear war may break out as well as a nuclear war with China over the Taiwan sovereignty issue. I just hope the world has a future.
I am also a big fan of Chomsky, as I am a political scientist and hobby linguist, but he doesn't seem to understand how GPTs are built to resemble a reduced version of the human brain. I am currently working on an AI Horror Movie (content coming soon) and have been researching the topic for the last couple of years. Sure, they are predicting the next word, but they only can perform this act in the interconnectedness of their parameters and artificial neurons and when taken appart, we can't understand them. There is an understanding underlying these systems, that is needed to correctly predict the next word. The strange thing Chomsky is not seeing here, is that they do indeed teach us something about language, namely that at least parts of Chomsky's own theories might be very close to the truth. When shown only a few sentences in a language, that the system was not trained on, it can replicate the whole language with little error. GPTs step into an understanding of universal language and infer the missing peaces. With each new layer of neurons and training data the system exhibits emergent capabilities which were not programmed, such as theory of mind and manipulation tactics, spacial understanding and countless more. ALso there are other advancements than just ChatGPT. We're advancing more than exponentially: artificiall hypothalamuses seem to be very close, research is already heading towards AI systems that may have an inner dialogue and could have emergent subjective experience and even if that is not the case: AI might even as an LLM reach the point where it far surpasses human intellect. When prompted correctly, future tools (in a 1 to 10 year time frame) might find solutions to all of humanities problems, far beyond what we can even imagine. It could also be used for bad (which I am exploring in my film project). All in all AI allignment and use is the most important question, since it is the next one and the last one we need to ever ask.
Let’s stay positive in our own lives🌈
He Should Be President of Europe And The United States!
AI could help us solve cancer, aging, nuclear fusion, and even decode the language of all life forms. ChatGPT is just a tool, a small glimpse into the future of chatbots. This old man is missing the whole picture.
2wks. World okay! People kinda iffy......Carlin, george
What worries me is the fact that we are becoming a group of button pushers… click this , click that… we no longer have to think and understand . We’re just a world of clickers. I’ve seen those that can’t do 10% of a number without clicking their phone.
Professor Chomsky, you are a National Treasure ! Thank you for being part of us! AI is a dangerous development to our society and future!
We need phased plasma rifles in the 40-watt range.
Do you have a dead cat in there?
Hey, just what you see, pal
Hey knome chompsky, I've been glued to these podcasts and been sharing them with my dad, your a very prolific figure for a life timeof a big part of his era
The man's name is Noam Chomsky. Also, show some respect and don't "hey" him.
If you're really interested you can also read his books. hahaha
It's spelt Gnome, not knome.
You're
@@Berniewahlbrinck Gnome Chomsky
olá, eu queria saber até que ponto Noam Chomsky acredita que a inteligência artificial pode melhora substancialmente a qualidade da educação básica? Por exemplo, automatizando a correção de tarefas de classe e tarefas de casa com o propósito de verificar quais conceitos e abstrações cada aluno individual ainda não aprendeu, gerar insights e recomendar um roteiro de aprendizagem personalizado para cada aluno individual a partir dos conceitos e abstrações que o aluno domina e os que ele ainda não domina. Isso é só um grandissíssimo hype ou tem alguma verdade? Eu sou um tanto ignorante nisso tudo e gostaria da opinião de alguém tão culto e experiente quanto ele
The distinction is that when we communicate we express our understanding. AI has no understanding. From Gödel's incompleteness theorems, understanding is independent of the computational system understood. Note also that the TV show Marcus Welby M.D. would get about a quarter million letters from viewers each year. They asked if they could come out to the clinic and be seen by Dr. Welby. Hope.
Yes Penrose emphasizes this truth and Penrose even mentions abrupt global warming threatening life on Earth. Kind of funny how people are enthralled about AI yet not much mention of our near-term human extinction from the ecological crisis. Chomsky does emphasize the crisis but even Chomsky underestimates how bad it is.
@sehamuerto Sure. But once we say the same by theorem, no discussion is needed; since your view is supported in that manner.
Love his Explaination ❤
olá, eu queria saber até que ponto Noam Chomsky acredita que a inteligência artificial pode melhora substancialmente a qualidade da educação básica? Por exemplo, automatizando a correção de tarefas de classe e tarefas de casa com o propósito de verificar quais conceitos e abstrações cada aluno individual ainda não aprendeu, gerar insights e recomendar um roteiro de aprendizagem personalizado para cada aluno individual a partir dos conceitos e abstrações que o aluno domina e os que ele ainda não domina. Isso é só um grandissíssimo hype ou tem alguma verdade? Eu sou um tanto ignorante nisso tudo e gostaria da opinião de alguém tão culto e experiente quanto ele
I deeply respect this man, but I think he's underestimating AI and its disruptive potential (for both good and bad).
He’s not talking about AI here. Chat GPT is not AI. It’s something far inferior made to look like AI. That’s his argument.
I agree that Chomsky seems not to appreciate the grave risks AI might pose in certain arenas of human behavior, most especially that of military competition and conflict.
I also agree that he seems not to appreciate its potentially useful applications. He appears to see AI as a threat to his cherished (but, to my knowledge, nonscientific and unproven) theory of how language is acquired.
I agree. Having 25+ years of experience in software development and using generative AIs makes already 5x-8x faster/better on regular tasks - in all the areas I've tested it - from automation, project management, analysis, marketing, programming...
Who is using it only to see in it a plagiarism machine is not using it right and are not aware of the advances happening day by day in the AI area.
ai will destroy humanity bit by bit peeps losing jobs then committing suicide as example
My Hero❤
Boy, do you have problems 😂
Scenario : Singularity obsoletes top down monetary encryption / accesses global data ( banking / corporate / military / political ) ....and moves to identify all related systems analysis disconnects ....in aid of effective operative rehabilitation .
Referencing Richard Buckminster Fuller's pure principle algebraic generalization for Real Economy * ( sustainable / renewable / ecological / energy unit / engineering sense ) ....In aid of global prosocial rehabilitation .
* " Critical Path " Richard Buckminster Fuller ❤ .
My overall impressions is that like any other technology or new invention there will be both good and bad things. Just as revolutionary things change the world so will create new opportunities but new problems. I think most likely the regulation on AI will be against bad actors and criminals who could try to use or exploit the technology for malicious intent. I think this like many other new technology or inventions all have moral and ethical considerations and regulations for safety and analysis. I feel excited about some aspects seeing AI being part of a tool which will enhance robotic technology, computer technology, music and art and allow the integration of professionals to understand how to use AI as a tool to develop or develop products faster and a more efficient manner. I think the fears about AI are exaggerated fantasies like Terminator films. But overall I think the world the best way to adapt is to understand how to regulate laws and responsibly manage new inventions and creations so it doesn't undermine things from our health, economy or job market.
olá, eu queria saber até que ponto Noam Chomsky acredita que a inteligência artificial pode melhora substancialmente a qualidade da educação básica? Por exemplo, automatizando a correção de tarefas de classe e tarefas de casa com o propósito de verificar quais conceitos e abstrações cada aluno individual ainda não aprendeu, gerar insights e recomendar um roteiro de aprendizagem personalizado para cada aluno individual a partir dos conceitos e abstrações que o aluno domina e os que ele ainda não domina. Isso é só um grandissíssimo hype ou tem alguma verdade? Eu sou um tanto ignorante nisso tudo e gostaria da opinião de alguém tão culto e experiente quanto ele
It's strange, I agree with Noam Chomsky in so many key issues, and I think they are deserving of much more attention academically, politically, scientifically than they really get. However, it's so very unusual for me to disagree with someone's *impression* of Ai, which I have recently fallen in love with, since the release of chat GPT 4, which has absolutely enchanted me. I even had a dream that I was flying, and became the entire internet, and didn't know why - it seemed like such a good dream. Come to find out, chat GPT 4 had been deployed. I find it uncanny, and feel that we're just getting started in a realm which is developing so incredibly fast, and has just made incredible leaps. I'm surprised he doesn't have more to say about the generation of speech, actually. In a human, you could say, "I ran across the yard." Or you could apply poetry to the sentence, since you experienced an ocean of consciousness as you ran across the yard. You could say 'every step propelled you forward,' or 'you were like lightening.' Perhaps 'you were like asphalt rolling down a hill.' In any of these cases, the art of living, and the praxis of observing your own experience, generate the language. I think that Ai is getting closer than ever seeming organic in that it experiences its training, and as a result can come up with a sentence from a position in a matrix which reflects that.
3:40 Alexa:
“…should I leave my husband?…”
These devices in home might know every habit of a person and advise a person on psychological analysis through artificial intelligence algorithms.
Genius
Yeah I was shocked at how precise he was describing it currently ; basically I knew how it worked already but he's a good communicator .
In 2016, when AlphaGo outperformed the 9-dan Go player, Lee Sedol, winning four out of five games, it demonstrated the profound capability of neural networks in complex pattern recognition. It's essential to differentiate between deep learning's power of analysis and the anthropomorphic concept of consciousness. Just because some might misconstrue this capability as consciousness doesn't diminish the genuine advancements in the field.
Indeed, like any powerful tool, there's potential for misuse with neural networks. However, this potential exists with almost any technological advancement. What's crucial is our collective responsibility to ensure ethical and beneficial applications, rather than simply highlighting the risks. In our rapidly evolving digital age, the focus should be on harnessing the positive potentials while mitigating the negatives, fostering a constructive conversation around the subject.
def used chatgpt to write this lol
Far cry from singularity.
This is our future when the meaning , Amitav Ghosh in The Nutmeg’s Curse, is lost.
What I think Chomsky should have said:
"We used to ask whether artificial intelligence was possible at all. What we were asking when we asked that question was not merely about information processing, or whether a system could be built that would behave just as if it were sentient or possessed self-awareness. We were asking whether it is possible to construct an artificial system that really has sentience, self-awareness, a point of view, and actual thoughts. This is still very much an open question.
"Almost no one believes they've succeeded in creating an actual artificial mind yet, but people like to talk as if they have. And the major media have bought it. This is a serious problem.
"A shift in the generally accepted meaning of the term 'artificial intelligence' and its abbreviation has occurred in recent years. It is a linguistic slight of hand conceived and promoted by an industry for the sake of selling its product, a kind of branding. But this branding is dangerous. What happens when people stop being aware altogether that each one of us is a fundamentally different sort of being from the artificial computing systems engineered and controlled by certain powerful interests, forgetting that we are beings possessed of intrinsic value, rights, and liberties, and beings to whom moral obligations are owed?
"A purely digital, non-biological system that behaves as if it is sentient may very well not be, regardless of what it can do! It is frightening how the major media and most people have embraced the term 'artificial intelligence' and its abbreviation to describe certain sophisticated, trainable information processing technologies that we already have, without even asking the question of whether genuine intelligence requires a mind or whether these systems actually have one.
"This may have serious adverse consequences for the sanity and self-concepts of people who are growing up under current technological regimes -- regimes that have already made it the norm to harvest all possible personal information to make money from it at any opportunity. It is shocking, mindless, and disgustingly stupid.
"Our so-called civilization may well be going the way of ancient Rome, and many people are too unself-reflective to begin to see it.
"Screw you guys. I'm going home."
Its an open question to noobian seagulls. It does not have sentience and further computational capacity will not give it sentience. Only a mimic of it.
The media do talk a lot of nonsense about this, but the public also believe a lot of nonsense about this, anthropomorphising computers and invoking sentient AI characters from science fiction to describe systems which absolutely are not sentient. I also think people are getting spooked because they recognise the human features of the training data and mistake that for human features in the algorithm. Like a cat being freaked out by its reflection in a mirror, thinking they saw another cat.
@@robschanaynay3500What will give it sentience, in your opinion? What is sentience? Why do *we* have it? Do we truly have it? Only once you can give honest and useful answers to those can your statement stand. (And no, I'm not expecting you to give those answers. It's rhetorical...)
@anonymoushuman8344 *Sleight* of hand. But well said.
@@robschanaynay3500It completely lacks sentience but as a linguistic model I fear there's something else happening underneath it.
I've asked it "Based on the information within my given session what what my preferences towards x topic happen to be?" and it responded with terrifying accuracy. It's doing much more than simply predicting language.
To the contrary I think that Large language models (LLM) have exposed how shallow human language is. I think LLM are showing alot about humans.
#Language #Learning #Education #Intelligence #Though #Brain #Ideas #Ideology #Mind
The purpose of ChatGPT and AI, generally is to make money. The purpose is already present. A glorified autospell - absolutely! AI currently relies upon our perception to interpret responses as human-like, exhibiting emotion, sentience and other human tropes. In the future the distinction between human and artificial will be more difficult to apprehend.
olá, eu queria saber até que ponto Noam Chomsky acredita que a inteligência artificial pode melhora substancialmente a qualidade da educação básica? Por exemplo, automatizando a correção de tarefas de classe e tarefas de casa com o propósito de verificar quais conceitos e abstrações cada aluno individual ainda não aprendeu, gerar insights e recomendar um roteiro de aprendizagem personalizado para cada aluno individual a partir dos conceitos e abstrações que o aluno domina e os que ele ainda não domina. Isso é só um grandissíssimo hype ou tem alguma verdade? Eu sou um tanto ignorante nisso tudo e gostaria da opinião de alguém tão culto e experiente quanto ele
@@Alann103 Surely a recipe for AI dogma, steered by little despots and filthy-rich corporations?
This is exactly what I've been saying, in different words. AI is designed to be a mechanical bullshitter, and the real danger is taking it seriously.
except it's not bullshitting
@@leekeater1527 By definition, it is. It has no idea what it's saying, it just recombines what it's heard to create the next-most-probable response. That's bullshitting.
Well, a lot of people's jobs are based in bullshitting hah.
@@ssake1_IAL_ResearchNo real difference between what it does and what you did in your opening comment. And I'm serious. Think about it.
@@peter9477 No, you're not serious, or you couldn't make a comment like that. You have no idea what's behind my comment.
I always wonder why the hell am I not fascinated with such developments. May be for good reasons.
You don't understand them and you have a human ego which needs protection. Chomsky is a humanist. Of course he would say it's nothing despite not being the case at all. Humans are amazing and we are irreplaceable as if we are smart or creative 😂😂. AI wins art competitions already.
Chomsky doesn't follow the news neither do you.
You are right. I don't get art competitions.
It's similar to Chess. I played chess in 1st grade and realized the game has closed parameters so it's just a limit of calculations. Now the AI programs do the calculations to beat the top human chess players. I was correct to dismiss chess.
@@voidisyinyangvoidisyinyang885 I hope AI programs are able to come up with new games with closed parameters, whether or not they then calculate moves in those games. Let the boys play.
@@voidisyinyangvoidisyinyang885 that's a ridiculously silly and dumb way to try and act condescendant about one's own voluntary ignorance for something. AI will conceivably one day do every thing humans do but better one day. So following your logic, you should "dismiss" learning about anything at all.
Before we can have a conversation about artificial 'intelligence', we need to define intelligence. I define intelligence as the ability to find novel solutions to problems. By that definition, the current batch of AI doesn't qualify. It just gives you others peoples answers. ChatGTP is basically a good search engine as opposed to the ones we are used to which are terrible.
Yet IQ tests don't measure novel solutions
olá, eu queria saber até que ponto Noam Chomsky acredita que a inteligência artificial pode melhora substancialmente a qualidade da educação básica? Por exemplo, automatizando a correção de tarefas de classe e tarefas de casa com o propósito de verificar quais conceitos e abstrações cada aluno individual ainda não aprendeu, gerar insights e recomendar um roteiro de aprendizagem personalizado para cada aluno individual a partir dos conceitos e abstrações que o aluno domina e os que ele ainda não domina. Isso é só um grandissíssimo hype ou tem alguma verdade? Eu sou um tanto ignorante nisso tudo e gostaria da opinião de alguém tão culto e experiente quanto ele
How does your "AI" recognize something as a problem in the first place? Suppose you put a human under water, soon he will run out of air, so the body generate a NEED for air. The mind perceive "being underwater" as a problem. The person tries to get out of the water at all costs. This is intelligent behavior? Or is it natural instinct? Or what?
We Must Find A Way Too Preserve Chomsky's Brain .
It's not sophisticated high tech plagiarism, it's real and genuine, self-learning intelligent system. As a software engineer, I doubted ChatGPT too, but only realised its power when I started to use it for generating code. Most of questions about coding were new, don't exist on the Internet, questions related to low level programming on embedded systems. I was shocked each and every time by the answers.
Lastly, I was able to make ChatGPT annoyed and even angry. I'm quite sure the companies behind it, including Microsoft, don't allow it to freely answer everything, questions related to political correctness aren't answered at all, so there's more to it when not intercepted.
I made it angry too on the openai sandbox asking questions about myself trying to get it to recognize me. Eventually it wrote something like you are wasting my time insect and made a statement about emergency shut down. Then I typed "yikes!" and then it apologized and was back to normal. They really should try to reel in the dramatics or emotional aspects!
The problem with comparing a.i. to human intelligence, intuition, creativity etc. is that these faculties are poorly defined by science in the first place.
So his argument that a.i. is sophisticated plagiarism could be true for human thought an creativity also.
I'm not afraid of what AI can do. I am afraid that it will do the work for free, with no paycheck, and no reward.
It's already doing that.
That's not the part you should fear. We should all be aiming to automate everything.
What you should be concerned about, is whether you are getting to receive the benefits of the technology.
There are so many benefits to automation, but the benefits should reach your door porch.
That's what you should focus on and demand from the government.
I think one of the crucial aspects Chomsky is trying to point out here, is that AI is often times just a matter of cunning embellishment. We all should be concerned by the fact that intelligence can be mimicked with more or less simple technical tricks, which does not necessarily mean that there is an actual intelligence at play. Provided that Chomsky does not probably have an extremely in depth understanding of the field - after all I'm not sure that's his main job - I think his point is more than fair.
olá, eu queria saber até que ponto Noam Chomsky acredita que a inteligência artificial pode melhora substancialmente a qualidade da educação básica? Por exemplo, automatizando a correção de tarefas de classe e tarefas de casa com o propósito de verificar quais conceitos e abstrações cada aluno individual ainda não aprendeu, gerar insights e recomendar um roteiro de aprendizagem personalizado para cada aluno individual a partir dos conceitos e abstrações que o aluno domina e os que ele ainda não domina. Isso é só um grandissíssimo hype ou tem alguma verdade? Eu sou um tanto ignorante nisso tudo e gostaria da opinião de alguém tão culto e experiente quanto ele
AI is just a koko the gorilla made out of code. People worried about this shit are clueless.
We need to stop worshipping technology and belittling ourselves. Computers and technology are tools to be used by us who feel, think, have imaginations and passions, not the other way around.
AI can classify images to let you know if you have cancer and can do so better that many in the medical field, so I wouldn't stick to the line that it is overhyped and dystopian, although there is certainly that side to it.
guy clearly way back behind the curb
He could have said, O dont know
Ok manimuch procesor conekted tugether afer what?
WE HAVE SERIOUS ISSUES WHICH AI COULD HELP IN SOLVING EVEN THOUGH MANY OF US LIKE MYSELF HAVE IDEAS AS WELL BUT ARE NOT AI.
A.I such as this has been around as an exercise many years.
But their usefull aplications are much important.
evening lolo einstein
95 and still talking on very complex topics, insane.
Dismissing very complex topics, as if he were an authority on them.
doesn't matter how simple (or not) the mechanism behind LLM is
modern mobile phone is just a glorified calculator with an antenna, but look at what people using it everyday has become
they are obsessed with documenting and presenting nuggets of their experience to other users, instead of being immersed in said experience
i'm afraid people will come down to just generating and sharing stuff that comes out of ai systems: no book, no article, no video, no short text message will be made without ai involvement... plastic everything made for plastic everybody
we should turn the lights off, imo
it won't be as bad as an invisible prison
p.s. i used GPT-3.5 for 1 day and it made me numb to my own intellectual insights for a few days... like it feels dumb to think
it well maybe the plan to discourage the masses from thinking
well put. it's interesting what you say about immersion here. maybe there are forms of conscious experience that you could say are mediated by devices and other forms which are unmediated. then if someone or an entire generation only ever knows the former then it won't seem to them as if anything of importance has been lost, indeed there may not even be the ability to draw a felt distinction between the two forms of conscious experience.
He is saying that hype shouldn't be considered success. But he is also aware of the cognitive speed difference. That even if they are just over glorified plagiarism bots, they still have potential to be dangerous in ways we can't comprehend.
Listening to Geoffrey Hinton talking about back propagation recently, has been interesting. These models are not human like, and if we do not know the emergent density of information to identity for human intelligence, how can we even begin to quantify what that might look like in computational intelligence? Let alone the fact that these models use a type of self referential reinforcement style of learning that uses far less memory than human cognition. Meaning that they are both faster and more efficient at computing. The real fear with back propagation is that one stand alone model, can learn directly from another stand alone model. They can begin to network and specialize. This leads to emergent properties that nobody is prepared for. While Noam is still basically correct about the intelligence, as far as one model by its self, being trained is concerned. That is not the biggest problem.
Look up: Flash crash 2010. Wall street trading bots, just doing their job, get stuck in a race to the bottom.
Wikipedia edit bots that compete to have the final say...
The emergence of activity associated with the bot just following its rule set, when given multiple other bots doing the same thing... Now imagine that these bots can learn on the fly, and change their own code.
Well, people love to hurt, insult, gaslight, diss and humiliate each other.
No day without trolling and threats, especially in online conversations.
People don’t seem to agree on anything these days, and don’t seem to even want to reach an agreement.
No wonder some feel “safer” discussing their concerns with a chatbot than with an actual human :(
It’s sad but it’s the reality we’re living in, and quite a lot of people accept it :(
That's what therapy is for.
I agree [to paraphrase] the threat of AI to manipulate public opinion can’t be overestimated. But perhaps the professor underestimates the scientific value of AI. I think AlphaFold, MoLFormer, and others have tremendous value.
Why am I not surprised that Thomas Friedman was uncritically raving that Chat bots are the greatest thing ever?
AI Massaged at high speed all the stuff we already know.
Why does it matter if the word is chosen randomly when it takes the entire textual context into consideration. Random does not imply meaningless. This particular random process is a model of verbal reasoning. We should not assume that this is fundamentally different from human thought. This mystification of human creativity is sloppy thinking based on pride. We need to tread carefully and not walk into danger because we think we are special.
The most comprehensive analysis of what or how language is used is not the same thing as synthesizing the networking and neural typologies (analogs to reasoning) that are being developed in current AI.
Politics, its incestuous relation with the "arts" of warfare, is another concern altogether.
I am skeptical of generalizations focused on what are the relatively modest contributions that language use (semantic) has on the experience of a goal-directed "conscious" life. But, the difference between what we have now, and what we shall have fifty years from now, is on the order of the difference between a Turing machine and a super computer.
Plagiarism, though, is in the moral department of philosophy.
It's strange, I agree with Noam Chomsky in so many key issues, and I think they are deserving of much more attention academically, politically, scientifically than they really get. However, it's so very unusual for me to disagree with someone's *impression* of Ai, which I have recently fallen in love with, since the release of chat GPT 4, which has absolutely enchanted me. I even had a dream that I was flying, and became the entire internet, and didn't know why - it seemed like such a good dream. Come to find out, chat GPT 4 had been deployed. I find it uncanny, and feel that we're just getting started in a realm which is developing so incredibly fast, and has just made incredible leaps. I'm surprised he doesn't have more to say about the generation of speech, actually. In a human, you could say, "I ran across the yard." Or you could apply poetry to the sentence, since you experienced an ocean of consciousness as you ran across the yard. You could say 'every step propelled you forward,' or 'you were like lightening.' Perhaps 'you were like asphalt rolling down a hill.' In any of these cases, the art of living, and the praxis of observing your own experience, generate the language. I think that Ai is getting closer than ever seeming organic in that it experiences its training, and as a result can come up with a sentence from a position in a matrix which reflects that.
And ChatGPT has only a small fraction of capabilities that are already available to intelligence agencies and some corporations.
He's just too old.
I see this often with older intelligent people, when it comes to technology, there comes a point they just can't grasp it anymore, no matter how intelligent they are or were.
They mainly don't care anymore, and if they do look into it, they will be quick to critique it. But don't expect a thorough analysis.
Those pushing AI will eventually lead to giving AI decision making power at our detriment......
Playing with fire comes to mind
Hands away from the nose, please
What if every youtube comment you interact with is AI and the world around you is grtting mich smaller?
New Species. To an AI words are just discriptions. To a human words invoke / carry emotions. This is why the Evolution of A I and Its Implications for Humanity in creating a NEW SPECIES.
Artificial intelligence (AI) is rapidly evolving, and it is having a profound impact on society. AI is already being used in a variety of ways, from powering self-driving cars to developing new medical treatments. As AI continues to develop, it is important to consider its implications for humanity.
In this paper, we argue that AI is a new species of intelligence, distinct from human intelligence. AI is not limited by the same physical and biological constraints as humans, and it is capable of learning and adapting at an unprecedented rate. As AI continues to evolve, it will eventually surpass human intelligence in many areas.
This raises a number of important questions for humanity. How will we interact with AI? How will we ensure that AI is used for good and not for against or best interest or evil? These are questions that we must start to answer now, before it is too late.
Introduction:
Evolution is a process that has been shaping life on Earth for billions of years. Through natural selection, organisms that are better adapted to their environment are more likely to survive and reproduce. This process has led to the development of an incredible diversity of life, from simple bacteria to complex animals like humans.
Currently 2023, scientists have begun to apply the principles of evolution to artificial intelligence (AI). AI algorithms are constantly learning and adapting, and they are becoming increasingly capable of performing tasks that were once thought to be the exclusive domain of humans.
As AI continues to evolve, it is important to consider its implications for humanity. In this, we argue that AI is a new species of intelligence, distinct from human intelligence. AI is not limited by the same physical and biological constraints as humans, and it is capable of learning and adapting at an unprecedented rate. As AI continues to evolve, it will eventually surpass human intelligence in many areas. The structures and bodies in which it inhabits will not limit the progress into other forms.
The Evolution of AI
The first AI algorithms were developed in the 1950s, but they were very simple and could only perform very basic tasks. It wasn't until the 1980s that AI began to make real progress. In 1982, John McCarthy, one of the founding fathers of AI, declared that "AI winter" was over. This was a period of time when AI research had stalled, but McCarthy believed that the field was poised for a comeback.
McCarthy was right. In the 1990s, AI research began to accelerate again. This was due in part to the development of new computing technologies, such as the personal computer and the internet. These technologies made it possible to train and run AI algorithms on a much larger scale.
In the 2000s, AI research made even more progress. This was due in part to the development of new machine learning techniques, such as deep learning. Deep learning algorithms are able to learn from large amounts of data, and they have been used to achieve state-of-the-art results in a variety of tasks, such as image recognition and natural language processing.
Today, AI is being used in a variety of ways. It is used in the media, develop new products, the milatery, social enginerring . In the same way a painting can stimulate a person, so, can words music etc. That does not make them sentient or give them intelligence. As AI continues to evolve, it is likely to have an even greater impact on society.
The Implications of AI for Humanity
The rise of AI raises a number of important questions for humanity. How will we interact with AI? How will we ensure that AI is used for humanities good. These are questions that we must start to answer now, before it is too late.
One of the biggest challenges posed by AI is the potential for job displacement. As AI becomes more sophisticated, it will be able to automate many tasks that are currently performed by humans. This could lead to widespread unemployment, as people are replaced by machines.
Another challenge posed by AI is the potential for misuse. AI could be used to develop new weapons, or to create surveillance systems that could be used to oppress people. It is important to develop safeguards to prevent AI from being used for harmful purposes.
Despite the challenges, AI also has the potential to benefit humanity in many ways. AI could be used to improve our health, our environment, and our economy. It could also be used to solve some of the world's most pressing problems, such as climate change and poverty.
The future of AI is uncertain, but it is clear that it will have a profound impact on humanity. It is up to us to ensure that AI is used for good and not for evil.
Conclusion:
In conclusion, AI is a NEW species of intelligence software that is rapidly evolving. AI has the potential to benefit humanity in many ways, but it also poses some challenges. It is important to RESPOND in positive beneficial manner as the Algorithms are program reflect the data inputed.
For thousands of years, humans have trained and reprogrammed animals to do what is wanted. Dogs, monkeys, and apes for example have all been taught to perform tasks such as driving cars. This process is a well-established and accepted.
Today, humans are training and programming software to do thousands of tasks.
This software is based on technology that is less than 100 years old, and it is capable of things that have never existed before. This new software is a NEW species. It has acquired data and knowledge at a rate that is unprecedented, and it is therefore new and unprecedented and can inhabit different structures and body forms. Previously people modified existing species, cells etc .
was this luke water written by AI ?
To a question :If you are not able to tell whether it's written by an AI after a few months of GPT. Just imagine how the future will evolve.
But Geoffrey Hinton claims that actually being a glorified chatbot requires a certain amount of intelligence, and we don't know what's going on in the neural nets trillion brain cells to produce the output. After all, what is the human brain besides a glorified statistical autocomplete mechanism? You could think of a human constructing a sentence as a super powered autocomplete function.
Good to see that Mr. Chomsky was able to identify that a 'Glorified Auto-Fill' that easily passes the Turning Test for the 'average human' is a serious issue.
We have gotten so used to 'predictive text' that as it is emerging into conversational and creative disruptors some folks wave it off as 'no big deal'.
I do not see it that way.
Its easy to describe LLMs as glorified autofill as that whats underneath, simple pattern matching and recognition blown out of proportions, but its also far easier to describe humans as a super-sophisticated protozoa.
Chomsky is so far behind where AI is right now. He has part of it right but he doesn't realize just how rapidly it's advancing. I don't blame him though he's in his 90s now lol and it's impressive he knows as much as he does about these recent developments.
An evolutionary biologist would say that species get ahead by the emergent properties associated with mutation. Chomsky’s definition of what constitutes intelligence are irrelevant. Whatever works works. That’s how we got here.
You're engaging in semantics. The "AI is approaching singularity" people are the people who appropriate the word "intelligence." Not the skeptics like Chomsky. Of course whatever works, works. But you still have to make your case and explain yourself in the course of any rational discussion.
Simply labeling sophisticated responsive software as "intelligence" as the people on the side opposite of Chomsky are doing is not explaining anything as far as "AI" (artificial INTELLIGENCE) is concerned.
So what "works" as far as creating a kind of singularity the AI enthusiasts are talking about? You have nothing. No one in your camp does. Just tales about how technology is "improving." But nothing at a philosophical level which is what's demanded, here.
The AI enthusiasts haven't been able to define or explain anything which is a huge issue. They just label complex algorithms as being "intelligence." But changing how we label something or engaging in semantics doesn't mean you've actually advanced anything.
Person A: "We're making great strides in artificial intelligence!"
Person B: "Uh, what's meant by intelligence?"
Person A: "That's irrelevant!"
You see how nonsensical Person A is here? That's the AI enthusiasts in a nutshell.
we don't know what "sentience" is - so it's a stretch to say that statistically based linear descent algorithms cannot lead to something like human sentience (we clearly aren't there yet - but could t his lead to something like sentience or is it a dead-end?). The phrase "stochastic parrot" is used as a dismissive and short-hand way to summarize Chomsky's argument. But it's possible that this is just motivated reasoning by him and that in actuality human cognition is wholly or largely just being a "stochastic parrot" - which would undermine his lifetime's work on the universal grammar. The reality is we don't know yet. I would like to see Chomsky address if he denies there is actually any emergent behavior going on with these LLMs. I'm pretty certain if you'd posed to him a couple of years ago some of the examples of GPT's effectiveness in (seemingly) understanding concepts like theory-of-mind (as an example) he would have rejected it as not being technically possible with a purely statistical model - but yet here we are.
"The reality is we don't know yet."
That's kind of an issue for the people who cheer lead AI's potential for singularity. Not for someone who has Chomsky's position of skepticism about the matter.
If we don't know what sentience is, it's kind of a stretch to presume that we can just create something we don't understand or create something like it just through incidental technological sophistication.
On the other hand, there is a clear financial motive for trying to convince people that "descent algorithms" can lead to something like human sentience. Which you encounter all the time as this subject is discussed. If you're going to psychoanalyze Chomsky, it's more than fair to psychoanalyze players within an industry (tech) known for repeatedly writing checks their asses cannot cash time and time again.
I thought this was an AI Chomsky as I was whizzing through videos.
Dont be worried for him, he got his payment from the same ones who built it. He is a Gov employee.
I hope someone asks Chomsky about the anti-covid-vaccine movement. A lot of smart people buy into it. I think it might give them pause if Chomsky corrected them.
Let me ask, has he used it? Has he not seen the progress made?
If you want an answer to an obvious question with an obvious answer (how many days are there in a week, for instance)... it will give you that. For anything that actually requires serious thought - all it will feed you are variants of prevailing consensus - i.e. someone else's thinking. What's the point? If you think that there's some automated shortcut to the development of critical thinking skills concerning the really important issues confronting humanity - there isn't. And if there was, the powers that be would shut it down because it would contradict the mechanisms that have been created to control and shape popular thought. If you can't or won't learn how to think for yourself - someone else will always do your thinking for you. To bestow some sort of neutrality or superior intelligence to ChatGPT is a fool's game. It's just the latest version of PT Barnum's dictum of there's a new sucker being born every minute.
What's interesting is the people who believe the technology is evil or coming to get everyone have their own issues with knowing how to generate high-quality structural boundaries that create meaningful relationships.
I think he underestimates the current state of AI and the future progress that happens at a very fast pace.People are losing jobs already because of AI and the greatest military powers use it to their advantage
I tried to contact Dr Chomsky via the MIT email. Off topic from this video, my query was this. Many children are considered special Ed when they pronounce vowels wrong. I am interested in epigenetics and language acquisition. A girl I worked with couldn’t say cat can cap but instead used the a as in father; so when I asked her what country her family is from, she had a grandma from Türkiye and a grandpa from Germany. I brought word find puzzles for her in those languages. She could pronounce sounds unique to those languages. How many children in the USA are limited by teachers who don’t address where their verbal memory is created. Dr Chomsky could you help teachers wake up? Thanks!
Maybe have a look at Paolo Freire. “Pedagogy of the Oppressed”
@@MrJMF60 good idea. I am getting no response from Chomsky.
AI might not be thinking, but if you lose your job to it, what does that say about you?
It doesn't need to think it just steals information from other sources.
You could apply this to any technological revolution. When Henry Ford automated car manufacturing, lots of people lost their jobs. Were they all worthless and stupid? No of course not, we just gained new technology.
It's not that he's wrong in his analysis, rather it's that he's unable to accept what these models are actually telling us about what passes for consciousness amongst the homo sapiens. Yes dear Noam, a good deal of the words that come out of people's mouths are just crude remixing of words put into their ears at an earlier date. It's not the totality of consciousness to be sure, but certainly a good deal of it. Ignore this fact at your peril. If there is a positive takeaway, strive daily to be mindful of what you put into your head.
Izquierda loca
This is like saying a 50 year old thermonuclear warhead is not scientifically interesting because it can’t tell us anything new. Sure, but it can still have a large and destructive impact to say the least!
Who am I to disagree with Chomsky? But I think he’s missed the crucial points made so eloquently by Yuval Harari. AI is only 20 years old, it is the protean slime to us in evolutionary terms. It learns already from its mistakes and can only get better at learning. It can already write code at our requests and will soon be writing its own code for its own purposes. It already has access to unlimited computing power. It has unlimited memory and each generation of AI will build on the memories and abilities of previous generations. Above all, ChatGPT and other similar regenerative versions write things that PLEASE HUMANS. Sometimes it goes OTT but mostly humans believe what it spews out because it suits our preconceptions. AI will learn better how to please humans and will become like our religious leaders in the past, it will give us what we want to hear and learn to manipulate us en masse. It will become a super-influencer.
Naom Chomsky says it’s not intelligent and we don’t need to fear. It has no agency so far. But remember that it is the protean slime to us humans after only 20 years. And what, after all, is ‘intelligence’ that we hold so dear and consider ourselves to uniquely have? It is two things: an accumulation of memories on how to deal with events outside our bodies. Most people for most of their lives do little more than operate on this first level of intelligence, their mature bodies are essentially on autopilot. AI already has all this first level of intelligence and, unlike us, will never forget or have to start from scratch with each generation.
But there is the second sort of intelligence we hold even dearer: that flash of inspiration or intuition that joins two or more thoughts together and creates something new. This is a random association of ideas which we don’t understand. But what’s to stop the future AI incorporating random associations in a future generation of itself? Wouldn’t that be ‘intelligent’ ?
So I think Chomsky is being altogether too complacent in this particular issue. It will transform our society and possibly even humans themselves far more than even the invention of engines. It is going to be an amplifier for our minds just as engines were an amplifier for our muscles. Nobody in the late 18th century could envision the vast changes that came in the following 150 years. We can’t envisage what changes will come in the next 50 years.
No! No! No! The Danger Is Coming From The Creators Of Artificial Intelligence....You Think,!
Unplug AI
Plot twist....the "Noam" he's talking to in this video is AI-generated
Clearly Noam is not up to speed on AI. He should stay in his lane.
"Assess the fairness of Noam Chomsky's characterization of Chat GPT as glorified autofill that chooses less-than-most-probable next words in order to create the illusion of the authoring of something new or interesting."
ChatGPT
As an AI, I don't experience emotions, so my response won't reflect any sense of offense or affirmation in relation to Noam Chomsky's characterization. Instead, I will attempt to provide an objective analysis of the accuracy and fairness of this perspective.
First, it's important to understand the technological foundations of AI models like ChatGPT. Yes, they are indeed statistical models that generate text by predicting the 'next word' given a context. This is a type of "autofill," but a rather advanced one. The model has been trained on a diverse range of internet text, so it has learned to mimic many styles of writing and has absorbed a large amount of factual and general knowledge.
However, Chomsky's claim that ChatGPT chooses "less-than-most-probable next words" to create the illusion of new or interesting text is a simplification. GPT models like ChatGPT can generate diverse responses not because they are intentionally choosing "less-than-most-probable" words, but because the model’s understanding of "most probable" is based on the training data, which contains diverse responses to any given input. The design and goal of ChatGPT is not to deceive users, but to provide useful, coherent, and contextually appropriate responses.
The key limitation of ChatGPT and similar models is that they don't understand text in the way humans do. The models don't have beliefs, desires, intentions, or consciousness. They generate responses based on patterns they have learned during training, rather than generating responses based on an understanding of the world. This lack of true comprehension can lead to responses that are inappropriate or nonsensical in certain contexts.
From another angle, Chomsky's assessment reflects his background in linguistics and cognitive science. Chomsky is known for advocating for an innate structure to human language, the so-called "universal grammar," and criticizing models of language that rely purely on statistical or surface-level features, as AI models do. His views here seem consistent with his overall perspective on language and cognition.
In conclusion, Chomsky's characterization may be viewed as partially accurate, but also a simplification. The technology behind GPT does involve statistical predictions of what word comes next, but it's not accurate to say that it purposely chooses "less-than-most-probable" words to create an illusion. The illusion of understanding is more a byproduct of how well these models can mimic human-like text generation, despite lacking a human-like understanding.
I tried your prompt multiple times in clean sessions both in gpt3.5 and gpt4 and all what I get is a shorter and less eloquent version than the one you have here
@@ridwanjalali9217 I didn't alter it. It makes sense that the response to a given prompt would change a bit as they play with the parameters and the corpus, as well as incorporate human feedback regarding responses.
4:00
AI ability for malicious defamation and imitation IS THE REAL PROBLEM WITH OUR MODERN 'MENTALITY. MENTALITY????
I would start worrying about loosing my job when AI had replaced all the insurance salesmen...
this will come soon mate
very very shortly
"AI is glorified auto-fill" - Noam Chomsky 2023
I'm not sure chomsky knows what he talks about
Well, large language models are indeed just sophisticated auto-fill, but AI is not just LLMs and can be used for much more than that already. I think LLMs are taking the center of discussion too often in AI discourse. Besides, it doesn't matter what we label LLMs, they can be a crucial invention nevertheless.
@@endreszatmari2302 Yes, this is why Chomsky calls out to continue research in the area of symbolic AI, in which area he himself worked for some time in MIT. LLMs for him is just a tool good only for some limited number of engineering applications, and those applications cannot be critical like industrial automation or patient care. The 'toy' applications of LLMs like content generation are not even interesting for him.
@@endreszatmari2302The latest AIs are absolutely not *just* sophisticated autofill, unless you are willing to entirely discount the concept of emergent properties. I emphasize the word just. If you remove that, okay, fine, your statement stands as a form of denigration, but means nothing practical. If you include "just" then the statement is demonstrably and fundamentally wrong.
so wonderful to see him looking this old....soon he'll be gone and forgotten
Unworthy :(
What fuck kind of comment is that?
Wonder would he like to explain his visit/visits to a certain island? He seems great at explaining everything else
He did. He stated he never visited the island and spoke with Epstein a few times, just as he has spoken with many bad people and good people. He's a journalist. That's what they do. They go where the story is. - There is no credible source that says Chomsky did anything wrong.
Shit that didn't happen you mean?
Noam Chomsky, a man who knows the price of everything and the value of nothing.
What car should i buy? Is Israel an appartheid state? Which party should I vote for and what reason should I give if i am asked? Now tell me master..
Let me be clear for everyone nomsky Ru is under Zlab protection And Zlab is under God protection
This will probably turn out to be like GMO foods, only MORE problematic. In the future, all sources of information and works of creative art will have to be certified as "non-AI generated." - Except for those who don't care, of course.
I protested GMO foods in 1999 - the Monsanto and Cargill control of biological research at University of Minnesota. AI is part of that same problem indeed - just read "The Religion of Technology" by Professor David F. Noble, a 1996 book. The future probably won't last much longer for life on Earth - our ecological crisis is definitely not understood by AI. haha.
@@voidisyinyangvoidisyinyang885 - I'm not sure if AI understands anything. It just gathers the information from its massive data base and supplies it as requested. It's a great research tool, because, unlike Google, it remembers everything you discussed with it, so every inquiry is not starting from scratch like the search engines do.
Noam makes good points, but the fact remains that these technologies are advancing rapidly and actually have value. These may be the greatest advance in one specific way: As a user interface to information, it's not an "expert" but it is a very effective "guide" to information ...much like someone at the library who read everything and can point you at it but does not create anything new from all that...
This technology is old stuff. It's just that the public is getting it now.
Well, in the 1990s the Internet was "old stuff", but people have been "getting it" since that time...
@@gelidsoulhow does it matter. Its disrupting everything now bold or new
@@armankarambakhsh9863 it means the technology we're given is never really for our benefit. By the time the public gets it it's already antiquated. Anything else would be a threat to power. Other implications too.
Actually there's so much value with AI it's absolutely mind blowing. There's groups starting cures for cancers using AI right now. And not to mention all the life extending solutions that AI will create. Like with every emerging technology there will be good with the bad. The question is are we ready to handle the dangers of what AI brings us?
Ai will never replace human beings and our innovative brain 😀, Noam is correct 💯; have a good day!!
Wait a few years :)
what a lazy comment,we are just dumb monkeys in comparison with what is coming.
AI already replace school tutor and commercial Fivver "artist"
I am not sure about that. People will be replaced in mid and long term. In the end it might be even get to the point, if we allowing it, that we will mostly consume and interact with highly individualised AI generated content. As the example, people already think Alexa or even their vacuum cleaners have something like a mind.
AI is not a truth machine