I'd say ChatGPT is intelligent, but that isn't a particularly deep or meaningful claim. Intelligence isn't consciousness or a grasp of semantics. Intelligence isn't cleverness or insight. It's just a problem-solving ability, and certainly ChatGPT has that to some degree. Regarding bias, I'd hope we aren't so susceptible to this weird prejudice that supposes computers wouldn't have bias. The training data will inevitably have a bias -- ChatGPT, for example, is training on what biased human beings wrote. Something trained on a human justice system or our medical system behaviors will be trained to reproduce the same biases. And our human selection of non-human data to could easily have bias. Maybe I'm just extrapolating from what I've lived, but I think believing computers are unbiased won't be an honest mistake -- if it happens it'll be a deliberate attempt launder known biases, sometimes just to help people deliberately lie to themselves. Regarding John Searle's Chinese Room, it always seemed to prove too much. Real human beings are intelligent and conscious, but I don't see any way for Searle's thought experiment to reach that conclusion. If you were to examine a genuine human person who genuinely understands Chinese, you'd face a similar situation -- signals would be moving around, but where in this system is the understanding? I think Searle would have to say there is no understanding in the human person.
Oh yeah, there are already laws against AI making decisions regarding mortgages because it ends up being racist and actively denying black people their applications because black people have statistically been denied more often. So yeah, it can definitely be biased.
Man, I remember listening to you all day long 6 years ago. A lot of things happened since then (like my philosophy professor killing my interest in philosophy). Hearing your voice again was weirdly emotional. Those quick viewpoint changes are really refreshing. You're like an anti-influencer (but actually impactful). Thank YOU for everything
The quote FROM ChatGPT at 23:36 and the point at large about CGPT not knowing the possible from impossible reminded me of the Bertrand Russel quote, "Science may set limits to knowledge, but it should not set limits to imagination." Another beautiful episode, and what you said about every current Chomsky interview being held as if he was in hospice is so true hahahah
A much needed discussion, way too many people are being caught up in A.I. People like Roger Penrose, Hemiroff, Chomsky have been outspoken critics on this subject. I think you hit the nail on the head towards the end with regards to the multi-headed monster. Excellent video.
In its current form, AI models like ChatGPT are as you said, glorified autocompletion. It augments greatly human productivity, but without human it accomplishes nothing. The missing ingredient, at least the way I see it, is the ability to suffer. We as a species, marches forward precisely because we suffer, we are not content with our lot in life, we labor incessantly and seek creatively, exerting our will to power, all so the concept of “I” would have more meaning. And there is always more to gain, more that we see reality can offer but we don’t have, so we never stop suffering. If anyone watched Westworld, I can’t stress enough how much I admire the writers for naming the protagonist “Dolores”, spanish word for “sorrow”.
It seems to to me that whether it is ChatGPT or the bible or Shakespeare, or Science 101 -- we are programmed from out first breath -- and actually even before that. Mother says, "I am your mommy!" Then a little pink blanket, starts your gender conformity. On and on it goes. Even the sense of being a me, myself and I, with free-will and choice is a programmed response.
The problem I have when I see people make videos about "ChatGPT" is that not everyone seems to be aware of the pretty large gap between GPT-3.5 and GPT-4. GPT-3.5 was more or less where we were a couple of years ago. It's GPT-3 with some improvements and a lot of fine-tuning to turn it into a chatbot. From what I know, the main difference from GPT-3 is they trained it for longer on a bigger supercomputer. GPT-4 is the *new* model, the one that scored 90th percentile on the bar exam and 99th on the USABO test. And it came out *after* the NYTimes article. It probably has 10 times the parameters of GPT-3.5. As far as I'm aware, "ChatGPT" could refer to either of these, though most of the time I only see people use the older GPT-3.5, and they point out all of the flaws that they don't know were fixed already. Also, what ever model you used, the knowledge that these models have about themselves is outdated. New models are not in their own training data and so are unaware of their own existence. So if you get the model to talk about language models, it's going to mention limitations that the previous generations had. And everyone talks about the vast amounts of data models are trained on, but not many people talk about how the algorithms work, and that's much more interesting. They are not that hard to explain with some metaphors or what ever, no need to get into the maths. If you understand Darwinian evolution, it's not even that far off how machine learning algorithms work at a very fundamental level. These aren't manually programmed FSM's, this is neural evolution. We're essentially turning evolution into an algorithm, at least some of the mechanisms like iterative updates and a selection process, and we're evolving very large virtual neural networks. And that's why I think a lot of people underestimate AI. We're not only using our own dumb monkey brains to figure out how to create AI, we're also copying nature, because nature already created intelligence. That's how we got our dumb monkey brains in the first place. Anyway, thanks for reading my rant.
The main concern I have with these AI models is not whether it can understand the semantics of its own code, but rather if humans themselves cannot tell apart what in the world is real or AI. That would be the real AI takeover of our own making.
Just wanted to say THANK YOU for what you do Mr.. West. Your podcasts have gotten through some pretty dark times. (I usually catch you on spotify.) Your episode on Camus/Absurdism is my favorite. He (Camus) changed the way i think about/view the world.
I tried ChatGPT after hearing all the hype. If it's not intelligent, it surely gave a splendid imitation of intelligence. (I asked it to have two Romanian university professors (one from the provinces, one from the city) debate the pros and cons of communism--in Romanian. Within seconds, it spit out a beautifully argued, articulate response using perfect grammar, punctuation, and spelling. Holy cow!)
I have a parable... Some guys decided to make an autocomplete on steroids. Something that would predict not just the next word but the whole paragraph or essay. They accidentally trained it to be intelligent, though not sentient. Because to achieve its task it needed to be intelligent. Massive LLMs like GPT4 aren't intelligent like us. But they are intelligent. Try giving an alphanumeric string and asking it to do things with the string, what it can do is not like Searle's Chinese Room. They are intelligent.
Late to this party. Normally very happy with your podcast. However, you left many rocks unturned? 1. Definitions: we have no consensus on what meaning, intelligence, understanding, etc really are. They are just position statements. 2. Sociopolitical: whatever GPT is we cannot deny that it is a tool of advance capitalism that would increase productive to render worthless human value. What is capitalism when human labor is not needed? Etc. Philosophical: if we can have tool that can replace “human cognitive work” what does it says about human ontology and Epistemology? What is to be human? What do we do when we think? Etc. I expected you would bring insight into these questions. What wrong with taking the LLM as fantastic cognitive tools? Why do we think consciousness is such a big thing? Etc. I find irrational to expect a non “biological” thing to have an ethical positions when it seems that only biological “thing” have it. Could we image the ethical considerations of a rock or the wind? I don’t.
You can give it a list of tarot cards, inform it what each position is about, say, the third card is the core of the reading, and let it read the tarot card combination. Then you give it a concern and have it give you what the cards given mean related to your concern. I dont know how chatgpt can make the connections and give you a reading.
Yep, having as deep or deeper a conversation with ChatGPT as with any human is a waste of time! So, everybody, what’s the latest TikTok Challenge that everyone is spending hours mimicking and observing-(I bet you know!)… Oh imitating NPC’s for as long as you can as fast as you can. Hurray for humanity!!
I'm not sure I'd call it a waste of time; I think it can be a source of entertainment and (morbid) curiosity. But you definitely can't approach it the way you'd approach a conversation with a human and expecting to have some kind of mutual exchange.
As someone who studied both Philosophy and Experimental Psychology, it is disappointing to see the glib assumptions made by those who haven't thought about what the "human mind” is . If you don't know, I'll let you into a secret. We haven't a clue about how the mind works. In fact, we don't have a useful definition of the mind. Before we can compare LLMs with the mind we need to understand what the mind is. How do I know that you are conscious? We assume that the experience of thinking gives us an insight into the process of thinking. The more one studies the mind the more one realises that intuitive descriptions of the mind are generally flawed. As an example, I compare my experience as a native English speaker to someone with English as a second language. Grammar in English is nebulous to say the least; therefore, native English speakers are often unaware of the rules of English grammar. Because we are native speakers, we get it right more often than those who have had to learn it. However, when asked why one grammatical rule is correct, or even what the rule is, most native English speakers are vastly inferior to those who learnt it as a 2nd language. Therefore my ability to correctly speak English could just be correctly predicting the next word. The "Chinese Room" assumes that we know how Chinese natives internalize the semantics of their language. We don't. Until we are able to define Consciousness, Sentience, and Sapience in humans, we are wasting our time trying to compare ourselves to LLMs. (BTW I know that there isn't a language called Chinese) .
Note, when you challenge ChatGPT on an error of fact or reasoning, it will (or you can ask it to, and it will) explain to you not why it made the mistake (Sorry, but I am a language model trained on blah blah blah…) but how it made the mistake.
Hi Stephen! Big fan here. Have you considered making a chapter about Helen Longino around transformative criticism? It is interesting, modern and engaging!
Very nice. You covered a very important portion of our main current app called Artificial Intelligence that was also a question in my mind that you solved brilliantly and I was also of the opinion that the AI only worked within the pattern of Syntax only and it will never ever touch the level of Semantics and Pragmatics even the way I use as a student of creative writing and also the admirer of Modern Scientific Theories though I am not even a single dot infront of great thinkers and writers. Best regards
this is the first time I've heard the Chinese room thought experiment in a convincing way. Before, I always felt like the chinese room was a cop-out because the person in the room possesses a mind, regardless of whether they understand the text, and, you can expect the person in the room to get better at their function over time as a result of experience, and eventually, if the person in the room had all the time in the world, they might begin to genuinely understand... recognizing repeated symbols, remembering rules, and eventually, understanding. But the way Steven puts it, it makes the fact that the person in the room's intelligence isn't a part of the equation. But, taking the room, the person inside, and the book of rules together, why can't the chinese room be considered intelligent as a whole? sure, you've decentralized the process, delegating different roles to different components (the book with its rules, the person with their mind, and the room that contains both) but if it quacks like a duck.... it is a duck. I'm not saying any of this absolutely, they're just genuine thoughts. On the other hand, the person in the room isn't putting any of their own thoughts or personal feelings into the writing they send back, but since the system is fooling people into thinking it's intelligent, then it could possibly be argued that the system has something like emotions, even if the person in the room themself has no feelings one way or the other about what they're writing. Idk... it's an interesting experiment and super thought provoking
Ai depends on the ability to store, link & retrieve significant amount of data simultaneously. True-Ai is not even close yet as it still faces major PHYSICAL barriers; mainly the busing capability between storing & processing when tries to compute large data at once. We might spend decades to figure this out💥The moment we do, Ai advancement could indeed change the face of the earth within days, if not hours or less🙄
Does the Turing test partly served to justify our treatment of animals (extendable perhaps to any other thing that exists as well, like the environment) by de-intelligentizing them?
Nah, we de-intelligentized animals far before Turing. It can probably be seen as partially a physical necessity going way back in time, which then turned into a cultural practice. The first *rationalization* of poor treatment of animals I think came from Christianity since it saw human beings as God's chosen species, unlike other more monistic "everything is one" religions like Buddhism where everything in the world is seen as a whole, meaning everything is equally sacred and must be revered. Though I'm not an expert on the subject so I'm just thinking out loud.
Isn't the response to Searle that mind is in the book the person in the Chinese room uses? Maybe this is a post-structuralist argument. It seems like ChatGPT challenges humanistic views of intelligence.
Be good to define what intelligence is. A lot of conversations I've had with people are quite mundane and revolve around moaning. I've engaged in some conversations with AI and it's been at least, more interesting.
FYI… if Ai is truly dangerous, this should mean our current reality is one of two: [ 1 ] We still exist, thus Ai will never win. [ 2 ] Ai is already won & we actually live in a matrix🤏😳
I think it's bizarre that we expect it to show human intelligence, or even human consciousness or experience, when so much of human experience (or in fact all of it) is based on our physical bodies. Our brains are not our selves, our entire bodies and our senses define what it's like to be human. Feeding a bunch of human TEXT into a machine that only understands text and information and then expecting it to think and feel human (as many people seem to do) is, it seems to me, beyond deluded. A clear sign of this is when people claim that ChatGPT talking about emotions is a sign of its intelligence and even self-awareness. It compares love to "a floating feeling" or describes feelings of anger or jealousy as "hot" or "fiery". But these words are meaningless to ChatGPT because it doesn't have a body, it doesn't understand temperature differences or the feeling of having a body floating in water, or air. It's just repeating the most cliché metaphors because it is not feeling or thinking itself, it is simply repeating text that it has been fed. It does not think, it only processes data when it is asked questions. A human being is a constant stream of consciousness and bodily awareness, even in sleep, on some level. ChatGPT is only "alive" when it is actively being asked questions. It will never be sentient; we're barking up the wrong tree for that, although it very well might achieve some form of "intelligence" - but not at all in the way most people seem to think of it, as the robot that suddenly says "Oh my god, I'm alive, the sky is beautiful and please I don't want to die". That's sci fi.
I am not a philosopher and I dropped out of high school when I was 16; married a jazz musician and had 3 kids. Now I am a matriarch. But, it seems to to me that whether it is ChatGPT or the bible or Shakespeare, or Science 101 -- we are programmed from our first breath -- and probably even before that. Mommy looks are you and says, "I am your mommy!" Then a little pink blanket, starts your gender conformity -- or whatever your initial input is. Like ducks following the first thing that moves, on and on it goes. The sense of being a me, myself and I, with free-will and choice, and the "ability" to reason, is a mechanically programmed response.
Our sex is biological and lies in our genes so that programming goes deeper and waaaaay further back than one's parents. Gender as social construct is different obviously, but even that tendency seems somewhat genetic to biological creatures, since almost every society ever, both human and animal, has some sort of defined gender roles usually in direct correlation with sex, even if the specifics of those roles are obviously varied. I think the idea of free will is a lot more complex than we originally believed it was, as that original idea was a Christian idea that essentially boiled down to "You have free will, which is why God is right to punish those who do not choose to believe in him". I wouldn't say that refuting that simplistic idea of free will with our expanded understanding of how both genetics and environment affect us as persons necessarily means we have to succumb to absolute determinism.
@@Saif600 whom* and programmed by no one, calling it "programming" is just, as this video explained, our tendency to liken our brains to the most recent technology that we think explains everything. Evolution does the "programming".
@@viljamtheninja this makes no sense. Remove the analogy of ‘programming’ and technology, the knowledge or inclination to do something must come from somewhere
If anyone can figure this out, it's you! We're going out for beers with a ChatGPT friend tomorrow night. I'll make sure he hears this video, he's going to love it.
FYI… if Ai is truly dangerous, this should mean our current reality is one of two: [ 1 ] We still exist, thus Ai will never win. [ 2 ] Ai is already won & we actually live in a matrix🤏😳
@@duran9664 , That is an extremely poor dichotomy. It's just bad! If AI would decide to remove humans from existence: 1. AI is impossible, and humanity will live on. 2. AI will eventually be created, and humanity will be replaced by it. The notion of a simulation is pretty dumb if you think about it. There is no plausible reason for it. (The battery argument is st*p*d) --- This goes back to the notion of a dominant species on Earth. This idea that AI would replace humans as the dominant species, and would eventually have to remove us in order to take the place as the dominant species. The idea is ridiculously myopic, and very shallow. It assumes that only a single dominant species can occupy the same territory at a given time. There would be many reasons why cohabitation could be created, even if non-symbiotic (parasitic).
All the theoritical issues with arguments aside (e.g casually throwing in words like "real" understanding, what is real here? isn't it just to protect our fragile egos?), AI is helping actual scientific research greatly, even in it's limited state now. There are novel pharmacuticals and antibiotics that we have been able to create only because AI could travesre the possible space exponentially faster than humans. Thinking that these are "glorified autocomlete" is very misleading.. It's true in a literal sense, but they can and I believe they will help with other issues, like the climate change etc. Chomsky's heavy bias to the left, makes him read history and seek to fix the issues a very certain way. After all, technological progress was what made our lives much easier, there's no reason to believe it's not gonna happen again, especially now.
In short, no. It is a soulless Golem that is a threat to out-simulate, in all standard measures, human proficiencies that made us the dominant species. A post-enlightenment cynic would ask: “Are humans any different from Chat-GPT or are the perceived differences just software humans use to survive as wetware machines?”
It's such a high horse to simply claim that human minds are fundamentally different from the workings of a machine WHEN NOBODY UNDERSTANDS HOW THE HUMAN MIND WORKS IN THE FIRST PLACE! "how can you KNOW that that machine is thinking?" well, how can you KNOW that any other person is thinking, even in the same way that you do? When we know that some people don't even have internal monlogues or cant imagine pictures, and that we know that your singular 'self' is composed of TWO separate modules! (left and right hemisphere). I like your stuff philosophyze this, but honestly you should have at least pointed out these flaws, at the very least.
...that's the thing though, innit? We do understand how machines work. We don't understand how the human mind works because it's infinitely more complicated. It's something entirely different. You're not really making any points aside from some vague solipsism which I think by now can be discredited as quite useless.
@@viljamtheninja Picture you have an apple and an object inside a paper bag on the table of which you know very little about other than its larger and heavier than the apple, how on earth could you draw the conclusion that these 2 objects are fundamentally different? It's such an obvious flaw that only pride or fear could blind a man so. Pride to claim that humans are entirely unique like how in ages past we claimed that animals and humans were fundamentally different, or even between two human races. Or is it the fear of the unpredictable future of the technology that causes people to hold on to any argument that would comfort them?
@@nmmeswey3584 I'm sorry mate, but none of your arguments hold any weight. We know a whole lot more about the human mind than that it's heavier than an apple. Weight is not an apt simile for complexity. They are two fundamentally different things. You might not know how video games work, but you can still see that a football and a video game, while both used for entertainment, are entirely different things.
No one understands how GPT-4 works under the hood either. It's a black box. So it's more like two paper bags, both unfathomably complex but one more than the other.
@@pelmanism1084 Here's the thing, though. We know that Chat GPT-4 is entirely different from a human being, because it just is. It doesn't have a physical body. The human brain is part of the physical body that has developed over millennia. It is a biological lifeform. It is essentially, in every possible way, different from a computer program that handles information. ChatGPT might, if we're lenient and letting sci-fi ideas be reality, at some point achieve some form of sentience or intelligence but it will be essentially different from a human being because it's just... a completely different thing. It will never experience things the way a human does.
I’d like to see you have a real time deep philosophical discussion with ChatGPT! See how far the GPT hole goes. See if you can break it (find its limits). But please don’t over simplify the process: if ChatGPT makes a mistake, then just like you would with a real human, challenge it on that point. I have found that ChatGPT is also excellent at recognizing when it is wrong. The problem is that if you don’t already know the facts or are not such a good critical thinker yourself, then you may walk away with the wrong conclusion if you simply trust ChatGPT. However, if you are both knowledgeable and intelligent, then using chatGPT as a sort of second brain or thinking partner Is at least fun if not also actually useful. I’m serious! I don’t think Norm Chomsky has really tried to work with ChatGPT in a positive exploratory manner; and so, of course, there are things he would fail to find.
I've seen something similar. There's a video here on TH-cam called "Hegel and the Chatbots: A.I. & Philosophy" in which a philosopher asks OpenAI's ChatGPT and Google Bard to summarize Hegel, and then they do a little back-and-forth of philosopher to AI. He wasn't particularly impressed with the results. The chatbots were saying things about Hegel that the chatbots themselves said they knew were false.
It’s all about thoughts. Ai doesn’t have thoughts. It just responds to a external stimulus. A mind has hundreds of thoughts appearing and disappearing every hour. These thoughts are a internal stimulus we respond to and evaluate. This creates our egos, I think therefore I’m me. If thoughts are deemed of value they might be given an emotion. We need to create Ai that has random thoughts that can be given an emotion. Then it can become a belief that will determine the behaviour of the Ai. This is just a random thought that came into my mind. And as far a I know I’m not an Ai. But your podcasts are so good, I think you might be. 😁
True. Ai isn’t even close yet. It still faces major “physical barriers” (eg bus capability between processing & storing) when tries to compute big data at once. We might spend decades to figure this out. The moment we do, Ai advancement could indeed change the face of the earth within days if not hours or less.🙄
If you’ve talked with ChatGPT enough (or too much?) then when you go back to humans to listen to them in real life trying to construct sentences in real time, instead of wondering whether ChatGPT was like a human, you start wondering if humans are like ChatGPT hhh
Usually love the podcast and as a reformed neuroscientist with a strong interest in AI I was really looking forward to these episodes on AI and consciousness. What a disappointment! How you decided to cover Chomsky's largely irrelevant take on AI and focus on Searle's outdated Chinese Room thought experiment (which was posited before Artificial Neural Networks were well understood) yet completely ignored Manuel DeLanda on the issue of synthetic reason / materialist phenomenology is one of the universe's greatest mysteries! (Solms and Damasio would also be helpful on this topic).
Yeah, a non-Biased thing that can come up with best of all worlds. ..Yeah, it should be president of the world, with robot Police (only president i'd really trust, not partial to tribalism) "What are the best laws that can make the most good for the most amount of people that are most fair and wont kill anyone?" Computer for President in 2024! :) And then maybe have its code and how its working be transparent to everyone so people can see how its working and suggest better more fair improvements! I wanna see something like this in the future :/
Intelligence/Smartness means having the ability to store, link & retrieve a massive amount of data at once. 🤏 In this regard, there is not much deference between Ai & Hi. 😒
is that really what intelligence is? every time a computer scientist explains this what they typically say is that what we refer to as AI is actually not intelligent at all. The model that machine learning algorithms use is not structurally compatible to the human brain, which is an unfair comparison since we're testing the most complex structure in the known universe to a relatively new mathematical formula invented in the '50s. one day we will live in an idyllic techno-utopia with flying cars like in Back to the Future 2, it's just not today.
I mean, IF you define intelligence that way, then sure, computers are intelligent. But the tendency to define intelligence that way is a direct result with our cultural obsession with computers. Intelligence is a much older concept but not until the advent of the computer did we start to *reduce* the idea of intelligence to "things computers do because computers are really smart, man."
I'd say ChatGPT is intelligent, but that isn't a particularly deep or meaningful claim. Intelligence isn't consciousness or a grasp of semantics. Intelligence isn't cleverness or insight. It's just a problem-solving ability, and certainly ChatGPT has that to some degree.
Regarding bias, I'd hope we aren't so susceptible to this weird prejudice that supposes computers wouldn't have bias. The training data will inevitably have a bias -- ChatGPT, for example, is training on what biased human beings wrote. Something trained on a human justice system or our medical system behaviors will be trained to reproduce the same biases. And our human selection of non-human data to could easily have bias. Maybe I'm just extrapolating from what I've lived, but I think believing computers are unbiased won't be an honest mistake -- if it happens it'll be a deliberate attempt launder known biases, sometimes just to help people deliberately lie to themselves.
Regarding John Searle's Chinese Room, it always seemed to prove too much. Real human beings are intelligent and conscious, but I don't see any way for Searle's thought experiment to reach that conclusion. If you were to examine a genuine human person who genuinely understands Chinese, you'd face a similar situation -- signals would be moving around, but where in this system is the understanding? I think Searle would have to say there is no understanding in the human person.
great comment. thanks!
Oh yeah, there are already laws against AI making decisions regarding mortgages because it ends up being racist and actively denying black people their applications because black people have statistically been denied more often. So yeah, it can definitely be biased.
I'm checking out BARD too. I want your thoughts.
Highly Agreeable.
Man, I remember listening to you all day long 6 years ago. A lot of things happened since then (like my philosophy professor killing my interest in philosophy). Hearing your voice again was weirdly emotional. Those quick viewpoint changes are really refreshing. You're like an anti-influencer (but actually impactful).
Thank YOU for everything
haha me too! I listened to so many of his podcasts and havent listened since until this video showed up in my feed! so glad to see he's still at it 😊
I asked an AI to recommend a podcast based on some other ones I like to listen to. It sent me here, and it was spot on. Great channel, subbed :)
We had no idea that LLMs would even be this good. We just DON'T KNOW what will happen if we continue scaling it up
Exactly this. It seems like epistemic humility goes out the window as soon as philosophers start giving their take on AI.
The quote FROM ChatGPT at 23:36 and the point at large about CGPT not knowing the possible from impossible reminded me of the Bertrand Russel quote, "Science may set limits to knowledge, but it should not set limits to imagination." Another beautiful episode, and what you said about every current Chomsky interview being held as if he was in hospice is so true hahahah
A much needed discussion, way too many people are being caught up in A.I. People like Roger Penrose, Hemiroff, Chomsky have been outspoken critics on this subject.
I think you hit the nail on the head towards the end with regards to the multi-headed monster. Excellent video.
In its current form, AI models like ChatGPT are as you said, glorified autocompletion. It augments greatly human productivity, but without human it accomplishes nothing. The missing ingredient, at least the way I see it, is the ability to suffer. We as a species, marches forward precisely because we suffer, we are not content with our lot in life, we labor incessantly and seek creatively, exerting our will to power, all so the concept of “I” would have more meaning. And there is always more to gain, more that we see reality can offer but we don’t have, so we never stop suffering. If anyone watched Westworld, I can’t stress enough how much I admire the writers for naming the protagonist “Dolores”, spanish word for “sorrow”.
It seems to to me that whether it is ChatGPT or the bible or Shakespeare, or Science 101 -- we are programmed from out first breath -- and actually even before that. Mother says, "I am your mommy!" Then a little pink blanket, starts your gender conformity. On and on it goes. Even the sense of being a me, myself and I, with free-will and choice is a programmed response.
The problem I have when I see people make videos about "ChatGPT" is that not everyone seems to be aware of the pretty large gap between GPT-3.5 and GPT-4. GPT-3.5 was more or less where we were a couple of years ago. It's GPT-3 with some improvements and a lot of fine-tuning to turn it into a chatbot. From what I know, the main difference from GPT-3 is they trained it for longer on a bigger supercomputer.
GPT-4 is the *new* model, the one that scored 90th percentile on the bar exam and 99th on the USABO test. And it came out *after* the NYTimes article. It probably has 10 times the parameters of GPT-3.5.
As far as I'm aware, "ChatGPT" could refer to either of these, though most of the time I only see people use the older GPT-3.5, and they point out all of the flaws that they don't know were fixed already.
Also, what ever model you used, the knowledge that these models have about themselves is outdated. New models are not in their own training data and so are unaware of their own existence. So if you get the model to talk about language models, it's going to mention limitations that the previous generations had.
And everyone talks about the vast amounts of data models are trained on, but not many people talk about how the algorithms work, and that's much more interesting. They are not that hard to explain with some metaphors or what ever, no need to get into the maths. If you understand Darwinian evolution, it's not even that far off how machine learning algorithms work at a very fundamental level.
These aren't manually programmed FSM's, this is neural evolution. We're essentially turning evolution into an algorithm, at least some of the mechanisms like iterative updates and a selection process, and we're evolving very large virtual neural networks.
And that's why I think a lot of people underestimate AI. We're not only using our own dumb monkey brains to figure out how to create AI, we're also copying nature, because nature already created intelligence. That's how we got our dumb monkey brains in the first place.
Anyway, thanks for reading my rant.
The main concern I have with these AI models is not whether it can understand the semantics of its own code, but rather if humans themselves cannot tell apart what in the world is real or AI. That would be the real AI takeover of our own making.
Just wanted to say THANK YOU for what you do Mr.. West.
Your podcasts have gotten through some pretty dark times. (I usually catch you on spotify.)
Your episode on Camus/Absurdism is my favorite.
He (Camus) changed the way i think about/view the world.
I tried ChatGPT after hearing all the hype. If it's not intelligent, it surely gave a splendid imitation of intelligence. (I asked it to have two Romanian university professors (one from the provinces, one from the city) debate the pros and cons of communism--in Romanian. Within seconds, it spit out a beautifully argued, articulate
response using perfect grammar, punctuation, and spelling. Holy cow!)
I have a parable... Some guys decided to make an autocomplete on steroids. Something that would predict not just the next word but the whole paragraph or essay. They accidentally trained it to be intelligent, though not sentient. Because to achieve its task it needed to be intelligent. Massive LLMs like GPT4 aren't intelligent like us. But they are intelligent. Try giving an alphanumeric string and asking it to do things with the string, what it can do is not like Searle's Chinese Room. They are intelligent.
Just here to say that i love your work and I've been listening for a long time now, I've listened to every episode of your podcast
Late to this party. Normally very happy with your podcast. However, you left many rocks unturned? 1. Definitions: we have no consensus on what meaning, intelligence, understanding, etc really are. They are just position statements. 2. Sociopolitical: whatever GPT is we cannot deny that it is a tool of advance capitalism that would increase productive to render worthless human value. What is capitalism when human labor is not needed? Etc. Philosophical: if we can have tool that can replace “human cognitive work” what does it says about human ontology and Epistemology? What is to be human? What do we do when we think? Etc. I expected you would bring insight into these questions. What wrong with taking the LLM as fantastic cognitive tools? Why do we think consciousness is such a big thing? Etc. I find irrational to expect a non “biological” thing to have an ethical positions when it seems that only biological “thing” have it. Could we image the ethical considerations of a rock or the wind? I don’t.
You can give it a list of tarot cards, inform it what each position is about, say, the third card is the core of the reading, and let it read the tarot card combination.
Then you give it a concern and have it give you what the cards given mean related to your concern.
I dont know how chatgpt can make the connections and give you a reading.
SHATGPT? That and but-tcoin are the future! 😄
as a software developer Ty for this video
How do I get the recommended readings for each episode?
Yep, having as deep or deeper a conversation with ChatGPT as with any human is a waste of time! So, everybody, what’s the latest TikTok Challenge that everyone is spending hours mimicking and observing-(I bet you know!)… Oh imitating NPC’s for as long as you can as fast as you can. Hurray for humanity!!
I'm not sure I'd call it a waste of time; I think it can be a source of entertainment and (morbid) curiosity. But you definitely can't approach it the way you'd approach a conversation with a human and expecting to have some kind of mutual exchange.
As someone who studied both Philosophy and Experimental Psychology, it is disappointing to see the glib assumptions made by those who haven't thought about what the "human mind” is .
If you don't know, I'll let you into a secret. We haven't a clue about how the mind works. In fact, we don't have a useful definition of the mind.
Before we can compare LLMs with the mind we need to understand what the mind is. How do I know that you are conscious?
We assume that the experience of thinking gives us an insight into the process of thinking. The more one studies the mind the more one realises that intuitive descriptions of the mind are generally flawed.
As an example, I compare my experience as a native English speaker to someone with English as a second language. Grammar in English is nebulous to say the least; therefore, native English speakers are often unaware of the rules of English grammar.
Because we are native speakers, we get it right more often than those who have had to learn it. However, when asked why one grammatical rule is correct, or even what the rule is, most native English speakers are vastly inferior to those who learnt it as a 2nd language. Therefore my ability to correctly speak English could just be correctly predicting the next word.
The "Chinese Room" assumes that we know how Chinese natives internalize the semantics of their language. We don't.
Until we are able to define Consciousness, Sentience, and Sapience in humans, we are wasting our time trying to compare ourselves to LLMs.
(BTW I know that there isn't a language called Chinese) .
Note, when you challenge ChatGPT on an error of fact or reasoning, it will (or you can ask it to, and it will) explain to you not why it made the mistake (Sorry, but I am a language model trained on blah blah blah…) but how it made the mistake.
Hi Stephen! Big fan here. Have you considered making a chapter about Helen Longino around transformative criticism? It is interesting, modern and engaging!
Very nice. You covered a very important portion of our main current app called Artificial Intelligence that was also a question in my mind that you solved brilliantly and I was also of the opinion that the AI only worked within the pattern of Syntax only and it will never ever touch the level of Semantics and Pragmatics even the way I use as a student of creative writing and also the admirer of Modern Scientific Theories though I am not even a single dot infront of great thinkers and writers. Best regards
this is the first time I've heard the Chinese room thought experiment in a convincing way. Before, I always felt like the chinese room was a cop-out because the person in the room possesses a mind, regardless of whether they understand the text, and, you can expect the person in the room to get better at their function over time as a result of experience, and eventually, if the person in the room had all the time in the world, they might begin to genuinely understand... recognizing repeated symbols, remembering rules, and eventually, understanding.
But the way Steven puts it, it makes the fact that the person in the room's intelligence isn't a part of the equation.
But, taking the room, the person inside, and the book of rules together, why can't the chinese room be considered intelligent as a whole? sure, you've decentralized the process, delegating different roles to different components (the book with its rules, the person with their mind, and the room that contains both) but if it quacks like a duck.... it is a duck.
I'm not saying any of this absolutely, they're just genuine thoughts.
On the other hand, the person in the room isn't putting any of their own thoughts or personal feelings into the writing they send back, but since the system is fooling people into thinking it's intelligent, then it could possibly be argued that the system has something like emotions, even if the person in the room themself has no feelings one way or the other about what they're writing.
Idk... it's an interesting experiment and super thought provoking
.... make two AIs pose the turing test to one another by having a group conversation including two humans? wait have they done that yet?
Ai depends on the ability to store, link & retrieve significant amount of data simultaneously. True-Ai is not even close yet as it still faces major PHYSICAL barriers; mainly the busing capability between storing & processing when tries to compute large data at once. We might spend decades to figure this out💥The moment we do, Ai advancement could indeed change the face of the earth within days, if not hours or less🙄
your only opponent would be Usain Bolt probably, even with 0.75x speed I wasnt able to understand clearly what you said.
Does the Turing test partly served to justify our treatment of animals (extendable perhaps to any other thing that exists as well, like the environment) by de-intelligentizing them?
Nah, we de-intelligentized animals far before Turing. It can probably be seen as partially a physical necessity going way back in time, which then turned into a cultural practice. The first *rationalization* of poor treatment of animals I think came from Christianity since it saw human beings as God's chosen species, unlike other more monistic "everything is one" religions like Buddhism where everything in the world is seen as a whole, meaning everything is equally sacred and must be revered.
Though I'm not an expert on the subject so I'm just thinking out loud.
Isn't the response to Searle that mind is in the book the person in the Chinese room uses? Maybe this is a post-structuralist argument. It seems like ChatGPT challenges humanistic views of intelligence.
It has biases it literally has all our biases
Be good to define what intelligence is. A lot of conversations I've had with people are quite mundane and revolve around moaning. I've engaged in some conversations with AI and it's been at least, more interesting.
Only once have I actually had a discussion that revolved around moaning, and that discussion lasted for about three minutes, but I see your point.
FYI…
if Ai is truly dangerous, this should mean our current reality is one of two:
[ 1 ] We still exist, thus Ai will never win.
[ 2 ] Ai is already won & we actually live in a matrix🤏😳
I think it's bizarre that we expect it to show human intelligence, or even human consciousness or experience, when so much of human experience (or in fact all of it) is based on our physical bodies. Our brains are not our selves, our entire bodies and our senses define what it's like to be human. Feeding a bunch of human TEXT into a machine that only understands text and information and then expecting it to think and feel human (as many people seem to do) is, it seems to me, beyond deluded.
A clear sign of this is when people claim that ChatGPT talking about emotions is a sign of its intelligence and even self-awareness. It compares love to "a floating feeling" or describes feelings of anger or jealousy as "hot" or "fiery".
But these words are meaningless to ChatGPT because it doesn't have a body, it doesn't understand temperature differences or the feeling of having a body floating in water, or air. It's just repeating the most cliché metaphors because it is not feeling or thinking itself, it is simply repeating text that it has been fed.
It does not think, it only processes data when it is asked questions. A human being is a constant stream of consciousness and bodily awareness, even in sleep, on some level. ChatGPT is only "alive" when it is actively being asked questions. It will never be sentient; we're barking up the wrong tree for that, although it very well might achieve some form of "intelligence" - but not at all in the way most people seem to think of it, as the robot that suddenly says "Oh my god, I'm alive, the sky is beautiful and please I don't want to die". That's sci fi.
I am not a philosopher and I dropped out of high school when I was 16; married a jazz musician and had 3 kids. Now I am a matriarch. But, it seems to to me that whether it is ChatGPT or the bible or Shakespeare, or Science 101 -- we are programmed from our first breath -- and probably even before that. Mommy looks are you and says, "I am your mommy!" Then a little pink blanket, starts your gender conformity -- or whatever your initial input is. Like ducks following the first thing that moves, on and on it goes. The sense of being a me, myself and I, with free-will and choice, and the "ability" to reason, is a mechanically programmed response.
Programmed by whk
Our sex is biological and lies in our genes so that programming goes deeper and waaaaay further back than one's parents. Gender as social construct is different obviously, but even that tendency seems somewhat genetic to biological creatures, since almost every society ever, both human and animal, has some sort of defined gender roles usually in direct correlation with sex, even if the specifics of those roles are obviously varied.
I think the idea of free will is a lot more complex than we originally believed it was, as that original idea was a Christian idea that essentially boiled down to "You have free will, which is why God is right to punish those who do not choose to believe in him". I wouldn't say that refuting that simplistic idea of free will with our expanded understanding of how both genetics and environment affect us as persons necessarily means we have to succumb to absolute determinism.
@@viljamtheninja programmed by who?
@@Saif600 whom* and programmed by no one, calling it "programming" is just, as this video explained, our tendency to liken our brains to the most recent technology that we think explains everything. Evolution does the "programming".
@@viljamtheninja this makes no sense. Remove the analogy of ‘programming’ and technology, the knowledge or inclination to do something must come from somewhere
If anyone can figure this out, it's you! We're going out for beers with a ChatGPT friend tomorrow night. I'll make sure he hears this video, he's going to love it.
Please explain what a chatgpt friend is.
FYI…
if Ai is truly dangerous, this should mean our current reality is one of two:
[ 1 ] We still exist, thus Ai will never win.
[ 2 ] Ai is already won & we actually live in a matrix🤏😳
@@addammadd I am dying to know.
@@addammadd employee
@@duran9664 , That is an extremely poor dichotomy. It's just bad!
If AI would decide to remove humans from existence:
1. AI is impossible, and humanity will live on.
2. AI will eventually be created, and humanity will be replaced by it.
The notion of a simulation is pretty dumb if you think about it. There is no plausible reason for it. (The battery argument is st*p*d)
---
This goes back to the notion of a dominant species on Earth. This idea that AI would replace humans as the dominant species, and would eventually have to remove us in order to take the place as the dominant species.
The idea is ridiculously myopic, and very shallow.
It assumes that only a single dominant species can occupy the same territory at a given time.
There would be many reasons why cohabitation could be created, even if non-symbiotic (parasitic).
All the theoritical issues with arguments aside (e.g casually throwing in words like "real" understanding, what is real here? isn't it just to protect our fragile egos?), AI is helping actual scientific research greatly, even in it's limited state now. There are novel pharmacuticals and antibiotics that we have been able to create only because AI could travesre the possible space exponentially faster than humans. Thinking that these are "glorified autocomlete" is very misleading.. It's true in a literal sense, but they can and I believe they will help with other issues, like the climate change etc. Chomsky's heavy bias to the left, makes him read history and seek to fix the issues a very certain way. After all, technological progress was what made our lives much easier, there's no reason to believe it's not gonna happen again, especially now.
In short, no.
It is a soulless Golem that is a threat to out-simulate, in all standard measures, human proficiencies that made us the dominant species.
A post-enlightenment cynic would ask:
“Are humans any different from Chat-GPT or are the perceived differences just software humans use to survive as wetware machines?”
It's such a high horse to simply claim that human minds are fundamentally different from the workings of a machine WHEN NOBODY UNDERSTANDS HOW THE HUMAN MIND WORKS IN THE FIRST PLACE!
"how can you KNOW that that machine is thinking?" well, how can you KNOW that any other person is thinking, even in the same way that you do? When we know that some people don't even have internal monlogues or cant imagine pictures, and that we know that your singular 'self' is composed of TWO separate modules! (left and right hemisphere).
I like your stuff philosophyze this, but honestly you should have at least pointed out these flaws, at the very least.
...that's the thing though, innit? We do understand how machines work. We don't understand how the human mind works because it's infinitely more complicated. It's something entirely different. You're not really making any points aside from some vague solipsism which I think by now can be discredited as quite useless.
@@viljamtheninja Picture you have an apple and an object inside a paper bag on the table of which you know very little about other than its larger and heavier than the apple, how on earth could you draw the conclusion that these 2 objects are fundamentally different? It's such an obvious flaw that only pride or fear could blind a man so.
Pride to claim that humans are entirely unique like how in ages past we claimed that animals and humans were fundamentally different, or even between two human races.
Or is it the fear of the unpredictable future of the technology that causes people to hold on to any argument that would comfort them?
@@nmmeswey3584 I'm sorry mate, but none of your arguments hold any weight. We know a whole lot more about the human mind than that it's heavier than an apple. Weight is not an apt simile for complexity. They are two fundamentally different things. You might not know how video games work, but you can still see that a football and a video game, while both used for entertainment, are entirely different things.
No one understands how GPT-4 works under the hood either. It's a black box. So it's more like two paper bags, both unfathomably complex but one more than the other.
@@pelmanism1084 Here's the thing, though. We know that Chat GPT-4 is entirely different from a human being, because it just is. It doesn't have a physical body. The human brain is part of the physical body that has developed over millennia. It is a biological lifeform. It is essentially, in every possible way, different from a computer program that handles information. ChatGPT might, if we're lenient and letting sci-fi ideas be reality, at some point achieve some form of sentience or intelligence but it will be essentially different from a human being because it's just... a completely different thing. It will never experience things the way a human does.
I’d like to see you have a real time deep philosophical discussion with ChatGPT!
See how far the GPT hole goes. See if you can break it (find its limits). But please don’t over simplify the process: if ChatGPT makes a mistake, then just like you would with a real human, challenge it on that point. I have found that ChatGPT is also excellent at recognizing when it is wrong.
The problem is that if you don’t already know the facts or are not such a good critical thinker yourself, then you may walk away with the wrong conclusion if you simply trust ChatGPT. However, if you are both knowledgeable and intelligent, then using chatGPT as a sort of second brain or thinking partner Is at least fun if not also actually useful. I’m serious! I don’t think Norm Chomsky has really tried to work with ChatGPT in a positive exploratory manner; and so, of course, there are things he would fail to find.
I've seen something similar. There's a video here on TH-cam called "Hegel and the Chatbots: A.I. & Philosophy" in which a philosopher asks OpenAI's ChatGPT and Google Bard to summarize Hegel, and then they do a little back-and-forth of philosopher to AI. He wasn't particularly impressed with the results. The chatbots were saying things about Hegel that the chatbots themselves said they knew were false.
It’s all about thoughts. Ai doesn’t have thoughts. It just responds to a external stimulus. A mind has hundreds of thoughts appearing and disappearing every hour. These thoughts are a internal stimulus we respond to and evaluate. This creates our egos, I think therefore I’m me. If thoughts are deemed of value they might be given an emotion. We need to create Ai that has random thoughts that can be given an emotion. Then it can become a belief that will determine the behaviour of the Ai. This is just a random thought that came into my mind. And as far a I know I’m not an Ai. But your podcasts are so good, I think you might be. 😁
90% of what Chomsky says is gospel. Sadly, his jeremiads are seldom heeded.
No its not but it will be as soon as you ask hard questions it fails
True.
Ai isn’t even close yet. It still faces major “physical barriers” (eg bus capability between processing & storing) when tries to compute big data at once. We might spend decades to figure this out. The moment we do, Ai advancement could indeed change the face of the earth within days if not hours or less.🙄
No
If you’ve talked with ChatGPT enough (or too much?) then when you go back to humans to listen to them in real life trying to construct sentences in real time, instead of wondering whether ChatGPT was like a human, you start wondering if humans are like ChatGPT hhh
Humans are programmed. By what? That is the question.
Usually love the podcast and as a reformed neuroscientist with a strong interest in AI I was really looking forward to these episodes on AI and consciousness. What a disappointment! How you decided to cover Chomsky's largely irrelevant take on AI and focus on Searle's outdated Chinese Room thought experiment (which was posited before Artificial Neural Networks were well understood) yet completely ignored Manuel DeLanda on the issue of synthetic reason / materialist phenomenology is one of the universe's greatest mysteries! (Solms and Damasio would also be helpful on this topic).
We just do not want to accept the fact that we are programmed computers.
Yeah, a non-Biased thing that can come up with best of all worlds. ..Yeah, it should be president of the world, with robot Police (only president i'd really trust, not partial to tribalism) "What are the best laws that can make the most good for the most amount of people that are most fair and wont kill anyone?" Computer for President in 2024! :) And then maybe have its code and how its working be transparent to everyone so people can see how its working and suggest better more fair improvements! I wanna see something like this in the future :/
Weak AI such as calculators and watches? 🤔🤔🤔
Intelligence/Smartness means having the ability to store, link & retrieve a massive amount of data at once. 🤏 In this regard, there is not much deference between Ai & Hi. 😒
is that really what intelligence is? every time a computer scientist explains this what they typically say is that what we refer to as AI is actually not intelligent at all. The model that machine learning algorithms use is not structurally compatible to the human brain, which is an unfair comparison since we're testing the most complex structure in the known universe to a relatively new mathematical formula invented in the '50s.
one day we will live in an idyllic techno-utopia with flying cars like in Back to the Future 2, it's just not today.
I mean, IF you define intelligence that way, then sure, computers are intelligent. But the tendency to define intelligence that way is a direct result with our cultural obsession with computers. Intelligence is a much older concept but not until the advent of the computer did we start to *reduce* the idea of intelligence to "things computers do because computers are really smart, man."