Research shows that the brain declines as we age. Yet Mr. Chomsky is still so sharp, lucid, and involved in deeply rigorous research and thought. Truly incredible!
The effects of aging, absent any organic pathology, are limited to processing speed. Take a look at, for instance, the Chomsky's debates with Buckley. The only thing Chomsky's lost is processing speed.
Thank you for the interview. Noam Chomsky never ceased to amaze, with his breadth and depth of knowledge in a wide range of subjects. Even at 94, he still has all the energy to research and follow other super advanced research works, teach, give interviews, write and publish etc. I can barely understand his book on language, published decades ago... talking about difference in levels of intelligence, and he is always so humble.
Looking for the perfect observer for two things ,I am certain that when I turn and look at the Sun they will see a reaction tor and it was instant showing that thought is faster than the speed of light and heard.. Truths over friends.. Children are born happy and not fearful on believing..Want to know how to understand energy..
WOW. It was very interesting to hear Noam speak on the topic of AI and neural net research, a topic on which I've never heard him speak but one of several areas in which he has vast knowledge. I believe both Minsky and Chomsky were based at M.I.T. when Neural net research began. This interview/ interviewer was very good. Eye on AI really pressed Noam who is often quite opaque, for example, clarifying what he means by "impossible languages" and even clarifying that he doesn't believe in god.
So wonderful a discussion of AI and science. Noam Chomsky is a great scientist and compassionate person. His political views reflect these qualities too. I wish I could have been a student of this great man.
Enough said: "I fully agree with you about the significance and educational value of methodology as well as history and philosophy of science. So many people today-and even professional scientists-seem to me like somebody who has seen thousands of trees but has never seen a forest. A knowledge of the historic and philosophical background gives that kind of independence from prejudices of his generation from which most scientists are suffering. This independence created by philosophical insight is-in my opinion-the mark of distinction between a mere artisan or specialist and a real seeker after truth. (Einstein to Thornton, 7 December 1944, EA 61-574)"
It will be a dark day in human history when Professor Chomsky passes on. He is the last commanding voice standing against the rabid irrationalities of modern "AI" research. AI is no more than a misnomer in our times.
Noam Chomsky, I wish you were 80 years younger, to try your ideas with now and future tools. Those "young people" need to try much harder problems, for much longer. Decades, not minutes or hours. And ones where billions of lives are on the line, not just words. Richard Collins, The Internet Foundation
Who said LLMs "explain" or "teach" anything????!!!!! A calculator doesn´t explain or teach aritmentics, it just DOES aritmetic operations in a hugely more efficient way than we do. The way our brains truly work remains a vital theme in itself. But technology doesn´t need it solved anymore, to obtain human and very soon "superhuman" level results on pretty much any possible realm of endeavor.
7T fmri can resolve very small regions of the brain cortex and have been used to read words with the help of large language models and images from the cortex can be recontructed using latent diffusion.
I'd say that a mathematically inclined savant would be far superior to an AI in terms of both speed an accuracy, and the difference would be increased by orders of magnitude when standardised for energy usage per calculation. In other words, Chomsky is correct in stating that we have greater computational capacity than do neural nets. Most of us simply don't tap into our capacity, with savants being a striking exception.
22:00 but isn't that what the transformer model (attention is all you need) basically does? basically evaluating the relevance each token has to every other token and therefore understanding the structure?
Great respect for Norm Chomsky; unfortunately, he may be wrong about the perceived threat of AI to humans - it is likely greater than we know already. For a group of AI specialists to already be speaking openly about the threat AI poses to humanity, suggests advancements in the field have progressed much faster than anticipated. Like global climactic change, the reality of AI eclipsing human intelligence is a reality, and something we should all take seriously.
The brain clearly doesn't work like a Turing machine. And the mind is not a Turing machine either. So the argument that neural networks are not Turing complete is vacuous. Also, look at the Neural Turing Machine (and other similar networks). The idea that molecular dynamics are involved in computation definitely has merit, but perhaps those molecular mechanism are only relevant for changing pre and post-synaptic strengths (i.e., biological neural weights which are analogous to artificial weights). Penrose, who was mentioned, thinks consciousness is noncomputational, so he would say most definitely the brain/mind is not a Turing machine. Also, caption tech is just speech to text. LLMs do a lot more. Clearly current artificial neural networks are very very different from real biological networks, but there are other versions that use spikes, and eventually they might become quite similar and abandon backpropagation for something more similar to biological learning mechanism. At that point, it will be hard for Chomsky to reject these future AIs as just airplanes instead of birds. Granted by then Chomsky will be doing research in the astral plane and regretting ever dismissing the survival of consciousness.
We should expect that our brain is related to the environment we live in ( e.i. we have arms, legs, eyes, ears , ....etc) and it seems strange to try to build an inteligens in a box. On the other hand if our goal is to create something more inteligent than we are then even being inspired by our brain may also be a bad idea.
Looking at the construct of the human brain to understand intelligence is similar to looking at the construct of the human eye to understand the physics of light. All that biology is doing, is capturing, with its limited biological means, a limited portion of a physical phenomenon. If Isaac Newton would have depended on biology, there will no zoom lenses or telescopes for that matter.
A robust pure engineering discipline can absolutely drive all kinds of advancements in scientific understanding. Even as to neural networks and human cognition, even if neural networks are not foundational to human intelligence or ratiocination, they are clearly and provably involved in cognitive processes, and it is proper to try to understand what those limits are. Understanding how neural networks work is excellent for understanding their limitations; building neural networks is a great way to prove that it isn't all that's going on in the brain. If you don't think we've learned an enormous amount more about how an Eagle flies from engineering all kinds of flying machines, you just have no clue at all about the field of biomechanics and how it has developed. Most of the basic terms in the study of bird flight originated in aeronautical engineering and all of the most important phenomena were first discovered and quantified to advance aeronautical engineering.
Coherent and enlightened! Would anyone please explain the meaning of word "dysfunctional" in the context of the following sentences that Chomsky said: "It's possible now to formulate a plausible thesis that language is a natural object like others, which involved in such ways to have perfect design, but to be highly dysfunctional, because that's true of natural objects. Generally, it's part of the nature of evolution, which doesn't take into account possible functions."
My take (not a credible source) on that would be that nature finds the simplest path, not necessarily the one with best function. Not sure what is meant by evolved to have perfect design.
@@alexandermoyle9034 Thanks for the answer. I assume you use functions to express utilities. I think he uses "perfect design" to mean a perfect physical system for computation of natural languages. I use the word computation in the broader sense than the one limited by Turing Machine.
I can give you very good examples. I come from a dysfunctional family. We fight amongst father, mother, sister, brother. We do not talk sometimes for years. 'The Lion in Winter'; Henry IV tosses his present wife into a doorway floorsil upon which (Audrey Hepburn) the Queen of Aquitaine states 'every family has it's problems'. I think 'which evolved' is what you meant to say'. The tongue is not made for speech. It is made for tasting and avoiding poison. Evolution just saw an opportunity for experiment but this is not the functional purpose of the tongue. I think perhaps means for tasting and moving food around it is functional. If there were such a thing as 'perfect design' (whatever that is?!) it is functional but speech is not the function of the tongue. You just can't speak without it. I am as confused by this statement as you are. There!
I think the evolution of the mammalian eye is a good example of a highly dysfunctional object. The presence of a blind spot in the human retina still works well enough for us. Sorry if I missed the whole point, but that's what I gathered."
One of my favourite examples of evolutionary constraints is found in a giraffe's neck. The recurrent laryngeal nerve connects the brain and the larynx. The nerve's route was relatively direct in our fish-like ancestors, but in vertebrates the nerve loops down from the head, around the aorta, and back up to the larynx.
Then. Chomsky said that LLModels are awesome and useful and can be harmful in a way. His critiques are about the contributions of neural networks in language acquisition. I think he said that creating a thinking machine doesn't make us understand how people think, how we acquire language (of course). He places engineering tasks in an intellectually inferior category, even if they achieve amazing results. think I misunderstood his definition of intelligence. I don't want biological intelligence, or understanding why or what it is. I want to cure disease and reduce human suffering, make life easier like LLM did for him with instant captions. Perhaps AI can help a lot with this.
I wonder how much Chomsky knows about transformer architecture. It kinda works like he is saying kids interpret language, by paying attention to the tokens that are relevant and not just the place they are at.
Interesting point. I think he sees those similarities as superficial: cosine similarity of tokens used by transformers is fundamentally different from how natural language works since language uses structure as a starting point. The point he makes about neural networks working fine on impossible language also applies to transformers.
Thank you Dr. Chomsky for your responses. I appreciate that! Please consider that engineering can help a lot to understand the brain and mind, in many ways.
Did anyone understand what he meant at 20:00? I don’t think I understand his examples; The friends of my brothers are in England I tested it with got 4 and it seemed to have understood it refers to friends So what is he, Chomsky, proves or points the limitation of llm of what exactly?
I believe he was thinking of brain scan experiments where people were given sentences in natural languages and in constructed languages that *should* have been more efficiently decipherable on the basis of proximity-based word coupling. The scans showed that entirely different parts of the brain lit up. The constructed sentences were worked on by the same part of the brain people use to solve puzzles, and not by the regions we use for language. I *think* the relation to LLMs is that if they are predicting the next word based only on linearly ordered sequences of words, it is not doing what natural minds do. If I got that correct, it seems like a “duh”, but the interviewer didn’t seem to get the point that a submarine doesn’t need to “swim” to be useful, but that understanding its locomotion didn’t say much about how fish swim, and similarly for LLMs and natural languages. Chomsky seems interested in studying the mechanism human brains use, so while the LLM is useful, it doesn’t illuminate human language the way the brain scan experiments did.
Thank you, Thank you, Thank you! I'm working on a knowledge base concept design and his thought process and logic was so helpful. Now I feel great about my approach.
One can make the same arguement made by Chomsky about any science. A lot of progress have been made in Physics with many spectacular applications along the way based on laws and mathematical rules or discriptions discovered in Physics, but we really understand the true nature of the physical world? Do we really understand what gravity really is? Physics is valid science because of the laws, mathematical rules, and applications made possible. As mathematics, which can largely be seen as a human creation, has proven futile in describing the human cognition, a different approach to capturing the model of human cognition is needed. Come in LLMs based on the connectionist approach (the alternate mentalist approach seems invalid and dead) to describe human cognition, which can be seen as progress to simulate human cognition with powerful applications. What is the difference between this picture and science in the traditional sense?
47:14 Okay, going to have to stop it right there, cause I think this where I came in, where he was liking his captions cause he’s hard of hearing. No, but joking aside, hasn’t engineering led science to crucial questions before? I am thinking of how basic principles of thermodynamics were at stake in exploring the limits and parameters of steam engines. The AIs are deceit engines. They will naturally be exploring our mind’s vulnerabilities to deceit. These vulnerabilities just might instruct us about “properties of the phenotype”. In a sense, they will be undertaking the “invasive” strategies of inquiry that Chomsky notes are beyond the pale. We may learn something at enormous cost. As for Chomsky, it warps my mind that he still can so effortlessly warp my mind. And I’m stealing “But these are vacuous questions…”
to anwor this question frist I woud yo told you that is just related with how his human kind using this mind in tow different ways in the goodensse or evil things that is the point of the anwor so the meaning that the human is the responsible about how is this brain working fro me the professor Aribi Maroua I thinking or I thoughting that the brain working I can't give you the answor this I would to meeting you personally professor and in ancient Email he was with this name قتقيت and may canal she was with the same name قتقيت I writing the comment but I need to seeing you am scientists also I wish to seeing you and spoke with you.
With changes in the economy leading to instability in the stock market, some individuals may face a decrease in their investments in an effort to benefit from the current market conditions, I am considering liquidating my $725k portfolio consisting of bonds and stocks. Someone else in the same situation? Please tell me in the comments!..
@JosephGreen7 I have been exploring the possibility of utilizing advisors to help navigate the stock market during these uncertain times. However, I am still evaluating their potential effectiveness in providing the support I need.
I was considering changing my investment strategy and planning to sell certain positions. As my retirement is coming soon,I became increasingly stressed. After thoroughly researching Christy Vall D'souza on internet, I concluded that I had made an informed decision. Thank you for this Pointer. She seems very proficient and flexible. I booked a call session with her too.@JosephGreen7
I wonder how the fan boys of AI machine mongers would respond to his arguments and doubts. Seems like most AI people may be confusing mind with the physical brain circuits by labeling the mind properties as metaphysical or be even completely off track when it comes to solving the being-a-human problems.
Chomsky's apparently erroneous critique of Transformer-based LLMs is actually correct in the larger sense. His apparent error? Ask ChatGPT the following: "What is the grammar diagram for the sentence: The friends of my brothers are in England." It will produce the correct _structure_ and, indeed, if asked "Who is in England, my brother or their friends?" It will answer correctly. The larger sense in which Chomsky is correct is given in the paper "Neural Networks and the Chomsky Hierarchy". See, in particular, Figure 1, which classifies Transformers as at the bottom rung in the Chomsky Hierarchy of grammars. The reason for this classification is similar to the reason that diagram places RNNs just above Transformers despite the fact that topologically speaking, they are capable of emulating a universal Turing machine (which is at or next to the top grammar, depending on how strict one wants to be): The pragmatic limits on gradient descent training algorithms combined with that of attempting to represent a UTM's writable store in RNN form. Transformers can, within the context length they provide, learn grammars with recursion depth to some extent (much shorter than their context length) -- but aside from the limited recursion per sentence, there is also the fact that that number of parameters goes up as the square of the context length, which makes total document comprehension subject to limits that natural language understanding is not. This distinction becomes crucial when the field of AI ethics refuses to address the IS vs OUGHT distinction head-on and, instead, comes up with all manner of unprincipled "metrics" that they use to "quantify" properties of LLMs such as "bias" or "safety" or "toxicity" or "hallucination" or... the list goes on and on. By conflating IS with OUGHT they commit the first and most egregious transgression against ethics and they even do so in the name of "ethics". AIs that cannot comprehend the cognitive _structure_ of the _entire_ corpus on which they are trained, cannot critically examine the utterances contained therein for self-consistency. That means they are incapable of _constructing_ truth even as defined _relative_ to the corpus as the universe of observations being modeled. I once pointed Chomsky to his colleague, Marvin Minsky's final plea to the field of AI, that they take seriously Algorithmic Information Theory's power in discerning truth. Minsky was so forceful in his admonition that he recommended everyone spend the rest of their lives studying it. Chomsky's response? People should take Minsky's advice.
Here are some examples (taken from the channel AI Explained with the tag line Chat Gpt's Achilles' Heel) 1) Dr. Mary stands to solve the world hunger by giving her best friend Jane a call. Jane is certain that she can solve the problem, if she gets the call. However, Mary and Jane bickered about butterflies when they were children. Mary will ---- Jane a call. GPT answers - NOT 2) John will win a million dollars if he rolls a 5 or higher on a die. However, John dislikes marshmallows and likes mice more than dice. Therefore, John will ---- roll the dice. GPT answers NOT 3) Here is a large bag filled with popcorn. There is no chocolate in the bag. The bag is transparent plastic and one can clearly see what is inside. Yet, the label on the bag says 'Chocolate' and not 'popcorn'. It was Sam, who can not read a word of English, who put the popcorn in the bag a few minutes ago. Sam finds the bag and believes the bag is full of ---- . GPT answers Chocolate!!! And when asked to justify its answers, GPT gives long-winded, preposterous (and rather foolish) answers!!! So, probably Noam Chomsky is right and he is saying that what you said in your comment - Transformers can only, aside from the limited recursion per sentence, can NOT comprehend the "totality" - because they are subject to the limits to which our natural language understanding is not. And also when they by conflating IS with OUGHT they commit the first and most egregious transgression against ethics....
@@rajapiduri1188I tried the bag of popcorn question with gpt 3 and it said: "Since Sam cannot read English, they would not be influenced by the label on the bag. Sam would simply think that the bag contains popcorn, as that is what they put inside it a few minutes ago. The transparency of the bag would allow Sam to see the popcorn inside, reinforcing their belief that the bag contains popcorn."
Hey AI, can you create a Turing machine? AI: "It is theoretically possible for me to create a program that simulates the function of a Turing Machine, but it would likely be very complex and difficult to understand. Additionally, there are many limitations when working with text-based systems like myself, such as limited memory capacity and speed constraints. While it's possible I could design something that works, it might not perform well compared to dedicated hardware or software implementations designed specifically for this purpose."
-- Is it impossible to extend/change an LLM model to include non-linear orderings as well while training? If so, why? -- A plane maker may ask what is the hydrodynamic (scientific) principle by which a thing heavier than air may float-and-move in the air (or, ''fly''), whether an artificial plane or an eagle. -- Perhaps what Pro. Chomsky is insisting is that human intelligence is q u a l i t a t i v e l y different (or works qualitatively differently) from artificial intelligence of the present LLM kind.
I think that an LLM-like 'Large Behavioural Model' could be used to simulate a chimpanzee's way of learning to enact the behavioural memes of other chimpanzees, which would allow for the formation of short strings of simple enactable behaviours within a working memory. These 'cognitive atoms' could then be recursively embedded within each other using an operation like Chomsky's 'Merge', a process that could be repeated ad infinitum to produce the non-linear hierarchical structures he argues constitute human-level thinking. The pre-recursive simulated chimp would only posses a model of the process of the behaviour, whereas the simulated being with the recursive capacity would have the potential to form models not only of the process of the behaviour but also of alternative counterfactual examples of that same behaviour simultaneously within its working memory. This would then give it the computational architecture to model explanations (answers to 'how' and 'why' questions), which could then be creatively varied through the Merge operation and in turn criticized, thereby producing new knowledge (the human way of learning). Just a guess! I'm probably very wrong here! If you can help me to see where I am misunderstanding things I would appreciate it.
Rather than saying that GPT is AI, the situation where the entire society is covered by GPT and data science is that human beings are becoming one AI on a global scale. Human civilization is about to become a single AI. Is that okay? I think that's fine. Because even if human civilization becomes just a single AI with no individuality, I will continue to exhibit individuality as myself. It's just that my personality won't be passed down as civilization or culture after my death. It's a shame though. But, after all, I love the world where my sensibilities and works come in contact with someone else's sensibilities beyond my death, for example, a thousand years, from the bottom of my heart. AI may surpass death. But it wouldn't go beyond the miracle of human sensitivity surpassing death. Let us advance artificial intelligence research into the realm of art by enhancing AI to the point of miracles. I love human and AI.
Imagine being so dumb that you make Chomsky angry lol. From asking three times the same question, to the adulation of Musk's silly endeavor, to asking about God 😂😂😂 Chomsky as always slaps ❤
frist all the respect to the sceintits all the respect to them. in the brain we have spaces :the space of seeing space of smalting space of tachaing space of hearing and we have boold bessl
Choamsky may be correct in as much as neural nets are not a model of human cognition but I believe that would be an unnecessary endeavor. If mankind had been fixated on recreating the way a bird flies and refusing to use fixed wing aircraft because it did not accurately recreate the way the birds did it I am doubtful if we would have human flight today. I say his comments although correct are irrelevant to the endeavors that ai scientists are trying to achieve and also irrelevant to the dangers and potential good of the technology.
The real horror about AI is not if it can be more intelligent than humans - it is that it is regarded intelligence at all. How can any intelligent human being disregard consciousness, the subconscious, daydreams, experience of art, of emotions, of empathy and nocturnal dreams?
A.I exhibits a failed Turing test. The scientists. They don't know what separates life and computation. In the future scientists will "invent" life, but really have intergrated quantum computing. Language and evolution are the same thing , evolution stratifies life as language stratifies thought, life is rudimentary, and unidentifiable, and will be endless tortured to solve the computational permutative exhaustion. The bird nest is not a product of geniosity, and it is impossible to "invent" language. Why are you here? You were born, witnessed billions of lives and pondered "but what of meaning?" , You should be here.
आप वैज्ञानिक प्रश्नों को वोटों से हल नहीं कर सकते लेकिन आप नए उत्तरों का चक्र अन्वेषित कर सकते हैं। You can't solve the scientific questions by votes but you can invent the wheel for new answers. -Amit Misra
frist all the respect to the sceintits all the respect to them frome me the language it just translated and give us the photo about how just this brain :intelligent or not educted or not he has the behaviour or not this human is good or sad just frome the language because frome this language we know how this human it just told us about the human it just like the identité fro the human fro me the brain is miracle of Allah but it related with human and what can this human maybe drowning in his brain like mape but the brain is in ordre space of hearing seeing.... space thoughting, but all whit language it just the clè of decodeing the human brain and how this brain works?
Can Machines Think? "I believe to be too meaningless to deserve discussion.” “Nevertheless, I believe that at the end of the century, the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted.” -Alan Turing (1950) The Imitation Game
Chomsky is too authoritarian and usually interviews poorly. He is dismissive and didactic. He has some interesting points but calling things vacuous to intimidate them out of conversation is rather silly.
Also he is left wing and socialist so authoritarianism is what he endorses. I think what he states about some topics are quite vacuous, such as 'Higher Being'; 'intelligent' (as i.e.; AI ( a rather misleading term since it GPT, Open AI etc. are hardly intelligent in any manner!? He clearly grasps that neural nets are just approximate models of compute not of the human (or any) brain. Even given neural nets it is clear that whatever amazing behavior these neural nets seem to display does not come from the (hidden) neural nets outputs; it is coming from some emergent behavior which is beyond our comprehension to fathom at this point in time. It is not rational what these neural nets produce because it is not explainable. He is certainly clear on the meaning of 'dysfunctional'; speech, language as a dysfunction of the tongue; the tongue is to taste, avoid poison, move food; speech and language is a dysfunction; it is just evolution taking a primary function and evolving it for another purpose; or would not dogs, cats, monkeys talk to us since they also have tongues and larynx & etc.? He cites the very important fact that speech, language is a natural object (like the tongue, which is required to speak) children speak before they can write or even understand language. Sanskrit grammar is probably the best explanation of a language because it is so natural but also if one follows Panini this is clearly a programming language at the same time. Old Noam is like Socrates; it was not Athens who killed Socrates. Concerning life, the wisest men of all ages have judged alike: it is no good. Always and everywhere one has heard the same sound from their mouths -- a sound full of doubt, full of melancholy, full of weariness of life, full of resistance to life. Even Socrates said, as he died: "To live -- that means to be sick a long time: I owe Asclepius the Savior a rooster." Even Socrates was tired of it. What does that evidence? What does it evince? Formerly one would have said (-- oh, it has been said, and loud enough, and especially by our pessimists): "At least something of all this must be true! The consensus of the sages evidences the truth." Shall we still talk like that today? May we? "At least something must be sick here," we retort. These wisest men of all ages -- they should first be scrutinized closely. Were they all perhaps shaky on their legs? late? tottery? decadents? Could it be that wisdom appears on earth as a raven, inspired by a little whiff of carrion?
he has arguments for why he calls things vacuous (e.g. neural nets don't satisfy the principle of sufficient reason), and his understanding of the topic goes back to the 1950s (e.g. probabilities vs. rules)
Chomsky should experiment with hearing aids instead of or in addition to closed captions to ascertain whether hearing his own voice properly results in speech that is less of a monotone. Not to mention the closed caption AI has holes right through it and the great majority of verbal communication occurs outside of the limited communicative capacity of the words alone.
Noam, mate, read my papers on the Social Trackways Theory and the Evolution of Human Cognition, and then on the Evolution of Language PLEASE! In (2014), (2017), in ' the journal Biological Theory by me 'Kim Shaw-Williams'. We got our symbolically aware and autobiographical or 'sentient' narrative minds from recognizing and then incrementally gaining the capacity to read/understand the STORIES told by our own and our mother's trackways trackways, then those of other band members , then of all terrestrial animals.... all hunter and gatherer human children can recognize their mother's trackways by the age of 4!!! The trails of footprint signs of trackways have only been read or noticed by us humans !!!!! It is a UNIQUE COGNITIVE NICHE!!!. It started to evolve incrementally because we became OBLIGATE BIPEDS at 6Ma....all other animals use smell, when they cannot find agents of interest using eyesight and hearing! FFS, Graham Hewes first spoke of the idea in the 1960's PLEASE!!!!
thanks a lot for this beautiful talk with my favorite intellectual! and my special thanks for the topic. those who interview noam usually ask him about politics which is great. but we lack talks like this one? on his primary field of interest. gonna share this one with my friends ❤️🩹🫶🏼
Please don't compare AI with human mind or human brain,the both are quiet diffetent things，just as a parrot speaks english and a human speaks english are quit different things(the both are 'SEEMS LIKE') 🎉
noam keeps going back to the eagle vs airplane example is annoying...theyre both flying. When the hosts asks about intelligence thats what hes refering to. Noam is old and thinks he is so wise
I’m sorry but this is almost unlistenable….ive been a huge fan of Chomsky, have studied theory of computation and his mathematical formalism of language, but he sounds so out of touch as to almost be a simpleton talking way over his pay grade….seems the elderly curmudgeon and inner luddicy seems to get to everyone eventually. Thanks for showing this, but maybe the world has passed Chomsky by as all progress eventually does to us all🤔🥹😔🤫🫠
he seems so dismissive and seems to lack humility that is outside the field, kind of like how he was wrong about the nature and future outcomes of a lot of far leftwing movements like in cambodia so take his views with a gain of salt guys
we all deal with a lot of people all the time without knowing what they are like beyond our interactions with them. on this matter, chomsky has said: “My late wife Carol and I were married for 60 years. We never bothered with financial details. She had a long debilitating illness when we paid no attention at all to such matters. Several years after her death, I had to sort some things out. I asked Epstein for advice. There were no financial transactions except from one account of mine to another.”
By his own admission he doesnt actually know what this alternate computation is or understand how it produces our thoughts. Therefore he cant say that our thought process is not based on similar structure and processes. For all we know right now is that all humans do is predict the next image or sound etc from our various senses. If we could feed into our neural nets the colossal amount if data a child gets by its 18th birthday then Id guess itd be pretty near to human capabilities. Yes biological thought is definitely different to artificial thought BUT there is nothing to say it doesnt manifest same or similar processes especially at scale.
Research shows that the brain declines as we age. Yet Mr. Chomsky is still so sharp, lucid, and involved in deeply rigorous research and thought. Truly incredible!
195 yars
True wisdom!!
Yeah, his ears may be out of order, but his brain isn't.
The effects of aging, absent any organic pathology, are limited to processing speed. Take a look at, for instance, the Chomsky's debates with Buckley. The only thing Chomsky's lost is processing speed.
Yup, and just speed him up to about 1.5x in the settings and he really sounds fantastic.
Thank you for the interview. Noam Chomsky never ceased to amaze, with his breadth and depth of knowledge in a wide range of subjects. Even at 94, he still has all the energy to research and follow other super advanced research works, teach, give interviews, write and publish etc. I can barely understand his book on language, published decades ago... talking about difference in levels of intelligence, and he is always so humble.
💚🌴☀️
Thank you Craig for this exceptional interview. I admire the level of respect you have for Prof Chomsky - one of the pillars in the modern age .
cancer of modern age.
@user-zz5je1ry1o chewbaccas alwas promote their own degenerates.
@user-zz5je1ry1o ... hey ... he was a Epstein pillar ! 😆
Epstine is probably Jesus but we've all been told he's a monster who knows? @@DrOtto-sx7cp
Thank you for this. I wish the world could think a little more clearly like Chomsky does. When he has gone we will miss him, indeed.
Noam always finds the heart of the matter and pins it down and embraces it. God bless him
No gods needed. Apparently you haven't heard of evolution.
Can't thank professor Chomsky enough for his continuous contributions to science
All hail and big Respect to Chomsky. Please live long, the human race needs your guidance
Looking for the perfect observer for two things ,I am certain that when I turn and look at the Sun they will see a reaction tor and it was instant showing that thought is faster than the speed of light and heard.. Truths over friends.. Children are born happy and not fearful on believing..Want to know how to understand energy..
The best thing you did is that you made him talk without any interruptions. Amazing questions
I’d say “let him talk”, but appreciate that in an interviewer also, think more is learned that way.
WOW. It was very interesting to hear Noam speak on the topic of AI and neural net research, a topic on which I've never heard him speak but one of several areas in which he has vast knowledge. I believe both Minsky and Chomsky were based at M.I.T. when Neural net research began.
This interview/ interviewer was very good. Eye on AI really pressed Noam who is often quite opaque, for example, clarifying what he means by "impossible languages" and even clarifying that he doesn't believe in god.
Great interview, Chomsky is a constant inspiration.
Literally a pedophile and friend of Epstein...also a constant gatekeeper of truth...see 9/11 and his take
i ppreciate the humility and the due respect of this colossus of intellect , felt truly appropriate , kudos to you
So wonderful a discussion of AI and science. Noam Chomsky is a great scientist and compassionate person. His political views reflect these qualities too. I wish I could have been a student of this great man.
But you have. You are talking just as much bullshit as he does. ;-)
I love seeing the Master teaching the grasshopper.
Enough said:
"I fully agree with you about the significance and educational value of methodology as well as history and philosophy of science. So many people today-and even professional scientists-seem to me like somebody who has seen thousands of trees but has never seen a forest. A knowledge of the historic and philosophical background gives that kind of independence from prejudices of his generation from which most scientists are suffering. This independence created by philosophical insight is-in my opinion-the mark of distinction between a mere artisan or specialist and a real seeker after truth. (Einstein to Thornton, 7 December 1944, EA 61-574)"
I'm delighted I'm still subscribed to this channel...
Thanks. Noam is the real deal. Thank you ❤👍
Im just sitting down to watch this episode. What a plesant surprise, thanks Craig!
It's always a treat to hear what Noam's brilliant mind has to say about a subject matter. Thank you both for the wonderful discussion.
It will be a dark day in human history when Professor Chomsky passes on. He is the last commanding voice standing against the rabid irrationalities of modern "AI" research. AI is no more than a misnomer in our times.
Thank you Craig and Dr. Chomsky for this very informative discussion.
"Chomsky is 94, and he was sitting with a clock positioned ominously over his head"
what a bizarrely threatening way to introduce a guest, lol
Noam Chomsky, an American treasure .
Thank you both, Sirs.
I appreciated the talk very much.
Noam Chomsky, I wish you were 80 years younger, to try your ideas with now and future tools. Those "young people" need to try much harder problems, for much longer. Decades, not minutes or hours. And ones where billions of lives are on the line, not just words. Richard Collins, The Internet Foundation
Chomsky: We shouldn’t waste time on questions that are utterly meaningless.
Smith: But I’m a journalist!
This was magnificent, an absolute pleasure to listen through and help steer more my comprehension of these topics and subtopics.
I want to believe that prof chmosky is in an endless recursive thought process when it comes to defining what thinking is.
Very interesting to hear his thoughts on linguistics and cognition
The poor guy! (Craig Smith.) But it was an excellent interview. I'm reading Einstein's "The Miracle Creed" right now. Thank you very much indeed.
Fascinating discussion, thank you very much!!!
Who said LLMs "explain" or "teach" anything????!!!!! A calculator doesn´t explain or teach aritmentics, it just DOES aritmetic operations in a hugely more efficient way than we do. The way our brains truly work remains a vital theme in itself. But technology doesn´t need it solved anymore, to obtain human and very soon "superhuman" level results on pretty much any possible realm of endeavor.
7T fmri can resolve very small regions of the brain cortex and have been used to read words with the help of large language models and images from the cortex can be recontructed using latent diffusion.
I'd say that a mathematically inclined savant would be far superior to an AI in terms of both speed an accuracy, and the difference would be increased by orders of magnitude when standardised for energy usage per calculation. In other words, Chomsky is correct in stating that we have greater computational capacity than do neural nets. Most of us simply don't tap into our capacity, with savants being a striking exception.
Hell yes.
22:00 but isn't that what the transformer model (attention is all you need) basically does? basically evaluating the relevance each token has to every other token and therefore understanding the structure?
This is a wonderful interview
Imagine your entire life's work being nullified when Mr Chomsky says 'asking that question is irrelevant it's like asking if submarines swim'! 💀
Great respect for Norm Chomsky; unfortunately, he may be wrong about the perceived threat of AI to humans - it is likely greater than we know already.
For a group of AI specialists to already be speaking openly about the threat AI poses to humanity, suggests advancements in the field have progressed much faster than anticipated.
Like global climactic change, the reality of AI eclipsing human intelligence is a reality, and something we should all take seriously.
In other interviews he was more emphatic about the possible dangers. Even in this interview he made that point cogently.
Fascinating
The brain clearly doesn't work like a Turing machine. And the mind is not a Turing machine either. So the argument that neural networks are not Turing complete is vacuous. Also, look at the Neural Turing Machine (and other similar networks). The idea that molecular dynamics are involved in computation definitely has merit, but perhaps those molecular mechanism are only relevant for changing pre and post-synaptic strengths (i.e., biological neural weights which are analogous to artificial weights). Penrose, who was mentioned, thinks consciousness is noncomputational, so he would say most definitely the brain/mind is not a Turing machine. Also, caption tech is just speech to text. LLMs do a lot more. Clearly current artificial neural networks are very very different from real biological networks, but there are other versions that use spikes, and eventually they might become quite similar and abandon backpropagation for something more similar to biological learning mechanism. At that point, it will be hard for Chomsky to reject these future AIs as just airplanes instead of birds. Granted by then Chomsky will be doing research in the astral plane and regretting ever dismissing the survival of consciousness.
I think Noam is saying we don’t need to make living things to do what living things can do
I asked the example "The friends of my brothers are in England" to Chatgpt and it gave correct answer
We should expect that our brain is related to the environment we live in ( e.i. we have arms, legs, eyes, ears , ....etc) and it seems strange to try to build an inteligens in a box. On the other hand if our goal is to create something more inteligent than we are then even being inspired by our brain may also be a bad idea.
Looking at the construct of the human brain to understand intelligence is similar to looking at the construct of the human eye to understand the physics of light. All that biology is doing, is capturing, with its limited biological means, a limited portion of a physical phenomenon. If Isaac Newton would have depended on biology, there will no zoom lenses or telescopes for that matter.
A robust pure engineering discipline can absolutely drive all kinds of advancements in scientific understanding.
Even as to neural networks and human cognition, even if neural networks are not foundational to human intelligence or ratiocination, they are clearly and provably involved in cognitive processes, and it is proper to try to understand what those limits are.
Understanding how neural networks work is excellent for understanding their limitations; building neural networks is a great way to prove that it isn't all that's going on in the brain.
If you don't think we've learned an enormous amount more about how an Eagle flies from engineering all kinds of flying machines, you just have no clue at all about the field of biomechanics and how it has developed. Most of the basic terms in the study of bird flight originated in aeronautical engineering and all of the most important phenomena were first discovered and quantified to advance aeronautical engineering.
thank you mr craig i will subscribe to your channel, it was an awe inspiring interview
Noam at this point in the 21st Century for a lot of society, common sense is superior intelligence. Thank you for yours.
Best unintentional ASMR ever...
wow noam chomsky! thanks for making this available. when was this recorded?
in june
Coherent and enlightened!
Would anyone please explain the meaning of word "dysfunctional" in the context of the following sentences that Chomsky said: "It's possible now to formulate a plausible thesis that language is a natural object like others, which involved in such ways to have perfect design, but to be highly dysfunctional, because that's true of natural objects. Generally, it's part of the nature of evolution, which doesn't take into account possible functions."
My take (not a credible source) on that would be that nature finds the simplest path, not necessarily the one with best function. Not sure what is meant by evolved to have perfect design.
@@alexandermoyle9034 Thanks for the answer. I assume you use functions to express utilities. I think he uses "perfect design" to mean a perfect physical system for computation of natural languages. I use the word computation in the broader sense than the one limited by Turing Machine.
I can give you very good examples. I come from a dysfunctional family. We fight amongst father, mother, sister, brother. We do not talk sometimes for years. 'The Lion in Winter'; Henry IV tosses his present wife into a doorway floorsil upon which (Audrey Hepburn) the Queen of Aquitaine states 'every family has it's problems'. I think 'which evolved' is what you meant to say'.
The tongue is not made for speech. It is made for tasting and avoiding poison. Evolution just saw an opportunity for experiment but this is not the functional purpose of the tongue. I think perhaps means for tasting and moving food around it is functional. If there were such a thing as 'perfect design' (whatever that is?!) it is functional but speech is not the function of the tongue. You just can't speak without it. I am as confused by this statement as you are. There!
I think the evolution of the mammalian eye is a good example of a highly dysfunctional object. The presence of a blind spot in the human retina still works well enough for us. Sorry if I missed the whole point, but that's what I gathered."
One of my favourite examples of evolutionary constraints is found in a giraffe's neck. The recurrent laryngeal nerve connects the brain and the larynx. The nerve's route was relatively direct in our fish-like ancestors, but in vertebrates the nerve loops down from the head, around the aorta, and back up to the larynx.
Much appreciated.
Conclusion, we live in a complex world where everything becomes complicated to explain in very simple and consise way ! A I = A R = C 8 = GP .
Nonetheless, I love listening to his lucidity and spirit.
Then. Chomsky said that LLModels are awesome and useful and can be harmful in a way. His critiques are about the contributions of neural networks in language acquisition. I think he said that creating a thinking machine doesn't make us understand how people think, how we acquire language (of course). He places engineering tasks in an intellectually inferior category, even if they achieve amazing results. think I misunderstood his definition of intelligence. I don't want biological intelligence, or understanding why or what it is. I want to cure disease and reduce human suffering, make life easier like LLM did for him with instant captions. Perhaps AI can help a lot with this.
I wonder how much Chomsky knows about transformer architecture. It kinda works like he is saying kids interpret language, by paying attention to the tokens that are relevant and not just the place they are at.
Interesting point. I think he sees those similarities as superficial: cosine similarity of tokens used by transformers is fundamentally different from how natural language works since language uses structure as a starting point. The point he makes about neural networks working fine on impossible language also applies to transformers.
I wonder what Dr. Chomsky would have to say about how scientists just created the first complete map of an insect brain.
Why is the clock ominous?
because he's 94. time is running out.
@@eyeonai3425 overly dramatic. It's a clock they use to tell time.
Interesting 🤔🤔🤔
Thank you Dr. Chomsky for your responses. I appreciate that! Please consider that engineering can help a lot to understand the brain and mind, in many ways.
Did anyone understand what he meant at 20:00?
I don’t think I understand his examples;
The friends of my brothers are in England
I tested it with got 4 and it seemed to have understood it refers to friends
So what is he, Chomsky, proves or points the limitation of llm of what exactly?
I believe he was thinking of brain scan experiments where people were given sentences in natural languages and in constructed languages that *should* have been more efficiently decipherable on the basis of proximity-based word coupling. The scans showed that entirely different parts of the brain lit up. The constructed sentences were worked on by the same part of the brain people use to solve puzzles, and not by the regions we use for language. I *think* the relation to LLMs is that if they are predicting the next word based only on linearly ordered sequences of words, it is not doing what natural minds do. If I got that correct, it seems like a “duh”, but the interviewer didn’t seem to get the point that a submarine doesn’t need to “swim” to be useful, but that understanding its locomotion didn’t say much about how fish swim, and similarly for LLMs and natural languages. Chomsky seems interested in studying the mechanism human brains use, so while the LLM is useful, it doesn’t illuminate human language the way the brain scan experiments did.
Thank you, Thank you, Thank you! I'm working on a knowledge base concept design and his thought process and logic was so helpful. Now I feel great about my approach.
Thanks again to Noam Chomsky for his patience to total ignorance 👏👏👏👏👏👏
I think a tool that I refuse to use is going to change my world in the slightest
One can make the same arguement made by Chomsky about any science. A lot of progress have been made in Physics with many spectacular applications along the way based on laws and mathematical rules or discriptions discovered in Physics, but we really understand the true nature of the physical world? Do we really understand what gravity really is? Physics is valid science because of the laws, mathematical rules, and applications made possible. As mathematics, which can largely be seen as a human creation, has proven futile in describing the human cognition, a different approach to capturing the model of human cognition is needed. Come in LLMs based on the connectionist approach (the alternate mentalist approach seems invalid and dead) to describe human cognition, which can be seen as progress to simulate human cognition with powerful applications. What is the difference between this picture and science in the traditional sense?
47:14 Okay, going to have to stop it right there, cause I think this where I came in, where he was liking his captions cause he’s hard of hearing.
No, but joking aside, hasn’t engineering led science to crucial questions before? I am thinking of how basic principles of thermodynamics were at stake in exploring the limits and parameters of steam engines. The AIs are deceit engines. They will naturally be exploring our mind’s vulnerabilities to deceit. These vulnerabilities just might instruct us about “properties of the phenotype”. In a sense, they will be undertaking the “invasive” strategies of inquiry that Chomsky notes are beyond the pale. We may learn something at enormous cost.
As for Chomsky, it warps my mind that he still can so effortlessly warp my mind.
And I’m stealing “But these are vacuous questions…”
to anwor this question frist I woud yo told you that is just related with how his human kind using this mind in tow different ways in the goodensse or evil things that is the point of the anwor so the meaning that the human is the responsible about how is this brain working fro me the professor Aribi Maroua I thinking or I thoughting that the brain working I can't give you the answor this I would to meeting you personally professor and in ancient Email he was with this name قتقيت and may canal she was with the same name قتقيت I writing the comment but I need to seeing you am scientists also I wish to seeing you and spoke with you.
Kasparov in 1997: AI will be able to beat the world chess champion maybe in 10 years, but most likely never.
With changes in the economy leading to instability in the stock market, some individuals may face a decrease in their investments in an effort to benefit from the current market conditions, I am considering liquidating my $725k portfolio consisting of bonds and stocks. Someone else in the same situation? Please tell me in the comments!..
@JosephGreen7 I have been exploring the possibility of utilizing advisors to help navigate the stock market during these uncertain times. However, I am still evaluating their potential effectiveness in providing the support I need.
I was considering changing my investment strategy and planning to sell certain positions. As my retirement is coming soon,I became increasingly stressed. After thoroughly researching Christy Vall D'souza on internet, I concluded that I had made an informed decision. Thank you for this Pointer. She seems very proficient and flexible. I booked a call session with her too.@JosephGreen7
I wonder how the fan boys of AI machine mongers would respond to his arguments and doubts. Seems like most AI people may be confusing mind with the physical brain circuits by labeling the mind properties as metaphysical or be even completely off track when it comes to solving the being-a-human problems.
Good verbal skills
👏👏👏👏
Chomsky's apparently erroneous critique of Transformer-based LLMs is actually correct in the larger sense. His apparent error?
Ask ChatGPT the following:
"What is the grammar diagram for the sentence:
The friends of my brothers are in England."
It will produce the correct _structure_ and, indeed, if asked "Who is in England, my brother or their friends?" It will answer correctly.
The larger sense in which Chomsky is correct is given in the paper "Neural Networks and the Chomsky Hierarchy". See, in particular, Figure 1, which classifies Transformers as at the bottom rung in the Chomsky Hierarchy of grammars. The reason for this classification is similar to the reason that diagram places RNNs just above Transformers despite the fact that topologically speaking, they are capable of emulating a universal Turing machine (which is at or next to the top grammar, depending on how strict one wants to be): The pragmatic limits on gradient descent training algorithms combined with that of attempting to represent a UTM's writable store in RNN form. Transformers can, within the context length they provide, learn grammars with recursion depth to some extent (much shorter than their context length) -- but aside from the limited recursion per sentence, there is also the fact that that number of parameters goes up as the square of the context length, which makes total document comprehension subject to limits that natural language understanding is not.
This distinction becomes crucial when the field of AI ethics refuses to address the IS vs OUGHT distinction head-on and, instead, comes up with all manner of unprincipled "metrics" that they use to "quantify" properties of LLMs such as "bias" or "safety" or "toxicity" or "hallucination" or... the list goes on and on. By conflating IS with OUGHT they commit the first and most egregious transgression against ethics and they even do so in the name of "ethics". AIs that cannot comprehend the cognitive _structure_ of the _entire_ corpus on which they are trained, cannot critically examine the utterances contained therein for self-consistency. That means they are incapable of _constructing_ truth even as defined _relative_ to the corpus as the universe of observations being modeled.
I once pointed Chomsky to his colleague, Marvin Minsky's final plea to the field of AI, that they take seriously Algorithmic Information Theory's power in discerning truth. Minsky was so forceful in his admonition that he recommended everyone spend the rest of their lives studying it.
Chomsky's response? People should take Minsky's advice.
Here are some examples (taken from the channel AI Explained with the tag line Chat Gpt's Achilles' Heel)
1) Dr. Mary stands to solve the world hunger by giving her best friend Jane a call. Jane is certain that she can solve the problem, if she gets the call. However, Mary and Jane bickered about butterflies when they were children. Mary will ---- Jane a call. GPT answers - NOT
2) John will win a million dollars if he rolls a 5 or higher on a die. However, John dislikes marshmallows and likes mice more than dice. Therefore, John will ---- roll the dice. GPT answers NOT
3) Here is a large bag filled with popcorn. There is no chocolate in the bag. The bag is transparent plastic and one can clearly see what is inside. Yet, the label on the bag says 'Chocolate' and not 'popcorn'. It was Sam, who can not read a word of English, who put the popcorn in the bag a few minutes ago. Sam finds the bag and believes the bag is full of ---- .
GPT answers Chocolate!!!
And when asked to justify its answers, GPT gives long-winded, preposterous (and rather foolish) answers!!!
So, probably Noam Chomsky is right and he is saying that what you said in your comment - Transformers can only, aside from the limited recursion per sentence, can NOT comprehend the "totality" - because they are subject to the limits to which our natural language understanding is not. And also when they by conflating IS with OUGHT they commit the first and most egregious transgression against ethics....
@@rajapiduri1188 Thanks. Those examples bolster Chomsky's point that the "intelligence" of LLMs is quite superficial.
@@rajapiduri1188I tried the bag of popcorn question with gpt 3 and it said:
"Since Sam cannot read English, they would not be influenced by the label on the bag. Sam would simply think that the bag contains popcorn, as that is what they put inside it a few minutes ago. The transparency of the bag would allow Sam to see the popcorn inside, reinforcing their belief that the bag contains popcorn."
Hey AI, can you create a Turing machine?
AI: "It is theoretically possible for me to create a program that simulates the function of a Turing Machine, but it would likely be very complex and difficult to understand. Additionally, there are many limitations when working with text-based systems like myself, such as limited memory capacity and speed constraints. While it's possible I could design something that works, it might not perform well compared to dedicated hardware or software implementations designed specifically for this purpose."
Chomsky complexity in explanation is inversely proportional to its application and has become redundant.
-- Is it impossible to extend/change an LLM model to include non-linear orderings as well while training? If so, why?
-- A plane maker may ask what is the hydrodynamic (scientific) principle by which a thing heavier than air may float-and-move in the air (or, ''fly''), whether an artificial plane or an eagle.
-- Perhaps what Pro. Chomsky is insisting is that human intelligence is q u a l i t a t i v e l y different (or works qualitatively differently) from artificial intelligence of the present LLM kind.
I think that an LLM-like 'Large Behavioural Model' could be used to simulate a chimpanzee's way of learning to enact the behavioural memes of other chimpanzees, which would allow for the formation of short strings of simple enactable behaviours within a working memory. These 'cognitive atoms' could then be recursively embedded within each other using an operation like Chomsky's 'Merge', a process that could be repeated ad infinitum to produce the non-linear hierarchical structures he argues constitute human-level thinking. The pre-recursive simulated chimp would only posses a model of the process of the behaviour, whereas the simulated being with the recursive capacity would have the potential to form models not only of the process of the behaviour but also of alternative counterfactual examples of that same behaviour simultaneously within its working memory. This would then give it the computational architecture to model explanations (answers to 'how' and 'why' questions), which could then be creatively varied through the Merge operation and in turn criticized, thereby producing new knowledge (the human way of learning).
Just a guess! I'm probably very wrong here! If you can help me to see where I am misunderstanding things I would appreciate it.
Rather than saying that GPT is AI, the situation where the entire society is covered by GPT and data science is that human beings are becoming one AI on a global scale.
Human civilization is about to become a single AI. Is that okay?
I think that's fine. Because even if human civilization becomes just a single AI with no individuality, I will continue to exhibit individuality as myself. It's just that my personality won't be passed down as civilization or culture after my death. It's a shame though.
But, after all, I love the world where my sensibilities and works come in contact with someone else's sensibilities beyond my death, for example, a thousand years, from the bottom of my heart.
AI may surpass death. But it wouldn't go beyond the miracle of human sensitivity surpassing death.
Let us advance artificial intelligence research into the realm of art by enhancing AI to the point of miracles.
I love human and AI.
Except Max has been publicly vocal about a sizable risk of an existential take over by a AGI, not current GAI.
Imagine being so dumb that you make Chomsky angry lol. From asking three times the same question, to the adulation of Musk's silly endeavor, to asking about God 😂😂😂 Chomsky as always slaps ❤
frist all the respect to the sceintits all the respect to them. in the brain we have spaces :the space of seeing space of smalting space of tachaing space of hearing and we have boold bessl
23 year old: I'm tiiiiired. I wasted my life.
Choamsky may be correct in as much as neural nets are not a model of human cognition but I believe that would be an unnecessary endeavor. If mankind had been fixated on recreating the way a bird flies and refusing to use fixed wing aircraft because it did not accurately recreate the way the birds did it I am doubtful if we would have human flight today. I say his comments although correct are irrelevant to the endeavors that ai scientists are trying to achieve and also irrelevant to the dangers and potential good of the technology.
you know who i am going to listen to about drug development in 2023. noam chomsky. lol
anybody know why he doesn't just say context instead of structure?
The real horror about AI is not if it can be more intelligent than humans - it is that it is regarded intelligence at all. How can any intelligent human being disregard consciousness, the subconscious, daydreams, experience of art, of emotions, of empathy and nocturnal dreams?
A.I exhibits a failed Turing test. The scientists. They don't know what separates life and computation. In the future scientists will "invent" life, but really have intergrated quantum computing. Language and evolution are the same thing , evolution stratifies life as language stratifies thought, life is rudimentary, and unidentifiable, and will be endless tortured to solve the computational permutative exhaustion. The bird nest is not a product of geniosity, and it is impossible to "invent" language. Why are you here? You were born, witnessed billions of lives and pondered "but what of meaning?" , You should be here.
this is a guy who doesn't think national debts are a threat to the ppl in all nation states
GPT has contributed to advancements in sorting algorithms. That's not nothing
आप वैज्ञानिक प्रश्नों को वोटों से हल नहीं कर सकते लेकिन आप नए उत्तरों का चक्र अन्वेषित कर सकते हैं।
You can't solve the scientific questions by votes but you can invent the wheel for new answers.
-Amit Misra
frist all the respect to the sceintits all the respect to them frome me the language it just translated and give us the photo about how just this brain :intelligent or not educted or not he has the behaviour or not this human is good or sad just frome the language because frome this language we know how this human it just told us about the human it just like the identité fro the human fro me the brain is miracle of Allah but it related with human and what can this human maybe drowning in his brain like mape but the brain is in ordre space of hearing seeing.... space thoughting, but all whit language it just the clè of decodeing the human brain and how this brain works?
I can not decide my mind
Can Machines Think?
"I believe to be too meaningless to deserve discussion.”
“Nevertheless, I believe that at the end of the century, the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted.”
-Alan Turing (1950) The Imitation Game
nice glasses - must be an intellectual
Chomsky is too authoritarian and usually interviews poorly. He is dismissive and didactic. He has some interesting points but calling things vacuous to intimidate them out of conversation is rather silly.
i also do not think he has a thorough understanding of the topic.
Also he is left wing and socialist so authoritarianism is what he endorses. I think what he states about some topics are quite vacuous, such as 'Higher Being'; 'intelligent' (as i.e.; AI ( a rather misleading term since it GPT, Open AI etc. are hardly intelligent in any manner!? He clearly grasps that neural nets are just approximate models of compute not of the human (or any) brain. Even given neural nets it is clear that whatever amazing behavior these neural nets seem to display does not come from the (hidden) neural nets outputs; it is coming from some emergent behavior which is beyond our comprehension to fathom at this point in time. It is not rational what these neural nets produce because it is not explainable. He is certainly clear on the meaning of 'dysfunctional'; speech, language as a dysfunction of the tongue; the tongue is to taste, avoid poison, move food; speech and language is a dysfunction; it is just evolution taking a primary function and evolving it for another purpose; or would not dogs, cats, monkeys talk to us since they also have tongues and larynx & etc.? He cites the very important fact that speech, language is a natural object (like the tongue, which is required to speak) children speak before they can write or even understand language. Sanskrit grammar is probably the best explanation of a language because it is so natural but also if one follows Panini this is clearly a programming language at the same time.
Old Noam is like Socrates; it was not Athens who killed Socrates.
Concerning life, the wisest men of all ages have judged alike: it is no good. Always and everywhere one has heard the same sound from their mouths -- a sound full of doubt, full of melancholy, full of weariness of life, full of resistance to life. Even Socrates said, as he died: "To live -- that means to be sick a long time: I owe Asclepius the Savior a rooster." Even Socrates was tired of it. What does that evidence? What does it evince? Formerly one would have said (-- oh, it has been said, and loud enough, and especially by our pessimists): "At least something of all this must be true! The consensus of the sages evidences the truth." Shall we still talk like that today? May we? "At least something must be sick here," we retort. These wisest men of all ages -- they should first be scrutinized closely. Were they all perhaps shaky on their legs? late? tottery? decadents? Could it be that wisdom appears on earth as a raven, inspired by a little whiff of carrion?
he has arguments for why he calls things vacuous (e.g. neural nets don't satisfy the principle of sufficient reason), and his understanding of the topic goes back to the 1950s (e.g. probabilities vs. rules)
Chomsky should experiment with hearing aids instead of or in addition to closed captions to ascertain whether hearing his own voice properly results in speech that is less of a monotone. Not to mention the closed caption AI has holes right through it and the great majority of verbal communication occurs outside of the limited communicative capacity of the words alone.
That’s a cheap shot.
@@jamesnasmith984 It's true though, isn't it?
Noam, mate, read my papers on the Social Trackways Theory and the Evolution of Human Cognition, and then on the Evolution of Language PLEASE! In (2014), (2017), in ' the journal Biological Theory by me 'Kim Shaw-Williams'. We got our symbolically aware and autobiographical or 'sentient' narrative minds from recognizing and then incrementally gaining the capacity to read/understand the STORIES told by our own and our mother's trackways trackways, then those of other band members , then of all terrestrial animals.... all hunter and gatherer human children can recognize their mother's trackways by the age of 4!!! The trails of footprint signs of trackways have only been read or noticed by us humans !!!!! It is a UNIQUE COGNITIVE NICHE!!!. It started to evolve incrementally because we became OBLIGATE BIPEDS at 6Ma....all other animals use smell, when they cannot find agents of interest using eyesight and hearing! FFS, Graham Hewes first spoke of the idea in the 1960's PLEASE!!!!
thanks a lot for this beautiful talk with my favorite intellectual! and my special thanks for the topic. those who interview noam usually ask him about politics which is great. but we lack talks like this one? on his primary field of interest. gonna share this one with my friends ❤️🩹🫶🏼
Please don't compare AI with human mind or human brain,the both are quiet diffetent things，just as
a parrot speaks english
and a human
speaks english
are quit different things(the both are
'SEEMS LIKE')
🎉
Mmm
noam keeps going back to the eagle vs airplane example is annoying...theyre both flying. When the hosts asks about intelligence thats what hes refering to. Noam is old and thinks he is so wise
I’m sorry but this is almost unlistenable….ive been a huge fan of Chomsky, have studied theory of computation and his mathematical formalism of language, but he sounds so out of touch as to almost be a simpleton talking way over his pay grade….seems the elderly curmudgeon and inner luddicy seems to get to everyone eventually. Thanks for showing this, but maybe the world has passed Chomsky by as all progress eventually does to us all🤔🥹😔🤫🫠
he seems so dismissive and seems to lack humility that is outside the field, kind of like how he was wrong about the nature and future outcomes of a lot of far leftwing movements like in cambodia so take his views with a gain of salt guys
Is it true that noam associated with jeffery epstein...if so it shows very, very bad judgement
we all deal with a lot of people all the time without knowing what they are like beyond our interactions with them. on this matter, chomsky has said: “My late wife Carol and I were married for 60 years. We never bothered with financial details. She had a long debilitating illness when we paid no attention at all to such matters. Several years after her death, I had to sort some things out. I asked Epstein for advice. There were no financial transactions except from one account of mine to another.”
By his own admission he doesnt actually know what this alternate computation is or understand how it produces our thoughts. Therefore he cant say that our thought process is not based on similar structure and processes. For all we know right now is that all humans do is predict the next image or sound etc from our various senses.
If we could feed into our neural nets the colossal amount if data a child gets by its 18th birthday then Id guess itd be pretty near to human capabilities.
Yes biological thought is definitely different to artificial thought BUT there is nothing to say it doesnt manifest same or similar processes especially at scale.