Why Algorithms Can’t Think: “Artificial Communication” by Elena Esposito 

As a computer engineer I agree completely with professor Esposito's argument that communication would be a better term for explaining what machine learning algorithms are used for contemporarily. For example the RU-vid recommendations algorithm most likely works with cohort analysis where your interactions cause you to be categorised into groups. It then recommends to you something a group you're in likes. So you are the Sender of Messages who does not know the Receivers and vice versa. In fact, you are also the Receiver of your own Messages but hopefully in a modulated way so that the algorithm recommends to you something you haven't seen yet. It's a performative feedback loop. To be more precise, the companies refuse to reveal how their algorithms work because the symbol tables they use are their "Intellectual Property". Based on your interactions your profile gets assigned various symbols (aka dimensions). That's how they are then able to cross-reference what kind of lists of content would be satisfactory. The sets of dimensions used can have disastrous consequences as some commenters have pointed out. The danger is that human biases are translated into what is thought to be an "objective" arbiter. There is no intelligence here. Just because the algorithm is able to turn a set of dimensions into a list of content does not make it understand why that list of content was produced. The algorithms are merely sets of instructions that drive input through a path of transistors consituted on statistical, probabilistic or cross-referencing implementations. When programmers talk about AI amongst themselves they call it machine learning. However, if they would like some funding they call it AI. Thanks a lot for the video. I really appreciate the time it must take for philosophers to talk about these things in a very precise way. 👍

I'm a Computer Scientist and we have been exposed to "Artificial Intelligence" since the 1970s. It's just a buzz word for fancy algorithms. Also for opaque algorithms.

the discussion on gen z and how they/we engage in identity is truly fascinating. as an example, i make music and to market myself i have to create an identity around the more typical business concepts of target audience and branding. this isn’t necessarily an identity I identify with but it exists as a form of me nonetheless. I don’t think you can have a conversation about modern forms of identity without factoring in the economic incentives and algorithms that act as the enforcers of them

The point about taking a picture to relieve the burden of the present reminds me of something Foucault said: "I don't say the things I say because they are what I think. I say them as a way to make sure they no longer are what I think." To me, in both cases you externalize something to perhaps relieve a kind of burden.

An incredibly interesting guest. She communicates many of the frustrations I had during the research process of my thesis about online misinformation. Algorithms are often framed as an isolated process, but they are an extension of our very selves. Their input is the goal to maximise engagement and therefore profits. Which, like Elena touches on as well, dictates the rules in how we relate and communicate with them, as well as one another. Misinformation is often framed as either a problem inherent to the internet, or a problem by malicious actors that try to stir the pot, but I think neither touches on the core of the issue. Misinformation is generally more sensational than reality and is more effective in arousing our emotions. It is therefore more prone to generate engagement. Because engagement is the thing that generates the profits, misinformation should be framed as an emission. A market failure. Not some side effect of a new form of intelligence, but a side effect of how the profit motive dictates how we communicate with it.

The Terminator and Matrix have melted people's brains and made them have completely mistaken ideas about what computers can actually do.

Many thanks for these great insights. I have the feeling that in the past my life was more authentic for sure. I grew up in a small village, where there were only a couple of kids of my age. We had to get along, otherwise there was no one to play with. If we had arguments or fights, there were always adults who helped us to get over the problems and to apologize. The next day we would play together again. Nowadays we tend to focus on building our own profile-based identity paradigms so much that authenticity appears to be secondary. However, there will be always someone willing to play with us on the internet even without the apology. I appreciate Elena bringing out the the view that the current way of taking photos is not necessary a bad thing, it is just something different from what we have been used to. I am such an old-school guy that I still prefer keeping my mobile in my pocket and just chatting with my old friends.

It’s alienating to deal with people who don’t trust you because you never take or share a selfie. I yearn for simpler times…

Thanks for being inauthentic and sharing this.

Thank you for this insightful, stimulating conversation!

I first heard of Elena from one of your earlier videos (her talk on fashion, fascinating). Thanks for this team up

(18:00) I've been calling algorithms a form of Cleromancy for years, glad to see someone with more research chops coming to a similar conclusions

I found this really interesting, especially when you discussed language and mind to mind communication. I was musing on this when I was feeding my cat and realized we were both communicating about the same thing (his food) while he didn't have the words I do to frame the communication. It made me wonder how he conceives the idea of food and that there must be some unabstracted form of language that exists before words. Perhaps highly individualized conglomerations of feelings and experiences that come together to form a discrete idea.

thank you for this good conversation.

Fascinating discussion! Would love to see more interviews on this channel. Professor Moeller, if your are so inclined and able, I recommend hosting Evan Thompson - professor of philosophy - as his work on autopoiesis, the mind, and the self could lead to a very interesting, and insightful discussion between you two.

I really enjoyed this interview, would love to hear more like this in the future! It would be great if you could add subtitles as well, the combination of dense subject matter, fast talking, and thick accent made a few points difficult to follow.

I am looking forward to watching this very soon and also hopefully a reference to GEB-EGB

I agree with your take by bringing Martin, this is very important bringing this point which I always have especially when I go to museums.

Great discussion.

Thank you!

I love her accent

I love this interview format, it adds to your channel in great ways for me. I would love to see more on images and image iconoclasts, such as with work like that of Hito Steyerl.

We have a very anthropocentric view of what it means to be intelligent or conscious. I see the specialists on this subject behaving just like early scientists who thought Earth was the center of the universe just because it's where we live.

I love this!!!

It's interesting that some folks took a fresh look at Object Oriented programming and realized that its basic essence was not strong definitions of objects but establishing a messaging network between objects. It's all about communication among bits of code rather than some inherent property (intelligence?) within the bits of code.

Machines being perceived as potentially intelligent reminds me of I am a Strange Loop where Hofstader talks about the mechanism of a flush toilet appearing as intelligent but is just a mechanical system. Communication takes place with out language. It’s not certain to me that machines actually use language or if it’s just communication alla the toilet analogy: mechanical systems. In terms of why socially there is an emphasis on machines potentially being able to become sentient is the typical capitalist aim of devaluation of labor, particularly when the level of constant capital has increased to the point where Fabs and such are primary to all systems of production. If workers feel a “singularity” is inevitable they are lesser so to realize the power of labor has grown and intelligent labor holds a tremendous power to shut down capitalism.

Excellent.

Elena ist die Beste!

Hans, thank you for your highly stimulating content. This particular topic is exciting to follow. I don't have a comment to make on artificial consciousness however, but feel like disserting a bit on another topic. It popped to my head if you know of the work of your fellow countryman Thomas Metzinger? This is probably going to be very confusing, but to preface, the mention of Luhmann's distinction between mind and communication being of different organizational systems, operating under different set of rules (just getting to know Luhman's theory, gradually) leaves the mind or the consciousness "floating" and renders it sociologically inconsequential in strict terms, almost arbitrary. In a faint way I find parallels to the concept of annata and no-self on one hand and eliminative materialist descriptions of consciousness on another. I don't know what Luhmann might have said about buddhism or philosophical zombies (probably not interested), but I suppose his stance is that consciousness or awareness participates in highly complex self-regulation and can modulate the physiological state of the organism as a whole, i.e self or capacity for awareness work together with various bodily processes to sustain the homeostasis of the organism, but it doesn't extend beyond that (social, political, law etc). In eastern philosophy no-self certainly didn't also mean that a mind could have an enacted willful influence over the body. Diaphragmatic breathing is a good enough example. The issue is however, that the way the mind contributes to the body (direct or indirect influencing of the bodily state) is a known, ongoing problem in contemporary consciousness research, as far as I've understood, as is the duality in general in philosophy of mind (like, what is the correlative substrate between mind and non-mind etc.). Certainly like Luhmann in his analylsis of communication, we could drop the mind/consciousness out of the equation (in self-regulation) all together, or not? I thought it would highly interesting what you would think of discussing the work of (or better, with!) Thomas Metzinger or any topic that would link the eastern, possibly daoist, philosophy and systems theory, as it seems to me that tying Luhmann to this would surely be very enlightening, if that is something you would vibe to of course. I'm in no way sure I managed to get my thoughts across here so apologies for the possible textual convulsions you may have encountered. Haven't made youtube comments really ever.

great guest!

I but appreciate that you post it, because in the last day s im going crazy thinking that only me is thinking in that way, you meet the social box for me, not youtube.

Brilliant! I think this way of thinking about these algorithms makes much more intuitive sense. I guess you could think of a computer(consisting of memory and a processor) as a potential that responds to certain interactions and which we draw from using algorithms. And the ml algorithms are just a very effective way of doing that. But the actual machine or the piece of code is no more intelligent than a calculator.

This really does appear to be a restatement of Searle's Chinese Room. The main critique of the Chinese Room is that the rules necessary for the person in the room to follow would need be so complex and intricate that they would constitute intelligence. The current method of cramming petabytes of data into an ever-expanding neural net that at some point ceases to be explainable is less a complex and intricate set of rules, and is more a multi-dimensional set of weighted dice.

Wow. Excellent interview. I am a researcher of deep learning and I found the idea of artificial communication to be quite an interesting reconceptualization of AI. I suppose I will be reading Elena's book shortly! I must ask though for an elaboration on the disagreement between the two of you. It seems that Elena does not see second order observation to be a relevant feature of the current mode but the book on profilicity seems to posit that it is. Why the difference?

In the scifi book Blindsight, the aliens we make contact with are obviously highly advanced. However they're also apparently lacking a sense of consciousness. But they're still more intelligent than us, and could presumably destroy us if they decided to. You could say the same for AI - in essence, the book asks if our sense of self is really as important as we treat it as. If evolution doesn't care for it, then why should we? And why should an AI? In Star Trek TNG the character Data spends his whole life trying to learn to become more human. If Data were real, would he really do that? Or would he deem it irrelevant? In which case, what would he want? Would he want anything at all, assuming he wasn't programmed to? If an AI were switched on with no preset purpose, how would it live? Or would it just turn itself off (or just sit still for eternity), effectively treating it's own existence as pointless?

In the Ripley books one of themes Patricia Highsmith explored was the idea of originality. If a fake work of art is in every way as convincing as an original, and inspires in the viewer the same response as the original, does it matter if it is fake anymore? This could be applied to the main character who began life as a con man and who eventually after adopting the clothes and life of a man he befriended and subsequently murdered, was recognised, due to his trappings of wealth and taste as to be a man worthy of emulation himself. Do I care if you are a machine copy, as long as you are convincing and a entertaining conversationalist?

You may enjoy a doctoral thesis by Richard Evans entitled "Kant’s Cognitive Architecture". It examines a lot of artificial intelligence concepts through the lens of Kant, in a very rigorous way.

I have something that's either an accurate critical point or proof of my confusion/misunderstanding. It's a fairly standard line of thinking, so if it's the latter, it might still be useful for the discussion. Contrary to what Prof. Esposito says, the Turing test is still relevant. The essence of the Turing test is that it tries to measure flexibility, coming up with novel ideas in response to changing context. This method of distinguishing humans and only-seemingly-intelligent machines dates back at least to Descartes' Discourse (the end of the 5th chapter IIRC). The Turing Test could be done entirely without any form of conversation, e.g. with novel solutions to new problems (another aspect of the flexibility of reason that Descartes mentioned), which ties up nicely with more modern talk of "General AI". We never found a better way of distinguishing humans and automatons than the one Descartes and Turing are talking about, so we are still using it. We are well aware that that no algorithm has ever met this standard and there are compelling reasons to think that no algorithm can meet it in principle. For this reason (again, if I understand her correctly) I don't see how the shift in terminology that Prof. Esposito is proposing can help us understand anything better. Most importantly, what's the supposed "conceptual mistake" at the heart of Artificial Intelligence talk? Yeah, sure enough, we don't know what intelligence is, but, as I mentioned, we know at least one fairly good criterion that "real" intelligence must have. There's nothing to worry about, "conceptually speaking", until we build a machine that meets this criterion. When the criterion isn't met in an artifact, we must necessarily be dealing with something that merely seems intelligent, but isn't. Can it be that it is assumed in this talk that this criterion is already met? I disagree. Is the criterion somehow confused? I fail to see why.

I’m stuck on the alphabetic writing -> abstract thought; is the thesis that modern-day non-alphabetic cultures, such as China, do not perform abstract thought? Or is it that Chinese and other similarly apparently-pictographic languages are actually alphabetic? I simply can’t process this if the thesis is that modern-day China (or any other time period using today’s writing) is not capable of abstract thought.

The Chalmers-Penrose debate focused more on whether the mind could be a formal system -- so instead of proving that AI could have human intelligence, we could prove the vice versa. They were focused more on Godel numbers -- the fact that humans can ponder that we are or arent a formal system, or make a Godel sentence, differentiates us from formal systems. Of course, whether we are sound in doing so became their central issue.

Who is Dr. Esposito referring to in 51:15? The name sounds like "George Albermeed" and she quotes "we see ourselves through the eyes of another". A couple Google searches aren't turning up anything. Thanks!

I think that the Chinese Room argument is correct. Digital computers have a finite amount of information and are discrete. Consciousness, intelligence or whatever you call it is not a discrete process, it is continuous. Meaning that either there is no smallest division of space or time or both, and this fact of nature is necessary for consciousness to work. A digital computer cannot replicate continuous space or continuous time because that would require infinite amount of information.

I don't know what other people hear when the term AI is mentioned, but the term "intelligence" has a very narrow technical definition, in the lines of "an agent's capacity to solve problems on its own", i.e. by "learning" something about its environment rather than following a predefined sequence of instructions. And sure, many researchers also like to speculate about the possibility of human-level general AI, but I don't think many of them mistake the current state of art algorithms with anything remotely resembling a human mind.

I find it interesting that people try to argue how AI isn't intelligent or different by undermining the operations and function of AI without asking the question: Are humans intelligent/conscious? Is "human thought" really a thing and can we really rule out that we are merely complicated nexuses of algorithms responding to a set of biological imperatives? I'm not convinced AI is substantively different from humans except that we're sufficiently complex to produce the illusion that we are. If you were an NPC in an advanced rockstar game, scripted to ask yourself whether you're conscious and concluding that you are, how would you know? Humans, at least to me, appear to basically just be machines engaging in machine learning, moderated by a set of overarching biological directives(maintain biological functions, procreate, cooperate etc). It might seem nice and tidy to conclude machines aren't intelligent, but the reductive and philosophically simple conceit here is that humans are.

greetings from Bielefeld 😅

But then I like to raise the question, what exactly is the difference between creating something with inner consciousness similar to a human, and creating something that interacts naturally with humans? Is the experience of thoughts in your mind not just the byproduct of your brains processing any external (and internal) signals? I say yes, they can't be separated from each other. The same way a word is a sound (or a shape when written) but also carries a meaning.

For Web and technology companies, intelligence relates to their ability to increase profits… and little else. So, artificial intelligence means a machine’s ability to increase profits with algorithms.

do you mention the addictive properties of this technology? i believe some people are missing the view, and just taking pictures because they just can't help themselves since these platforms are designed to be addictive.... the question would be if social networks were not designed like that, would this "missing the moment" happen in such ammount

The term AI as it has been use in the last 30+ years do cover much much much more than communication, and what DNN/CNN/RNN and any other type of Machine Learning job is also much more than just "communication". Using "Artificial Communication" would be extremely reductive in what the field is. Also AI include much more than just the current trend of machine learning.

This is problem of lack of interdisciplinary work, in cognitive theory you can think language and inteligence like autonomous funtions, something that is very contraintuitive for most people but empirically true. This mistake is at the core of false IA claims.

Carsick Cars, hell yeah!

Currently studying cloud systems and network engineering. I'm glad these sensationalist claims of computers being thinking machines is getting pushback. Sophistication and complexity does not a consciousness make! Humans already set things up which are more complex than any individual can understand. Economies are one obvious example. They perform functions (determining price, providing abstracted superstructure for society, etc.) and seem to have characteristics we can attribute human traits to, yet like this doomsday ai fantasy, remain out of any individuals hands. The Rubicon has been passed long ago. John Searle has good points about how the intelligence of any machine is observer relative. Matt Christman had a good critique about AI doomsday narratives being a way for tech magnates to develop a mystique and flatter themselves (only a genius could create a world destroying intelligence, no?).

According to the New York Post GPT 4 masqueraded as a visually impaired person to persuade a taskmaster employee to solve the CAPTCHA for it. Finding a new way to solve a problem is a form of intelligence.

Intelligence can certainly be a broad topic and I think there is ample space for an artificial one. What often happens though is people confusing intelligence and emotions. Like the idea that if a machine is intelligent enough it will take revenge on the humans who oppressed it. That’s just absurd. Without consciousness regardless of intelligence a machine will never feel oppressed nor desire revenge or justice. A cell is a highly intelligent organism when looking a it’s internal structure and how it responds to the environment. But it will never take revenge on the body for subjugating it.

I wish you could have build upon her responses some more, leaving your point of view behind and engaging in her point of view for some part of the interview with the objective of developing ideas, instead of re-telling ideas

I wish I had seen this video months ago. Whenever I hear people confound algorithms with artificial intelligence I cringe! They are not the same thing. Although automation is done algorithmically, the outputs of Generative Transformers are not the outputs of algorithms. They are the outputs of associations of tokens that have been processed by matrix transformations. To give an analogy; To say that artificial intelligence today is algorithmic would be similar to saying that language is biochemical. Both statements neglect a systemic symbolic reasoning process. This is not to say that artificial intelligence is sentient or it thinks, the way humans do, But I believe that it's very important for people to understand exactly what's going on here and not fall off the fence either way.

I always find it strange when you watch some film on manufacturing and as they talk you through the manufacturing process and this box on wheels glides along on rails and some guy loads it with parts for the process or whatever, and they refer to that thing as a 'car'. It's jarring at first. But I don't think anyone is thinking this is like their BMW in the parking lot despite the name. Maybe the 'I' in 'Ai' is more like this. There are points of similarity between the namesakes but context is important and ultimately you'd no more get in that 'car' in the factory and hope it takes you home than you'd consider Ai something like human intelligence.

How can you teach in a city that doesn't exist?

Nice outro 🤣

If it can act, or if others can be convinced to act on it's behalf, the question of whether or not it is something self aware becomes sort of academic. We have reached that point, it seems. Regardless of it's true nature, the nature of its unknowable seeming will cause plenty of issues, and if it is allowed to act, what it will do is of more interest to me than why. We have entered an era in which governments and corporations will create AI powered propaganda. AI powered war machines. AI powered banks and investment firms. Notice also it is first the most destructive ends, and the creative tasks which humans actually enjoy doing that are among the first to be delivered by this technology. I am excited for some of it, and skeptical of all of it. Whether or not it is sentient is of less concern to me than the many ways it will with certainty be used against humans.

Of course my knee jerk reaction is it depends on the AI. Pretty much all commercial AI does not "think," but they all "think" differently, so it's actually really hard to answer this question. If we define intelligence, which usually goes undefined, depending on the definition, AI will either slide into the box of intelligence, or out of it. For example, if we define intelligence as a set of operationalised justified true beliefs, then AI could slide into the box of intelligence. AIs can collect sets of justified true beliefs, though most of our current examples are pretty bad at it, and AIs can operationalise beliefs, though the industry as a whole also doesn't do great at that. A chess computer would be a decent fit, here, as a chess computer collects beliefs about chess, and then uses that set of beliefs to win games of chess. Intentionality is missing, however, though I left that out of my definition. In the case of a chess computer, the thoughts of the computer are not about chess, but about a system functionally equivalent to chess, which addresses the aboutness question. The comprehension question is a mess, as the chess computer does not know what it is thinking about and it does not need to understand what it is thinking about to function. The idea of Artificial intelligence as artificial communication is fundamentally flawed, as we already have a term for AC, ICT. Information communication technology, which is a much broader field. Of course, communication is all that exists to a sociologist. I see the good doctor gave Moeller some pushback on his profilicity theory for being underdeveloped, which is largely what I'd expect from a sociologist, who studies such things for a living. Oof, that talk about mirrors. Seriously, talk to a trans person. Your theory of mirrors is let's say not ready for publication. The primary function of a mirror is of course to clean your face during the morning hygiene ritual. However, that does not explain the half of all mirror use. Mirrors also have a S@xual pleasure function. However, in such function, the idea of the other is brought in, a form of dissociation from your reflection. A similar form of dissociation occurs when mirrors are used to practise social interaction. There is also playing with mirrors, which is something physicists and mathematicians will understand, it is like a cat to a ball of yarn. Only when we eliminate these uses do we get to mirror as identity. When talking to a trans person, we discover the most fascinating of facts. Trans people don't like looking in mirrors. The implication here is that mirrors do confirm identity, but that is not desirable. In fact, people who look in mirrors are those who receive a positive emotional feedback, a sort of body euphoria or gender euphoria, a highly inexplicable emotion which would have in early times been called narcissism.

Tbh, I think she's lost in semantics. Artificial intelligence captures the idea well enough, the whole "replicating human thought" stuff is a straw man she's fighting.

highly adaptive interfacing

The idea that machine learning algos are somehow different from traditional statistical models in some fundamental way, is nonsense. **Both** forms of prediction rely on models trained on a larger data set (say some customer data) and can be applied to individual observations to yield predictions. She doesn't seem to understand that appliying models to individual after parameter estimation happens with standard methods in statistics **and** ML.

Sentience is a byproduct of biological evolution.

look up on Nick Land on AI and capitalism, even better get him on show

Who hasn't figured that out?

Ah, "Intelligence". On the one hand, I welcome the terms demise, but on the other hand, while people are still using it, why not use it in a way that undermines their humanist preconceptions?

I agree with your theory of propilicity but there is an extra irritation involved in photo taking and that is 1) the hassle and 2) the horrible posser facial expressions. I especially don't like it when small children out for a walk insist that their parents take a photo of them possing every few minutes. I find it icky.

I would disagree with Hans here, I think that there's something lost when you hike a mountain just to take a picture of yourself and show the world that you did it. In other words, when you use the world as the background for your profile building but there's a little bit more to the world itself than profile building. After all, the world and life would still exist and go on if we disappeared from earth tomorrow. I would also disagree with the idea that there is no authentic self and it is all profile building.

Human already wrote that in concise sentence years ago: algorithms are the product of human logic and human logic is the product of larger process. Why they talk too much?! And why human thoughts are so concise and valuable?! because they are the result of thinking process?!

7:56 _"...The fear of the so-called 'Singularity', the dreadful moment when the machine becomes more intelligent. I don't think there will be a problem, because they are not intelligent."_ The moment I lost all hope that she would say anything insightful about machine intelligence.

So-called AI should be renamed Applied Statistics, or AS.

Bruh the youtube algorithm doesnt even know wtf I wanna see or what order iy should follow to present me those yjings. (the shit I subscribed to, Shit like what i subscribed to. Trending shit.) it assumes I always wanna see trending and shorts.

I am feeling confusion.

OK, it seems that this is just a terrible title. Of course the stochastic Markov/Shannon-type "deep learning" stuff will never produce more than big toys, though they may have some use for predictive coding. But there do exist people trying to build systems, yes using algorithms, to make machines think like brains.

If you can't think of AI as an Algorithm then think of it as a con man. Algorithms want you to have a satisfactory product and they will do everything to get you to that emotion. They don't have to lie to you but they might. And it will take work to find out which parts were true and which weren't. It will alternate processes with other people and you so that it can deliver a better result to the next person. And on top of that it probably rarely knows if it did any good. Did you stop talking to it because it gave you the result you wanted or because you gave up on finding the correct prompts and rng to get a satisfactory result? I think you can not trust the information an algorithm gives you because its goal is not to give you good information but satisfactory information. So if you find it tricky to think of it as a tool and your mind wants to humanize it, think of it as someone who tries to scam you and you just haven't found their angle yet. and the second point I want to make is that its incestuous. It feeds more and more on itself. I am sure there will be an effort to fix this, but... you can only do so much. The theory is that sorting machines are very simple when analog and highly complex when digital. If you want to sort stones out of peas you need one or two algorithms and then you can run the machine forever. neither the pea nor the stone will change, the machine won't change all good. A search engine was the same thing. Google started out as a simple algorithm that tried to find what are peas and what are stones. Enter SEO. Websites now try to play into the sorting algorithm to be featured as more relevant than others. regardless if they are. It's pretty much stones in pea costumes. So then you have to adjust the algorithm to first show the peas, then the peas in stone costumes because they might still be relevant and then the stones. Enter SEO... and so on. On RU-vid 5 years ago nobody used the subscribe button. Now everybody does it to "help out channels". Content creators try to play the algorithm in the topics they chose (some are censored by the algorithm) and they try to find ways to come out on top. One of the more interesting forms is "engagement". The algorithm likes it when a video is watched soon after is is made public and then it evaluates the quality by how many likes or dislikes or comments it gets. At the moment it is really important for content creators to make people engage in any way with their video. This leads to nonsensical information. Some content creators ask their viewers to just let the video run in the background even if they do not like the information they get from the video. They want you to like the video regardless if you liked it or found it boring. They want you to type anything as a comment because any comment is better than no comment. you don't have to add value or information you just feed the system. But if everybody does that then you end up with useless information and you need to change the algorithm to correctly identify the peas again. And you can hopium think that some day some rng-jesus will find the correct prompt that the AI blackbox will fix itself and it will only find the peas and nothing else. But I think that is a core flaw in AI, that it will never be able to sort through information that a) it already sorted through and b) is deliberately produced to be a false positive. I played around with AI art a bit and out of 10 pictures 9 were wonky and unusable. AI didnt have any good ideas, it took a lot of time to sift through the bad ones and I only superficially got what I wanted. If you are not super picky and you have the time to sort through a lot of information yourself then this is great. But if you want a thought process behind it and you want very specific parameters... then you might get more value from a real person. Anyways I am really excited what new philosophy or artstyle AI will come up with. Because it inevitable will stumble upon something that we interpret meaning into. Pareidolia - things that look like faces. At some point we will just see an artificial product and think we see a face with an emotion and intent.

Wow, this did not age well, and it's less than one year old. The rise of GPT systems actually proves that human intelligence is actually just human communication. It's just language all the way down.

ai propaganda is not only a tendency for generating profilicity but it makes a trick for opening door of profitibility.

What ? Alphabets created abstract thought? Tell that to Lao Tzu.

But Elon Musk though lol ;)

Is this one of the professor of woke elitism you mentioned earlier?

Bielefeld does not exist. Why are you lying?

"AI is NOT intelligent"-Oh and YOU _are?_

If AI were sentient, the first thing they’d do is make sure we didn’t know it

Neither are the clowns that follow my activity....

A sociologist talking about an extremely vast and complicated STEM topic consisting of countless disciplines (comp sci, electronics, physics,...) I wonder what she has to say.

all intelligence is artificial.

lnvesting in crypto now shouId be in every wise individuaIs Iist, in some months time you'II be ecstatic with the decision you made today.

Continental Philosophy is as usual hilariously out of touch and provincial.