Penrose says AGI will never be able to beat humans in some domains due to quantum interactions in our neurons. In other words, human understanding (and potential!) is fundamentally non-computable.Research: https://iai.tv/articles/penroses-theory-of-consciousness-supported-by-new-quantum-evidence-auid-3667
>>109846307He's not very smart
>>109846307what about quantum interactions in silicon?
>>109846307>Guy says>Quantum interactions
>>109846330n-n-no! becuase humans are different! why? b-b-because they just are!
>>109846307what if we replicate said phenomena in a machine?thats the thing, if these pehnomena can get generated in humans, that means there exists a configuration of matter where these phenomena are possiblei think someone along the line of mr Penrose -> journo -> opisafaggot is misusing the term "agi"
>>109846307dont we have quantum computershow is it non-computable
>>109846307AGI will never happen (in our lifetimes) because there is more to consciousness than matrices of neutrons
>>109846307He's senile.
>>109846328He'll probably calm the midwits down with an answer like that though. It's fine.
>>109846345theyre more quantum co-processors than quantum computersbut yeah, theoretically, we could replicate the quantum nature of the human brain in silicon
>>109846348why would one need to figure out consciousness for AGI?
>>109846307he's right, AGI is a marketing gimmick to spread fear, extract resources and get more power the common folks who believe in AGI are the same low IQs who believe in UFOs, which are made to increase the defense budget and subsequently empower the military-industrial complex >>109846330>>109846340they are really incomparable. the underlying quantum laws are the same, but physical systems and the quantum phenomena that matter are very different. I could go on on why the lack of quantum origin vs quantum coherence in silicon but to speak roughly, same quantum mechanics, very different materials, environments, particles, timescales, and emergent behavior
>>109846380based knower
cope, bbc won, bharat won, that funny indonesian piece of wood won, vibeGODS won
>>109846307Kek, we've gone from "AI beating us is science fiction" to "AI won't beat us because of science fiction"
>>109846490>subsidy money dries out>token cost goes up 10xnvidia won. anthropic won. openai is a faggotvibeshitters in shambles, scrambling for a response
>>109846328He is very smart if the definition is highest intellectual achievement. He isn't the first smart guy to believe some voodoo shit.
>>109846307Has nobody ever told him this quantum tubule nonsense he spews constantly is retarded? There is nothing scientific about it and he doesn't even seem to understand exactly what he is saying. He's been going on about it for decades now, maybe he's finally gone senile.
>>109846307>Replicate quantum interactions in the AGI Ok, now what?
Why do machines need to emulate a human brain to "beat" us? Calculators are simple as fuck but they're way more precise and faster than any living person at math.
>quantum mumbo jumbothat shit is not real dawg
>>109846307>t.
>>109846307>we can never build a purely physical system that fully replicates human brain function because of this aspect of the brain i made up which still is just a physical process
>>109846307
>>109846307>AI will never be able to understand natural language>Fine, but it will never be able to write functional snippets of code>O-o-ook, but but it still messes up simple math problems and logic puzzles>G-guys, it will never be able to work on big code bases>No no no, but it won't ever be able to solve real math problems (like the Erdos problems) or be able to be autonomous at coding>Ok ok guys, it can replace programming, but it will never be able to replace mathematicians. It won't ever solve a millennium problem like Navier Stok-AAAACK
AGI may exist one day, but it won't be in the form of an LLM. Adding more compute to an LLM only makes it a better LLM.Prediction of text is not intelligence. It's copying one thing intelligence can do.Imagine an experiment where you give an LLM a physical environment containing unfamiliar objects, unfamiliar physics, no language, no internet, no humans, no demonstrations, no textual descriptions.Then tell the LLM to figure the world out. A human child can do this.
"If we make this system sufficiently good at its current objective, eventually it will become a fundamentally different kind of system."
>>109846307> Glorified autocomplete that predicts the next word trained on existing data can't build anything new.Wao thank god we have someone smart to tell us that.
>>109846307>In other words, human understanding (and potential!) is fundamentally non-computable.Ok. And what exactly prevents AGI based on (fundamentally non-computable) analog computing?
>>109849131>w-w-we can build a purely physical system that fully replicates human brain function Why would you do that? Brains already exist.
>>109846345no, despite the billions in research funding and the ever increasing size/number of bits of the quantum processors, none have ever been shown to work.this should not be a surprise because at the quantum level all research is based on extrapolation of extrapolation of extrapolation of hypothesis. thats why everyone went nuts when some old guy said some particle existed because of some phenomena, and then they spent billions to make another phenomena also be wrong by a similar amount unless the particle existed.you know when you are in school they teach you about the bohr model of the atom, turns out science believed it was correct because some theory calculated some number and the real number was correct to 3 or 4 significant figures so they said ok good enough. turns out it was just coincidence. atoms don't follow the bohr model.science about things too small to observe are fundamentally flawed.
>>109850073>>AI will never be able to understand natural languageIt doesn't understand language. It copies language using statistics. These are extremely different things.All the other stuff you mentioned are things computers have always been good at.
>>109850073it cant do any of those things
>>109850313>All the other stuff you mentioned are things computers have always been good at.Yes anon, I used to solve millennium problems in basic in the nineties. I think Grigori Perelman just wrote a python script to solve the Poincaré Conjecture.
>>109846307Translation for those that need it:Our brains are complicated ass shit. We don't really understand this complicated shit in our brain. So we can't expect to copy that complicated shit with a machine.
>>109846345>>109850295Threadly reminder that quantum computers can't compute anything non-computable. They can be fully simulated on regular computers. Being quantum has basically nothing to do with being non-computable. Analog systems are technically non-computable even without quantum effects.
>>109850131Its all fake hype to get people to invest in A.I. It obviously has some good uses but its limited. We are not close at all to real A.G.IThis is a glorified composite system at best. Even then it requires so much training and raw horsepower to even work properly. Its gonna end up mainly being business and research software. The creatives will dump it once they see how repetitive it gets with art.
>>109850369>painstakingly exploring a search space defined by humans much faster than humans can do itYeah, that's something computers have always been good at.
>>109850313>It doesn't understand language. It copies language using statistics. These are extremely different things.Explain how they're different without fantasy fiction.
>>109846307His fragile ego couldn't handle the possibility of AI being smarter than him.
>>109846341quantum shit is based on probability, computers can't do randomness, just imitate it
>>109850611True random number generators already exist.
>>109846307Why do you stem bros worship physicists? They're all quacks with god complexes
>>109850449Explain how are they the same first.
>>109846307Cope more luddite Penrose trans
>>109850131>unfamiliar physicsI dont think humans could cope with this either, our brains were designed to learn regular physics, not some fucked 8 spatial dimensions 1 + 1 = 6 physics of other realities. The baby would lose
>>109850449Consider a calculator.A calculator "does math" better than you. You can "ask it" or "prompt it"789÷788And it can instantly "respond"1.0012690355329949238578680203045If you went on this demonstration alone you would have to conclude that a calculator "understands" math better than most/perhaps all humans.But demonstration of some result, doesn't say anything about understanding.If you "ask" a calculator, "what does 1+1 actually mean?" It can't even receive the question. All it can tell you is 2 if you narrowly prompt it with 1+1. It can replicate the result of understanding but it can't understand.LLMs are just more sophisticated because they have a human language layer in between you and their underlying machinery. But what are they doing? They're using statistics to generate text. They will be, or perhaps already are, better at generating coherent text than most humans. What does this imply about the understanding an LLM possesses? Nothing at all. Using the accuracy of the text result to the question being asked and using the behavior as proof of understanding would be like looking at "2" from the calculator and inferring that it must necessarily "understand what 1+1 means"
>>109850449Isn't it obvious? Copying a pattern does not imply understanding of the pattern, or understanding how the pattern first came to be. LLMs quite literally copy patterns since that is what they are designed to do. This does not imply or require any form of understanding at all. Why would you think the pattern machine "understands" anything in the first place? Do you also think your car understands why its wheels spin, simply because they do in fact spin?
>>109850653A baby experiences unfamiliar physics. Familiarity to you =\= familiarity to all babies lol
>>109850689Humans don't understand why they think either.
>>109850698Yes but my point is that even though the baby finds the physics of the world unfamiliar, its brain has already adapted to be able to learn it. The LLM wasnt designed to learn our physicsThe equivalent of putting LLM in our reality it has no adaptation to is the equivalent of sending a baby into a different universe with a configuration of physics it has no adaptation to which is a fairer comparison. Atleast until the software of LLM changes to increase its learning capabilities of our physics
ye, may be funny as those, remember, those retards who laughed at the theory of brain water change during sleep, like some siphon relaxes in there and the water changes.. i tell you, hes probably right, under the cold brain operates better, not like in the freeze, but in reaaal cold there are cleaner thoughs
>>109850724Is a calculator smarter than you? How about a gps system? Where do you draw the line? Exclusively with computations that can generate compelling human language? Why not other sophisticated computations that can generate things you can't?
>>109850749also evolutionary dumb people are in hot areas like African continent, so have to keep it cool. remove hair etc
>>109846307Quantum interactions != non-computableOur quantum theories ARE computable. We just don't have access to the internal quantum states.It is possible that some computations are done extremely efficiently using quantum effects, but AGI could just use quantum computers to do so. There's no way a Nobel prize winner can't entertain the idea that AGI would use a quantum computer.
>>109850748Your point misses my point, my point misses your point, round and round we go. Except I started it. So you're missing "the" point.Which is that humans are conscious, and can actually think and understand, whereas LLMs are not conscious , do not think, do not understand, and merely generate text.
>>109850808>whereas LLMs are not consciousBig claim there considering we can not collect any information whatsoever about consciousness including where it exists, how its made or what the fuck it even is. You just want a reason to feel superior to LLMs so you use things we have no understanding of to prove it
>>109850808>Which is that humans are consciousNow prove consciousness exists
>>109850832If you don't feel superior to a program that generates text I feel bad for you.
>>109850848Im not so insecure to feel threatened by a program to begin with. My identity doesnt shatter if a robot beats me at something, yours does
>>109850833I can't. I can only appeal to my qualia, which I must assume you possess for the sake of function.Now prove an LLM understands anything.
>>109850861Precisely the opposite. Machines have been consistently beating humans at many things for a very long time. Essentially since machines have existed.This isn't threatening to me at all. You're just incapable of understanding my argument so you're getting emotional. Maybe run my argument through chatgpt and see what it says. Probably something like, "there's real substance to that argument"
>>109850865NTA but believing in qualia is the greatest midwit filter
>>109850881>Believing in the existence of your own subjective experience is midwitI guess I'm yet to ascend then.
Humans are stupid af, it’s not exactly a high ceiling, llms are already superior in many ways and it won’t take long to beat us
>>109850891>existence of your own subjective experience>qualiaNice try
>>109850898Using different definitions is a funny basis to gloat or feel superior
>>109850917If you don't understand what qualia is claimed to be then there are plenty of great resources to learn about it
>>109850877Mhm, is that why your actions are exactly the same as people with a need to feel special? "Oh i have this thing but this robot doesnt!" is pretty much the only thing you have been arguing this entire time. Its like your subconscious is screaming "im special, im special, im special" over and over
>>109850929I believed I knew, and then looked it up again when you challenged me, which confirmed I did. So we're at an impasse unless you have more input than "heh. read a book maybe."
>>109850951Qualia are a special type of retardation that extends beyond simply "own subjective experience" (which plenty of theories incompatible with qualia also pose)
>>109850938>I'm not crying you're crying!>Take it back! LLMs can understand me! My AI boyfriend is real!
>>109850963Can you define it or point me to someone else's definition?
>>109850993Wikipedia suffices, it details a few properties of qualia that make it exceptionally retarded
>>109850980Believe it or not, i made no claims that LLMs could understand me. You must understand that just because someone is against you, it doesnt necessarily mean they belong to the opposite side. This is a common hallucination that some people have but the world isnt as simple as tribal unga bunga im afraid.
>>109851003>Common hallucination Like universalizing your emotive experience
>>109846307>in some domainsCould he be a little bit more specific? What domains and what kind of tasks exactly?
>>109850635>Explain how are they the same first.Why? I never claimed they were the same.>>109850687>If you "ask" a calculator, "what does 1+1 actually mean?" It can't even receive the question.But if I ask a token guesser instead, it can do so better than you.>What does this imply about the understanding an LLM possesses? Nothing at allI don't know. You haven't actually made any argument one way or the other besides a pants-on-head-retarded "analogy" about calculators not understanding arithmetic.
>>109851040Theres hardly a need for that. I can use christians as a reference, they believe they are shaped in the image of god. If they dont have a need to feel special, who does?They believe animals dont have souls and humans are the best (if you ignore things that dont matter like swimming speed, general strength, running speed and yadda yadda yadda)Oh wait, that sounds familiar. You believe programs dont have souls and humans are the best (if you ignore things like calculation, time keeping, location tracking and yadda yadda yadda)
>>109850963>t. no-qualia chabot
>>109851066>le chinese room is... le not real!
>>109851076>le chinese roomThat's not an argument you made.
>>109851066>You haven't actually made any argumentNot one you want to hear anyway.
>>109851075>circular reasoningExactly why qualia are the perfect midwit filter. Even under such a retarded framework it would be impossible to disprove the claim that chatbots have qualia.
>>109851076>>109851079Yes, because that's a different person.
>>109851089>it would be impossible to disprove the claim that chatbots have qualia.I think that's really simple: they only have what they're programmed to have, and they aren't programmed to have qualia. (They're programmed to generate text)
>>109851107Your schizobabble is valid if qualia are real, which they aren't
>>109851086>Not one you want to hear anyway.You literally haven't made one. A calculator doesn't simulate semantics.
>>109851119>You literally haven't made one.Of course I did.>A calculator doesn't simulate semanticsNo shit.
>>109851089Did you reply to the wrong post or are you having yet another psychotic episode? No argument was stated or implied in that post, circular or otherwise.
>>109851131>No shit.So what's your "analogy" about?>"simulated" semantics are not real semantics j-j-just like uhhh... un-simulated non-semantics are not real semantics You're literally 80-IQ.
What is the argument that any of you are willing to make and defend that LLMs can become AGI?
>>109851095Doesn't matter. The only thing the Chinese Room argument demonstrates is that you can't attribute understanding to the computer's processor.
>>109851140>If I s-s-stutter when I m-m-make shit up it makes y-you look s-stupid, not m-m-me!!You're literally average IQ, 98 or so.
>>109851148If by LLMs, you mean transformers trained on natural language text, then no. If you mean trained on tokens in general (like modern multimodal ones), then yes. OP argument falls flat because >>109850800
>>109851157Notice how you can't progress a single step beyond insisting that your 80 IQ "analogy" is good. You can't even explain what aspects are supposed to be analogous.
>>109851158So your argument for LLMs becoming AGI is, "then yes"?
Not sure how the Chinese room argument is relevant here.>>109851132I merely preempted the argument >>109851107
>>109851107Very few programs are designed to have bugs but they still do. Human intentions are the be all and end all of things
>>109851165What you would notice if you were sufficiently sophisticated is that I've judged you as not worth interacting with beyond trying to antagonize you with minimal effort.
>>109846307There's nothing humans can do that AI can't do, or fake.
>>109851175So are you proposing that AGI will be a bug?
>>109851170You should have preempted the entire exchange by taking your Thorazine.
>>109851158Multimodal is a meme currently. You send it a image, it transforms "what it sees" into text, then uses its own text description as basis, that's why even shit like astra can't replicate a apartment design and layout by image alone. Complete meme
>>109851178>or fakeIs doing literally all the work
>>109851177>i can totally elaborate on how a calculator is analogous to a LLM>y-you're n-not heckin' worth itWhy are nonwhites such bitter losers?
>>109851167Without circular premises there's no reason that a neural network couldn't approximate to any desired precision the function of a human brain given appropriate sensory instruments. The literature shows that self-play (which physical humans cannot easily do) results in exponential improvement, so yes, AGI seems quite feasible.>>109851183This is not true but Astra is specially autistic about media. Try Gemini they have a much better pretrain for media
>>109851194>if i had an infinitely big neural network and infinite computing power and a magical training regime i could simulate a brainWhat does that have to do with reality, though?
>>109851204It doesn't have to be infinitely big, just big enough>can becomebtw
>>109851190>ur brownlol good one
>>109851210>It doesn't have to be infinitely bigIt does, otherwise it can't simulate an analog system. But even ignoring that, the magical training regime doesn't exist, either.
>>109850687i can use the exact same logic to argue your brain doesn't understand math. textbook case of proving too much
Humans need to stay in their lane. AI is best for objective things. When you're looking for an answer, you talk to AI, but humans are better in subjective matters. Like futanari women should always be at least half-erect while in their most casual "least horny" state of being. They never truly go soft.
>>109851222Lemme see it
>>109851215It doesn't need to simulate an analog system exactly, it just has to be good enough>the magical training regime doesn't existThere's no reason it couldn't>>109851223>t. AI
>>109851256>it just has to be good enoughGood enough for what? It's already good enough to fool midwits without being anywhere near what you're referencing.
>>109851270What do you think AGI even is? Superintelligence is a greater bar to clear but generalist intelligence does not seem implausible, at least as well as humans
>>109851256>It doesn't need to simulate an analog system exactlyThat's your subjective head canon, not an actual argument.
>>109851251see what? what you call your brain is actually a bunch of elements blindly following rules they don't understand
>>109851277>What do you think AGI even is?I use basic definitionsIntelligence = The ability to acquire, understand, and use knowledgeGeneral = Not limited to or dealing with one class of things; diversifiedArtificial = Not arising from natural or necessary causesCurrently LLMs fulfill:ArtificialThey do not fulfill:IntelligenceGeneral
>>109851306>what you call your brain is actually a bunch of elements blindly following rules they don't understandYep. The difference between a calculator doing arithmetic and a brain doing it is that one system simply follows rules for doing arithmetic, while the other follows rules for a thought process about following rules for doing arithmetic. But you can't expect an 80 IQ specimen to keep track of meta-levels, hence retarded analogies between LLMs and calculators.
>>109851501If you gave a calculator infinite compute would it ever cease being a calculator?
Does a thermostat understand temperature?
>>109851306>what you call your brain is actually a bunch of elements blindly following rules they don't understandAre you sure? How do you know that?
>>109851552>i don't understand what i'm reading but i must reply
>>109851501Oh dear, your argument is almost convincing me that emulation is more advanced than the real thing. To possess not only your own rules but also a replication of another rules is almost like having an understanding of two rule systems at once. Would we be able to use our own rule system to replicate the rule system of an LLM? From what ive seen of their inner mechanisms, i highly doubt itPerhaps in a way, they have already surpassed us in understanding
>>109847764Because they are unable to come up with anything new or creative, they wouldn't even be able to tell what's new or creative, unless there was a strict criterion in the datasets
>>109851666fair enough. i don't actually know the individual neurons of your brain don't understand what makes your cognitive architecture work. maybe your neurons are very smart. but then i wonder why you'd even need more than one neuron. i guess you wouldn't. one neuron is probably all you have. all i can do now is bow before that solitary neuron for stumping me like this
>>109851697Can you? Everything people "create" are just derivatives of things they have already seen. If you are able to imagine a colour you have never seen before, that would truly count as creative in my book
>>109851685>i don't understand what i'm reading but i must reply
>>109851703>Can you? Everything people "create" are just derivatives of things they have already seen.Holy Mother of All Projecting Subhumans...
>>109851704Come now, knowledge is meant to be shared, what did i not understand?
>>109851710Your reply is so incongruent you probably didn't understand anything at all.
>>109846307How good would this be for a page filing A.i. model that spikes in a dyssynchronous polychronous golden rhythm in a spiking emulation running ubuntu in with brian2cuda.I want the a.i. to customize itself and preinstall modules into it's "brain"
>>109851717Well as much as id like to act smug and talk down to you from a high horse while poorly acting like a nice person, im actually rather tired so go bother someone else
>>109851735I wasn't talking to you in the first place, animal. Yeah, go bother someone else.
>>109851000Not him, but if you're not going to be specific about what Wikipedia says that supposedly proves you right, then he's more convincing than you. That's probably why he stopped replying.>>109851324Not him, but the trouble here is, that the AI companies themselves have put out so many conflicting press releases, they don't agree even in their own companies what AGI is.The closest you get is the weekly, monthly>We made an AI capable of destroying the world, please investor>an LLM that is human...ish, maybe? We don't know lol. What we can tell you is we just think it will change the way we live forever. And you should invest.
>agi>quantum consciousnessBoth meaningless terms."General intelligence" comes from the neocortex essentially being hundreds of thousands of minibrain neuron columns that each model a concrete or abstract object based on data coming from other regions of the brain, and they come together in a voting system to generate the most probable model of the whole environement. It works generally because each of these columns are the same hardware regardless of the data they recieve, and all data gets a geometric representation internally (even abstract concepts, it's probably why we love metaphors of location and distance). LLMs are inherently incapable of doing that by their architecture.Consciousness is a seperate issue entirely. The only decent physicalist theory of consciousness is Integrated Information Theory. Everything else is just "erm consciousness isn't real chud" or "le quantum!". Even abrahamic normie belief in a soul is less silly than any philosophy that could lead you to believe LLMs are even approaching consciousness.
>>109851750>delusional head canon
>>109851703Yeah, I just had a thought that I think was never had before. Or at least it wasn't expressed in memorable form.Which would explain the source of the so-called "ocean feeling" mentioned in an exchange of letters between two psychologists.It has to do with the necessity of incompleteness in the causation of such mental phenomena that would "bait" consciousness into "higher" states that couldn't exist simply by virtue of the classical biological explanations.It doesn't mean the idea is testable, but it's very likely a completely new idea, to the extent of my knowledge it hasn't been expressed before.
>>109851766Or maybe I should have said indeterminacy rather than incompleteness.
>>109851766NTA but I can get a mindless token guesser to generate more coherent and interesting stuff than that, so it's not really proving your point.
>>109851766I havent the fainted idea what this ocean feeling is or who these psychologists are. Probably a good idea to provide context when talking about niche events
>>109851784You can get a quick idea on wikipedia.>>109851781Very much doubt you can get anything usable in terms of metacognition from a chatbot.Here's an idea: ask it to write a new theory of gravity.
>>109846380Please do elaborate how quantum shit behave differently in silicon. I'm low IQ money tho pls use trump speak so I can understand
>>109851828Well, i normally dont use other peoples explanations but it seems like Feud himself made a theory on where this came from that fits well>Freud argues that the "oceanic feeling", if it exists, is the preserved "primitive ego-feeling" from infancy. The primitive ego-feeling precedes the creation of the ego and exists up until the mother ceases breastfeeding. Prior to this, the infant is regularly breastfed in response to its crying and has no concept that the breast does not belong to it. Therefore, the infant has no concept of a "self" or, rather, considers the breast to be part of itselfSo if he had experienced it before the concept of a self was created, it wouldnt be a new thing. It would just be an exceptionally jarring thing to experience when combined with the knowledge of a world that an infant would not have
>>109851881Yeah I think his theory was wrong.It just doesn't explain anything about the hybrid nature of human consciousness that produces a level of creative complexity that would be unnecessary just by virtue of biological drives.It's not really necessary to write a very complicated metacognition theory to secure resources or vagina. You could do it in much simpler ways and without even meaning what you write.So the explanation must produce a factor that makes human consciousness somehow rise to such a level of cognition as if it's an inherent necessity, motivated by the intrinsic drives, rather than a biological imperative.
>>109851703>Can you? Everything people "create" are just derivatives>>109851703>Everything people "create" are just derivatives of things they have already seenYour vague claim is either false or irrelevant, depending on how it's interpreted. But there's no need for this pseudophilosophy. The matter is settled by observing that a LLM models the distribution of human data it was trained, which gives you an objective metric for novelty: novelty is what doesn't fit in with that distribution.If a LLM could produce novelty, you could train it on its own slop and expect progress. Instead, you get model collapse. If humanity COULDN'T produce novelty, a LLM trained on all human texts produced before the 20th century would be roughly the same as a model trained on all human texts produced so far, which would be roughly the same as one trained on all that plus every human text produced in the next millennium etc.A retarded and untenable position, clearly. A LLM is literally the perfect tool to measure how wrong your opinion is.
>>109851910>It's not really necessary to write a very complicated metacognition theory to secure resources or vaginaI really want to argue against this but i deleted everything i wrote because->motivated by the intrinsic drives, rather than a biological imperative.Im far more interested in whatever this is. What in the world do you mean by intrinsic drive? Where would such a thing come from?
>>109852006Such as chasing a crazy new idea just for the sake of it, rather than thinking of just confirming your evolutionary blablablahs.Remember that guy Evariste Gaullois. He got into a duel with a guy over simping for some girl. And the night before the duel, he didn't get any sleep making sure he writes all his maths ideas down because he was expecting to get killed.That document was one of the most valuable in the history of maths and it helped create new fields in maths.The 'evolutionary imperative' would have dictated something else, like maybe going out and making sure he impregnates some girl before he dies. But he spent his last day on earth writing maths.
>>109852051Ahh alright then it all on how metacognition secures resources. The thing is, you are absolutely right, it doesnt. Philosophers themselves are often childless so it clearly isnt getting them much pussy eitherThe thing is that humans are pretty terrible at predicting whether something will turn out useful or not. So rather than pick and choose, we just uncover everything and anything in our path. The vast majority is useless garbage but the few useful things that we uncover from this relentless pursuit makes it all worth itTake bread for example, how in the world world we have ever discovered such staple food just by looking at the usefulness of grain? Its a shitty inedible plant, any sane usefulness focused species wouldnt have gave it a second glance but our constant theorising turned it into infinite caloriesI have less clear idea on how it helped cavemen since i dont actually know what they were in possession of or how much they knew but im sure it led them to quite a few discoveries that they otherwise wouldnt have found
>>1098520512/2Those mathematics? It had far more potential to help the tribe than one man breeding even though it was probably going to be useless garbage like the +90% of other things we discover. It was entirely coincidental that it wasnt and those coincidences are what makes us thrive
>>109846307Penrose is wrong. Look up Max Tegmarks arguments on the subject.
>>109846307AGI is not possible.I’m sorry.Even without rail guards.AI remains retarded and dumb
>>109846307ai already "hallucinates" and that introduces enough randomness to approximate whatever the hell he's talking about in humans. I love this guy btw.
