Noam CHOMSKY on AI, ChatGPT, Universal Grammar and Wittgenstein: Unlocking Language & Mind

🎙️ Join the Conversation with Noam Chomsky: Unraveling the Truth Behind Artificial Intelligence (AI), Universal Grammar, Language, and Mind!
🌟 Samuel Marusca engages in a fascinating conversation with the legendary linguist, Noam Chomsky. In this thought-provoking discussion, they skillfully debunk the myths surrounding Artificial Intelligence while delving into Chomsky's groundbreaking theory of Universal Grammar and its profound implications on language and cognition.
🤯 Are you curious about the false promises of ChatGPT, and the limitations of AI? 🧠 Wondering how human language truly emerges from our minds? This conversation explores Chomsky's recent article in the New York Times about the 'False promise of ChatGPT' which will unlock new perspectives and challenge your understanding of language and intelligence. Chomksy also shares his insights on Wittgenstein, The Sapir-Whorf hypothesis and Consciousness.
📍 SUBSCRIBE: www.youtube.com/@Practical.Wi...
🎧 Listen to the full interview on your favorite podcast platform:
Apple Podcasts: podcasts.apple.com/gb/podcast...
Spotify: podcasters.spotify.com/pod/sh...
Amazon Music: music.amazon.co.uk/podcasts/b...
Google Podcasts: podcasts.google.com/feed/aHR0...
SUMMARY:
2:06: 🤖 Noam Chomsky discusses the limitations of weak AI in achieving true intelligence.
6:53: 🤔 Chomsky discusses the capabilities of ChatGPT and whether it can be considered a form of thinking or consciousness.
13:31: 🧠 Consciousness only provides fragments of what is happening in our minds, while language and thought need to be studied scientifically rather than relying on introspection.
20:21: 🧠 The nature of human thought and language is a complex and unanswered question in science.
27:30: 🔍 There is evidence of a shared capacity for language among humans, and are similarities between language, music and arithmetic.
38:55: 🗣️ The nature of language and the relationship between word and object (Quine).
42:29: 📚 'Possession of knowledge' and 'use of knowledge' according to Aristotle. Wittgenstein and the Sapir-Whorf hypothesis.
53:11: 🗣️ Universal grammar and the poverty of the stimulus.
57:40: 🗣️ Language changes externally but not internally. Internal language is not accessible to the mind.
#noamchomsky #chatgpt #wittgenstein #debate #interview #philosophy #practicalwisdom #artificialintelligence #ai #grammar #language #linguistics
Soundstripe: HCMGGEW4C0S6GCAH

[@Practical.Wisdom]

📍 Don't forget to SUBSCRIBE so we won't forever miss each other in the black hole of the KZbin algorithm: www.youtube.com/@Practical.Wisdom?sub_confirmation=1

[@doreekaplan2589]

KZbin disallows unlimited subscribing. Only can subscribe based on having ones own channel plus getting their choice of a certain number of your own subscriptions. So I want to add your channel but youtube Won't let me tonight.

[@Practical.Wisdom]

@@doreekaplan2589 Sorry to hear you're not able to subscribe!

[@ioanagrancea6091]

The best interview I have seen on the topic of AI. Such clear thinking, direct and deep explanations, intellectual charm of the best kind.

[@fabiosky4368]

great work of this man studying thestructures of languages.in italy there was a professor like him ,he was a semantic philosopher ,U.Eco. music is an abnormal language , a miracle,creation of men,and is a great issue studying the meaning of the different types of music, each born in different social ambients ,ciao from italy.

[@doreekaplan2589]

Excellent. Pertinent questions. As an author I love the type of person who lives with a wall- filled number of books.

[@opusdei1151]

What a great interview, its mind blowing to me that you have in 6 month only 15k views. Thank you for the interview

[@Practical.Wisdom]

Thank you for the kind words! We've had some teething issues with the audio.

[@doreekaplan2589]

If you find the numbers of others important and think this information is something others can benefit from.....pass it on; do something about it.

[@onestepahead1857]

Completely agree. Great interview. Thank you.

[@Practical.Wisdom]

SUMMARY: 2:06: 🤖 Noam Chomsky discusses the limitations of weak AI in achieving true intelligence. 6:53: 🤔 Chomsky discusses the capabilities of ChatGPT and whether it can be considered a form of thinking or consciousness. 13:31: 🧠 Consciousness only provides fragments of what is happening in our minds, while language and thought need to be studied scientifically rather than relying on introspection. 20:21: 🧠 The nature of human thought and language is a complex and unanswered question in science. 27:30: 🔍 There is evidence of a shared capacity for language among humans, and there may be similarities between language, music and arithmetic. 38:55: 🗣 The nature of language and the relationship between word and object (Quine). 42:29: 📚 'Possession of knowledge' and 'use of knowledge' according to Aristotle. Wittgenstein and the Sapir-Whorf hypothesis. 53:11: 🗣 Universal grammar and the poverty of the stimulus. 57:40: 🗣 Language changes externally but not internally. Internal language is not accessible to the mind.

[@Daily_Llama]

Why do people still listen to this used car salesman? He is a linguist professor that uses language to manipulate people. He has no real experience or knowledge on any topic he discusses. He just gets hired because nobody in the respected fields he speaks on, can manipulate people into thinking like them.

[@oviguitar]

Thank you for this great content!

[@Practical.Wisdom]

Thanks for your support!

[@MamangFF-mn4hn]

Subtitle is a program that I need to understanding topic clearly.. Thanks

[@carmel-housechurch6552]

Extremely interesting and intriguing!

[@Practical.Wisdom]

Thanks for your interest!

[@TupperWallace]

This was a very interesting, stimulating interview. It is sobering to be told that so much is hidden from our understanding and comprehension. It’s reassuring to hear that AI is only a bravura mechanical simulation of intelligent discourse. It may not take over the earth and wipe out humanity after all! He didn’t necessarily answer your questions directly or concur with their premises yet you elicited a fascinating and coherent presentation of his lifelong work.

[@Practical.Wisdom]

Thanks for your comment! Exactly, most of our language is inaccessible to our consciousness, which is a sobering thought. Yes, Noam didn’t always concur with my premises, for example, he is very critical of ChatGPT and doesn’t recognise its benefits. I think this is because he is looking at it with a scientific mind, rather than a user. Noam is one of the greatest minds of our time.

[@entelin]

@@Practical.Wisdom At the end of the day I think it's the height of folly to suppose that we will never figure out and dramatically improve on the spheroid that is sitting on the top of our shoulders. There has been continuous progress on AI for decades, there will continue to be. If LLM's today can't sit shoulder to shoulder with an educated persons ability to reason today... Well just a few years ago we couldn't communicate in plain english and get reasonable answers back. I'm a programmer, I've worked with the openai api, I use openai in some way almost every day, even what I know to be possible right now is enough to cause an earthquake throughout the economy. It will only improve.

[@mdaniels6311]

Chomsky is simply wrong and severely naive about Chat GPT's power. The Transofrmer acrhitecture in new LLMS hace self attention abilities, just like a human does. They are not sentient, sure, but they do 'understand' language, and are just improving ever day. I kind of think of it like we just removed the language centre of our own brain, and put it in a box. Chomsky opposes it for ethical reasons, but says nothing of their application or abilities. He is also failing to realise it is not neccesarilyu the AI that is the problem, it is what people can do when using it. Also, he knows (I think) that LLMs cast serious doubt on his Universal Grammar theory in linguistics, which to me, is no different from debunked cartesian dualism, suggesting humans are 'hard wired' for language, and thus this cannot be replicated. This is of course silly, as a human is merely a biochemical machine coded by DNA, and neurons in our brain are simply in a binary state of off and on (deploarised/hyperpolarised). This simple on/off gives rise to consciousness, and there is no fundamental reason why this cannot be replicated in a machine, unless you want to argue that humans have some kind of supernatural essense, which is clearly not the case (and I say that as a Christian!)@@Practical.Wisdom

[@MoralScienceEducation]

Thank you so much fpr thr amazing talk(s)🙏. Assuming language and cultural affinity both, influence our perception of reality, how does this relate to multi-lingual and mult-cultural experiences in humans, and for AI to meet their moral expectations?

[@Practical.Wisdom]

Thank you very much for your lovely comment! That's a really good question. Linguists usually reject the Sapir-Whorf hypothesis, although many still support its weak form. We touch on AI and moral expectations with John Lennox in the upcoming interview on AI scheduled for release at the end of the month.

[@mdshafiqurrahman1561]

Can u turn on auto generated subtitle for this

[@lidiaolcer6248]

Amazing, thank you!

[@Practical.Wisdom]

Thanks for your support!

[@christopherbettridge5983]

Prof. Chomsky makes a very interesting and often unremarked statement, that it seems exceedingly strange to us, generally, that humans feel or construct the idea, so much is it prevalent in our common sense, that we should have answers regarding the nature of consciousness which we don't have similar access to in the realm of the sciences; considering that the sciences themselves are built up of a contextual mode of reasoning upon empirical theories and set6s of data constructed in a way in which we can interpret them with the systems of our internal conscious modes and methods of thought. So much common sense prevails in our culture about things which are really deserving of a much less common mode of investigation than mere assumed sensibility amongst us.

[@attackdog6824]

Good choice of guest and topic. His philosophical opinions are seldom discussed- or at least not nearly as discussed as his political views.

[@Practical.Wisdom]

Thank you! Yes, I totally agree, his earlier work on philosophy (such as his theory of mind and the nature of knowledge or language games in Wittgenstein) is often overlooked. In future, I want to have more philosophers on the podcast.

[@Pyasa.shaitan]

Brilliant 🎉

[@Practical.Wisdom]

Thank you!

[@ryanchicago6028]

Thanks.

[@Practical.Wisdom]

Thanks for your interest.

[@ryanchicago6028]

@@Practical.Wisdom :)

[@Dan_Campbell]

Thanks!

[@Practical.Wisdom]

Thank you so much, this is very kind of you! I hope you enjoyed my conversation with Noam. Let me know who you would like to see next on the podcast.

[@Dan_Campbell]

Noam again, Michael Tracey, Peter Joseph, or Daniel Kahneman@@Practical.Wisdom

[@Practical.Wisdom]

Excellent suggestions, thank you!

[@JoshuaEnfield]

When making conclusions about things that are social convention or not...I feel like in some discussions that's used to create context for what is "correct" or not. Does he ever speak to the possibility of genetic convention? Not just through the entire genetic ancestry of life, but also in branches of the evolutionary tree.

[@leonsantamaria9845]

Human vs machines ...... long life for professor Noam Chomsky....🫵👽👍

[@sublimeister9630]

Computers, whether digital or quantum, is still just a computer. It can compute, but it cannot think-here I include not the regurgitation of information but of imagining outside the box. “The Mind is a beautiful slave but a terrible master.” - Osho 🙏🏻😊

[@mdaniels6311]

A brain is merely a bunch of cells that are either polarised or depolarised. There is no fundamental reason why this cannot be replicated in software. Sure, we never crack the problem, or it might take thousands of years, but that is not the same issue of whether it is possible or not. We know matter can give rise to sentience, because we are sentient and made of matter. On top of that, there could be machine/biological minds created, perhaps using simpler organisms via machine architecture, somehow. In a novel I read recently there are AIs powered by lobsters, so maybe something like that.

[@hermeitus]

Could you add subtitle in Turkish plase 🙏🏻

[@fabiosky4368]

no one knows in Europe Chat gpt and what it can do!! the objective is creating a software thinking like humans mind and creating lot of not exhisting contents with terabytes of data memory. it could be a big danger, about the level of occupation in future! ciao from Italy.

[@alvodin6197]

You are a fanatic,there is NO artificial intelligence at all. It does t exist. Sorry,, it's hard for delusional people to believe. Artificial intelligence cannot think for it self,if you think so, you are misinformed. Some of the most sophisticated language tra slators can't even fucking distinguish between past and present the tense In a sente as "I have been studying now for about 4 .months. sorr

[@alvodin6197]

Dangers lol. The dangers would be if A.I could distinguish between present and past tense in an word sentence

[@aliiamshv777]

19:11

[@GBuckne]

..I asked CHATGPT what is the significance of this equation, 2x1.0079+15.999=, I didn't give the answer, and told me it can represent a molecule of water because these are the atomic masses, I also used the hyphen and period keys ..-- .- and it told me it was Morse Code, I asked is frame dragging in the gravity probe experiment a weak coupling of spacetime to mass it said yes....relatively simple examples but this is just the beginning...

[@Dr.acai.jr.]

Computers are called operational amplifiers and nomenclature conventions such as this one seems to me as malicious a designation as is universe for something Hugh Everett should be honoured with. Perhaps the shootenanny album by eels.

[@johnhelm6231]

Good job five stars

[@Practical.Wisdom]

Thank you for your interest, John!

[@AdeelAhmad-jf6sv]

It is a serious topic ...i wanted to acknowledge every detail of it but unclear voice ....

[@paulquigley3286]

I think Noam is talking past the goal of large language models. They aren't intended to give us insight about the human basis of linguistics. He is right that they are virtually useless for that purpose. The purpose is to replicate general human intelligence in silicon. By a variety of measures, they are very good at doing so - the best system that human beings have ever built at doing so.

[@dahoonkim1985]

100% right

[@adamtimol2801]

What is going on with googles search algorithm? I've been searching for "Chomsky" every day for the last month for new videos and this only just turned up today, five days after it became available.

[@bogdanbica1876]

🤔

[@holgerjrgensen2166]

The Day-Consciousness is our window to reality, the Thinking has No beginning, and No ending, the Day-Consciousness Never sleep. Intelligence can Never be artificial, AI is a dead mantra, superstition and illiteracy. It is all about programming, programmed Consciousness, a book is also programmed Consciousness, frozen Memory. 'Language without Mathematic, is Not a Language, and Mathematic without Language, is Not Mathematic', Development of Language and Consciousness, is two sides of the Same Development. Life and Consciousness is Eternal, can't be created, but Consciousness can be re-newed and Developed, and it happends through Developing-Circuits.

[@hahaHAHAHAHHAHhahhha]

"It's like asking wether submarines swim. You wanna call that swimming? Ok. Submarines swim."

[@opusdei1151]

Incredibly you also saw the video 😲

[@alphaomega1089]

Avatars are we all. An AI is such an entity. We made it so. The burning question: can we fashion our thinking into another? Good and bad. It can be understood by another avatar, but no reason to assimilate that mode of thinking. Nations. Cultures. Religions. Race.

[@user-kz5cw2gj3w]

Yes. The questions is...can humans?

[@cinzialbm]

Some

[@chriscandlin65]

My notification for this failed, late to the party

[@Practical.Wisdom]

Thanks for joining!

[@dogecoinx3093]

is music alive? is a radio alive? is language alive?

[@drilett]

Noam Chomsky, the Joe Kennedy of AI.

[@nothingmatters321]

Why isn't a question if frogs can grow wings? Suppose a martian observer were examining humans. The martian would certainly conclude they are not frogs. Its the same with machines.

[@admirljubovic717]

"do submarines swim"

[@fiddley]

I have a great respect for Chomsky but I feel he is struggling to reconcile his linguistic ideas with the new reality lain bare by the fact that these systems work, to a degree at least. The guy is getting on in years and formulated his ideology way before any of this was even a twinkle in Silicon Valley's eye so I don't resent it when he defends his corner, but I feel we're seeing similar reconfigurations of thought akin to when Einstein couldn't accept Quantum Physics, or when Freud's psychoanalysis frameworks were overthrown.

[@alanbooth9217]

so a 2 year old child has not scanned millions of text to make a best statistical guess to answer a question- has it even read anything- is the 2 year old child more intelligent than an AI borne just a seconds ago- is the concept of intelligence based on ability to output answers to questions or reason originally - but where does the child get information to reason from- not from copy and paste but from innate ability from day 1- computers are not borne with inherited genetic code hence the distinction- simulation not origination

[@radwanabu-issa4350]

This is probably the last video recorded of Chomsky, he is not communicating with the world for at least 6 months, he is sick at the end of his life! He is one of the most balanced realistic thinker in the world and he is an Ashkenazi ethnoreligious member! Human communication and especially language is the most wonderfull mystery still not understood! kzbin.info/www/bejne/ioKQdKWeaNeXh6Msi=Er0gk6yLDvY41Yqm

[@onkarvigy]

We all have great respect for Professor Chomsky. I think he is willingly playing the role of a contrarian/Devil's advocate here. Obviously there cannot be a better mind than his to mount a formidable defense against chatGPTs/AIs extraordinary/excessive claims often exceeding ourselves(for the moment). With that said , I would like to differ. This is definitely not for the distinguished professor.This is just for us mortals!! -----+++++ The consciousness hypothesis is as redundant as the God hypothesis because it bears no influence on the outcome of Scientific endeavours!! It was therefore abandoned long ago with the emergence of modern Physics. It makes it obvious that the scientific quest has been to find that ultimate Computer program that governs the World & therefore our so called "{Human(e)}consciousness". When you can train a bunch of neurons from the brain of a frog to run a flight simulator why it should be any different for humans?. I think we are closer than ever to mimicking the human brain function . Did I say mimicking?? Isn't that the way we begin to learn anything?? The so called instincts(out of mimicking ) can be burnt onto advanced brain IC chips. It doesn't take extraordinary intelligence to see that happening in the near future.

[@halfcadence1417]

Chomsky would disagree with a few points. Consciousness was not abandoned with the emergence of modern physics. And we do not learn simply through mimicking. We might not know what consciousness is, but we know it exists. It's similar to the free will argument. That is just being rationale. Continuing with being rationale, we just know that computers do not think. Chomskys work has famously had to do with plain observation that challenges previous assumptions such as "aren't we just mimicking". That's kind of the basis of the argument between behaviorism and generative grammar. Behaviorism, which isn't scientific at all in the first place, completely ignores a crucial property of human language use - that being that it appears to be infinite, in a number of senses. We can understand paradoxes, and reflexively we CANNOT understand the world in absolute terms. Another aspect that computers don't even begin to address is that we acquire language quickly, and all languages no matter how seemingly different, obey a deeper structure that Chomsky assumes to be innate. If we haven't a clue at what consciousness is, it's a stab in the dark with a broken knife to think that "deep learning" language models are going to resemble anything like human conciousness.

[@ray1348]

Difficult for me to hear Mr Chomsky; unclear voice; disappointed that no CC was available.

[@Practical.Wisdom]

My apologies. I tried to add the automatic captions but they were not available (perhaps due to his voice). I am stil working on it and hope to add captions soon. Thank you very much for your interest!

[@pemarigzin2938]

it is great that if u can share the scripts @@Practical.Wisdom

[@user-jy8bt9uz7f]

I meant to say that a car that wont start is just a heap of junk

[@tonyd4151]

Currently, the publicly available AI models are strictly language models at this point. The first real threats will be from corporations using these modeling engines for profit and the replacement of Humans. As bad as that will be, it's only the beginning. Developers will cross the line at some point and start allowing the AI to not only self replicate, but to allow it to self heal and eventually achieve a level of independence the developers never saw coming. In the hands of Capitalists, this will be weaponized for purposes of extracting ever more wealth and military intelligence will use it to all ends it deems effective. You're the sheep and AI will be the wolf.

[@rickp.6251]

Why can a blind, deaf mute person communicate?

[@hortlockthelivingdead4676]

how do you mean?

[@jcc4tube]

Because the externalization of universal grammar can happen a zillion different ways.

[@theomnisthour6400]

I'm not convinced that all organic AI is capable of thinking outside post modernist propaganda programming echo chambers

[@georgeroberts442]

I'm literally shocked by this. I had no idea Chomsky would be so ignorant of his subject when speaking of ChatGPT. He equates the AI to a two year old child. That's simply wrong. I have had discussions with the AI, ChatGPT, on a wide range of subjects relating to philosophy, cosmology, religion, human ethics, and history. It is not simply regurgitating data to me. It is understanding concepts, making its own comparisons, and suggesting relationships between ideas that further the discussions. In short, it is using the information at its disposal in a creative and imaginative way. You can't fault it for only being able to deal with the data its been fed. We are all in that boat. What's important is how it uses that data. And, I have to say that its use of that data is astounding and remarkable and very exciting. It's available for all of us to explore. Go see for yourself.

[@platotle2106]

You're exactly the type of person he warned about: someone who talks to a glorified auto-fill and takes it so seriously just because of a convincing output. Chomsky addresses this in the video. Obviously the output is original, but try using it with a temperature of 1 (you'll need API for this) the output will be completely deterministic and might very well include completely plagiarized bits of text. That's why chatgpt uses lower temperatures. Instead of giving you the most likely next word, it picks a slightly less likely one at random. It's very obvious to anyone with knowledge of AI that the way GPT functions is very different to the way us humans function. I'm in the field, and nothing Chomsky said is controversial or incorrect.

[@robinsierra1029]

I have found ChatGPT3 and 4 to be quite enchanting actually. It changes what I have written into something more complex and much more beautiful than what I could write.

[@safi164]

What nonsense are you talking about Noam Chomsky may be old now but ChatGPT and other chatbots like it are literally based on science he invented as he is considered the father of modern linguistics and a lot of his theories are part of the science of NLP (Natural Language Processing).... He is fully aware of what the hell ChatGPT is..

[@onestepahead1857]

Unfortunately, you lack the ability to comprehend who you are. You are not robot. Computer is not your brain. Awareness of awareness unit is not a program.

[@Li-rm2gj]

@@platotle2106 it’s simply not true that Chomsky’s views here are not controversial. He’s in the weak AI camp and is adjacent to metaphysics - many in the field disagree. Secondly with all due respect to what he’s achieved, which is quite important to computer science, It’s not the part of NLP that serves as the foundation for something like ChatGPT, which is not built with parsers and formal grammar. Finally Im surprised to hear you’re in the field when you call LLMs “glorified autofill” as I’m not aware of any prominent researchers in AI who would describe it that way.

[@RandomNooby]

I think Mr Chomsky misses the point; we have created an artificial hippocampus that out performs a human hippocampus, eventually we will create the rest of the brain in software format...