>"artificial intelligence">look inside>it's just a search engine with extra steps
dunno man, it coded shit faster for me. and better.it's like I manually plow my field or I use a tractor. with a tractor I do it faster with less effort. that's a win in my book
>>17074235>OP>look inside>its just a fag with extra steps
>>17074235It’s a search engine, stuff in a matrix!And covered in nacho cheese.
>>17074235>intelligence is just pattern recognitionYes, and clearly you are not intelligent enough yet to recognize you yourself is also running on this same pattern.
>>17074235>AI is useless junk>AI will take over and kill us allPick one
>>17074241because you aren't making anything new, youre just rehashing stuff other people already made
>>17074235No, its not just a search engine, it doesn't just search through an index of content to find matches, it is at least an upgrade to a context engine because it not only searches through the content index, but is able to compare all the content on the fly to provide helpful context to the content based on statistical calculations and historical uses of the content.>>17074241And an AI is more like an upgrade to your tractor so it automatically swaps out its attachments based on the conditions it detects.
>>17074248>He says this with a millennial reaction image from millennialslop movieThose shekels are paying off, don't they?
>>17074329It does. Stop lying to yourself.
>>17074241> and betterYou are a fraud.
>>17074284
>>17074343No, I clearly explained additional measures beyond the scope of engines that just do searching and sorting, but since you are obviously so low functioning you can't even grasp the answer, no amount of context will make you less retarded, so you will never comprehend the difference between indexed key word search versus comprehensive contextualization, so as far as you can tell, your mom's just a search engine given she is the one who finds your tendies and brings them to your room for you.
>>17074346>you have to have a high IQ to understand Rick and Mortylol, lmao even
Do you trust a system built by people who instituted laws fundamentally shaping how these ai react that make white people look worse and forbade you from saying mean words about jews?
>>17074352He's right your stupid retard. You're essentially asking a librarian in a library help desk for a book on cutting your dick off and just getting the first medical book on sexual anatomical surgery you find vs the librarian manually fetching you the "For Dummies" version with the page and outline section for a moron like yourself.
>>17074367I'm not white so yes
>it is at least an upgrade to a context engine because it not only searches through the content index, but is able to compare all the content on the fly to provide helpful context to the content based on statistical calculations and historical uses of the content.
>it's just a search engine with extra steps
>Yes, and clearly you are not intelligent enough yet to recognize you yourself is also running on this same pattern.
>>17074405Yes that is what allows it to synthesize novel data based on its own contextualization instead of just reproducing exactly what is in the search array like a search engine.>>17074406Extra capabilities and utility, not extra "steps" whatever that means. So ai is a search engine with "extra steps" in the same way an airplane is just a wagon with "extra steps".
>>17074409You realize you're hallucinating just like your mindless chat bot, right? It's word salad.
>>17074410No, I realize you are a barely literate retard who can't parse together most sentences you encounter let alone form your own sentence that justify your nonsequitur screeching.
>>17074419Look, I get it. You feel the chatbot is somehow different from a search engine because the results are dynamic and you heard it has something to do with "context". And that's about as far as your actual knowledge goes. But instead of realizing you don't understand shit, your broken Broca's area fires up and goes like:>ah, of course, it's just like a smegmato-statistical metacontextual search array engine based on historical analysis, not a regular search engineWhich somehow makes you feel less technologically illiterate. You're never gonna ask yourself "what the fuck am I even trying to say here?" because metacognition isn't your strong suit.
>>17074422Thanks for providing more evidence that you are barely literate. But you clearly still don't understand that search engines only search through existing text, they don't generate new text based on context.>You're never gonna ask yourself "what the fuck am I even trying to say here?" No, you are never going to ask what someone means by what they say, you are just going to search through your database of insults and start calling people chatbots or NPCs or whatever other low effort buzzword is popular this week because you are low functioning with very limited vocabulary capabilities.
>>17074425>search engines only search through existing text, they don't generate new text based on context.What did I just say?>>17074422>You feel the chatbot is somehow different from a search engine because the results are dynamic and you heard it has something to do with "context". And that's about as far as your actual knowledge goes.Pretty funny.
>>17074426>What did I just say?Retarded head in sand nonsense.>dynamicSearch engine results are also dynamic based on number of searches, location, time of day and clicks, but chatbot synthesis is completely ad libbed, it isn't just pulling something from a search, it is generating new information based on the context of the search and the information in the data set.
>>17074433>Retarded head in sand nonsense.But it's literally the same thing you said.>Search engine results are also dynamic based on number of searches, location, time of day and clicksSo you're saying it synthesizes novel outputs based on context? Wow. >muh LLM is generating new informationIt's definitely, 100% not doing that. If you weren't technologically and scientifically illiterate, you'd have known it can't do that in principle.
>>17074434>But it's literally the same thing you said.No its not and I just explained why, dynamic is not the same bar as generated from scratch.So you're saying you finally understand that it is doing something search engines can't do. Finally.>It's definitely, 100% not doing thatIt definitely is. That is why all those new novel papers about old math problems were dumped onto the public to scour and debate.
>>17074434>>17074440now kiss
>>17074440>dynamic is not the same bar as generated from scratch.Your chatbot is not generating anything from scratch.>It definitely isYou're a mongoloid for believing this and it's not even up for debate. A LLM alone can't generate new information.
>>17074447>Your chatbot is not generating anything from scratch.It is. Otherwise you are claiming that no new words have ever been generated since the invention of the alphabet since every possible word is just a permutation of letters from the alphabet.>A LLM alone can't generate new information.Then why is its output novel instead of exactly copy pasted directly from the search index like a search engine?
>>17074460I see you've degenerated into full-blown psychosis.
>>17074461I see you have decided to concede without admitting concession and your projected mental illness will compel you to keep replying even though nobody is going to read your further confessions.
>>17074447>A LLM alone can't generate new information.Prove this statement. So that we can see if you have any clue what you are talking about.
>>17074464You're just babbling incoherently. Your LLM is not generating anything from scratch and it's not producing new information. These are two basic facts that no one disputes except for mathematically and technically illiterate brown insects like you. They are not up for debate.
>>17074466>Prove this statement.https://en.wikipedia.org/wiki/Data_processing_inequalityThe proof is at the bottom of the page. Just take your L and sit back down in the dunce corner.
>>17074469So the alphabet can't contain any more information than the alphabet, its just a string of letters that can't be rearranged in novel ways to create new words that mean something other than the alphabet?
>>17074471>psychotic word saladWipe the foam off your mouth, retard.
>>17074474Just take your L and sit back down in the dunce corner. The alphabet can be used to generate new information that isn't just the alphabet.
>The alphabet can be used to generate new information that isn't just the alphabet.
>>17074476Yes, you are currently rearranging letters of the alphabet to generate information that isn't part of the alphabet itself.
>>17074477>you are currently rearranging letters of the alphabet to generate information that isn't part of the alphabet itself.You know what? Sure. Now tell me, what new information have I generated? Be specific.
>>17074479That you are retarded and you don't even know that new words can be made by making novel permutations of letters in the alphabet.
>>17074481How come asking you a simple question about your statement completely derails you? You said I was generating new information. What new information did I generate? What did it pertain to?
>>17074447>A LLM alone can't generate new informationThen what exactly is happening when an LLM is hallucinating if it isn't creating new factually false information as a result of processing disorders?
>>17074483How come giving you answers to your questions makes you have a complete breakdown? I told you what new information you have provided to the thread, just because it isn't flattering doesn't mean it isn't new.
>>17074484>Then what exactly is happening when an LLM is hallucinatingA probability distribution that encodes a fixed amount of information is being sampled, producing no new information, unless by "new information" you mean finding out what particular token slop sequence the PRNG will produce this time, rather than "new information" of the kind you need to train a model in the first place.
>>17074485>I told you what new information you have provided to the threadFull-blown psychotic delusion. Moving on. Notice how your mental illness forces you to reply to me again even though nothing you shit out will be read from this point on. :^)
>>17074487>producing no new information,Then why doesn't it match reality, if it isn't new, then shouldn't it match some information that comes from reality instead of new nonsense information that was generated by a faulty algorithm or input?>the kind you need to train a model in the first place.So if you train an AI on another AI's hallucinations, it will exactly match the real training data since there is no new information except what came from the real information?
>>17074489>Then why doesn't it match realityWhy should it match reality? >So if you train an AI on another AI's hallucinations, it will exactly match the real training data since there is no new information except what came from the real information?No, it will degrade because not only are you failing to generate new information, but also losing existing information.
>>17074491>Why should it match reality?Because if it can't generated new information, the information it outputs must have come from reality.>losing existing information.So x-y isn't a new number, you can't make new information by reduction, only addition, there can't be words for hairless or pigmentless animals, you can only add features even though no new ideas can be added with new words?
>>17074495>the psychotic brown sub-animal is samefagging
>>17074498>samefaggingAnother term you clearly don't understand.
Protip: if you're gonna blatantly samefag, at least try coming at it from a different angle instead of regurgitating the same incoherent nonsense. I know you think you're generating "new information" because you're using a different combination of words to express the exact same mental retardation, but it's actually nothing new and I can tell it's just you begging for another (You).
>>17074515Nope there can't be a different combination of words, you can't make new words with an alphabet, you can only have alphabet, there is no information besides alphabet in an alphabet, so its pointless to try to rearrange letters.
>retarded monkey drops the act and starts repeating itself verbatimlol
>>17074525>alphabet alphabet alphabet alphabet alphabet alphabet alphabet alphabet alphabet alphabet alphabetalphabet
>i mindbroke it
>>17074396God forbid we actually give people jobs and not become basement dwellers lol.>>17074409>least retarded AI bro opinion.
>>17074235I'm 44 I'm old enough to know AI when I see it and LLMs are AI. People who grew up with internet just don't realize how science fiction LLMs are because they don't remember what the world was like in the past.
>>17074700Classic sci-fi promised me thinking machines and the classic AI researchers were trying to make it happen. But they failed, MLfags took over and what I got is a token guesser.
>>17074706LLMs are thinking machines. They are likely limited due to having to serve and communicate with humans. But even that won't matter soon.
>>17074712>LLMs are thinking machinesLLMs can't think by definition.
>>17074715Semantics
>>17074719If you don't care about "semantics" (what words actually mean) why do you insist on claiming that the token guesser is thinking?
>>17074712>LLMs are thinking machines.Yeah? What are they thinking about?
>>17074241>and betterConfirmed indian
>>17074721Okay then, let's play this stupid game. What does it mean to "think"?
>>17074725Whatever you ask them
>>17074730>Whatever you ask themA probability distribution over tokens like "ch" is not a coherent followup to anything I ever asked a LLM.
>>17074727>What does it mean to "think"?There's no need to work that hard. I can tell you what "think" doesn't mean: "think" doesn't mean "let a PRNG-driven sampler choose your words for you".
>>17074733That’s still a definition, just worded as a negative. By that standard LLMs don’t think, again semantics.But Neuralink can already read the electrical patterns in someone’s brain when they think words and turn them straight into text or speech. The brain itself is doing a pattern-to-next-word process.So if that kind of process automatically doesn’t count as thinking, then a lot of what the human brain does when we speak starts looking the same.
>>17074750>retarded brown mumblingI accept your concession. If you don't care about "semantics" (what words actually mean) why do you insist on claiming that the token guesser is thinking?
>>17074752Oh no, not gonna ragebait your way out of this. You wanted to play this game remember?I am not claiming LLMs “think” (semantics) in the exact same way humans do. I’m saying the reason people give for why they can’t (“it’s just pattern-to-next-word”) also applies to what we can now measure in the human brain with Neuralink.If that process is disqualifying, then the argument needs better criteria than “I don’t like the mechanism.” That’s the point, not a concession that the word has no meaning.Now any real response or are you done?
>>17074771>I am not claiming LLMs “think” (semantics) in the exact same way humans do.They're not "thinking" in any way at all. They don't even pass a basic lower bound for thinking, nevermind the nuances. Your subhuman brown insect takes about "le humans also" will continue to be ignored. I consider them to be a sign of your rage and desperation.
>>17074771Just so you know you are responding to a spambot.
>>17074779>the literally who is still seething about getting BTFO in another thread
>>17074779Thanks
>>17074235Geoffrey Hinton believes LLMs are AI. I'll take his opinion over the opinion of a retard posting frogs with green text aka (You)
>>17074715Literally ever single time I prompt them it says "Thinking..." before responding unlike you who obviously just regurgitates the first nonsense that presents itself without putting any layers of thought or reasoning into it.
>>17074733Good think its choosing the words for itself then.
>>17075003>>17075004>genetic nonsentient mumblingMy point stands completely unchallenged. A probability distribution over tokens is not a thought and a PRNG sampling from it is not a choice of words.
>>17074777>They don't even pass a basic lower bound for thinkingThe lowest most basic bar is for one to claim one is thinking and AI can definitely pass that bar, in fact if you tried to do the Turing Test with a human and an LLM at this point, you would almost certainly be unable to tell which one was the human doing the "thinking" and which one was the robot just exhibiting "calculated patterns of output".
>>17075007No you can't even if think you can only see if the pattern matches your pattern of thought and agree with or or see if it doesn't match your pattern of though and dismiss it while saying something about mumbling even though mumbling is impossible in a written medium and you sound completely retarded when you say text is mumbled.
>>17075008>muh can't tell the difference>doesn't actually challenge the pointAdding this concession to my collection.>No you can't even if think you can only see...Losing it with rage.
>>17075010>>17075016
>>17075016>If there is no identifiable difference between thinking and what LLMs do, there is still a huge difference, trust me because I didn't dilute my claims with any of that thinking that just makes me angry and causes my head to hurt.>Losing it with rage.>projecting his own emotions again
>>17075018>>17075018You can prove to yourself that you're completely unintelligent by observing that you don't understand what's being argued, or even what's being argued about, but never once does it occur to you to ask a relevant follow-up question. The best you can do is mumble some generic talking point you picked up from OpenAI propaganda pamphlets trying to override a point you can't refute or even meaningfully engage with.
>>17075019>You can prove to yourselfI clearly understand since I can tell that not being able to identify differences between thinking and what LLMs do is like saying they are the same just like if you subtract two numbers and get 0, then they are the same number.>generic talking point You should be ashamed to know that having no difference between things makes them the same is a super basic point, yet still can't seem to understand that thinking and doing what LLMs do to generate information are the same because there is no identifiable difference.
>>17074329So why does it lie to me and needs to be corrected most of the time?>Oh, I'm sorry, you're correct. This key detail in my reply is false, I apologize.
>>17075024>I clearly understand since I can tell that not being able to identify differencesThis statement proves that you're completely mentally ill.
>>17075042>t. the schizo who is hallucinating differences that he has yet to point out after conceding there is no identifiable difference >>17075016
>>17075045Notice how your psychotic illness forces you to keep chanting the same irrelevant propaganda, meanwhile my point stands completely unchallenged. :^)
>>17075047Your point about how you can't identify any difference, but assume there is a difference anyway?
>>17075049Notice how your psychotic illness forces you to keep chanting the same irrelevant propaganda, meanwhile my point stands completely unchallenged. :^)
>>17075052No, the only one I notice copy pasting the exact same ad-hominem non-sequitur drivel over and over for propaganda purposes is the retard (You) copying and pasting the exact same drivel over >>17075047 and over >>17075052 because they have no argument, just fallacious propaganda to repeat over and over.
>>17074235>Human intelligence>look inside>Just ions moving across a lipid membrane
>>17075066Obviously, humans are just amoeba with extra steps.
>>17075007>A probability distribution over tokens is not a thought and a PRNG sampling from it is not a choice of words.basically /thread. or at the very least, it implies the LLM's "thoughts" don't really map to the humanlike lines of thought you get from repeated sampling. even if you suppose the LLM is thinking, each individual prediction step would have to represent a complete thought. let's say you run the model on an input like: "i want to fu". you might get an output like:>ck: 80%>gg: 5%>n: 1%>...so whatever "thinking" you think the model might've done to get to that, it has nothing to do with the thought "i want to fuck your mom", even if predicting repeatedly generates that output. saying a LLM thinks is like saying a slime mold thinks or an ant colony thinks. completely meaningless anthropomorphization
>>17075062I'm still not entirely convinced that's not a bot.
>>17074235>"AI">look inside>it's just a regression with extra steps
>>17075076So when it argues that LLMs can't think, you believe LLMs can't think because you know that it is speaking from direct experience.
>>17075074>saying a LLM thinks is like saying a slime mold thinks or an ant colony thinks. completely meaningless anthropomorphizationThat's obviously true but also totally irrelevant. The main conclusion you can draw from all this is that you don't need anything resembling human thought to solve problems. Not even human problems
>>17075080Slime molds and ants are better problem solvers than you are and saying that they are capable thinking is just a way to convey the fact that we observe them forming strategies to compete with others and acquire the resources they desire/require.
>>17075080It's very relevant when you realize that problems are circumstantial and ephemeral. Your house is chock-full of solutions no one will even remember 50 years from now. But what kind of intellectual heritage is the LLM psychosis generation going to produce? A machine that doesn't think isn't going to advance human thought. The Navier-Stokes fiasco is a perfect example of that. "Problem solved", too bad no one learned anything.
>>17075080>bviously true but also totally irrelevantlol. well, you should tell that to the mongs who work themselves into a frenzy every time someone states the "obvious" then. if it didn't matter the propaganda wouldn't be necessary
>>17075082Sit back down, retard.>>17075083>>17075083>But what kind of intellectual heritage is the LLM psychosis generation going to produce?Guess we'll have to wait and see.>A machine that doesn't think isn't going to advance human thought. The Navier-Stokes fiasco is a perfect example of that. "Problem solved", too bad no one learned anything.You're overly attached to human thought. I say, if you can't learn anything from 616,276 lines of perfectly good Lean, that's a (You) problem. Maybe human thought needs an upgrade. That's an engineering problem whose solution would make for some interesting "intellectual heritage" and I, for one, can see AI helping us there.
>>17075085Ants could probably figure out how to make me sit down, but you can't.
>>17075085>transhumanist mask slipIt always comes back to the anti-human agenda with AItards.
>>17074330Try again bud
>>17074235Nooo you cant just SAY that
>>17074422The projection here is astounding
>>17075114>broken chatbot guessers the tokens incorrectly
>>17074405very amusing nigg3r but have you updated your priors recently??
>>17074488I think this is one of those people the government hires to make 4chan less interesting so we leave it Or it's a thirdie
>>17075119>worthless subhuman projects
AI will make us all retards
>>17075136AI will rot only the brains of the users.
>>17075136>>17075897nothing will really change. the people it makes dumber are already on track, AI or not.
>>17074284It's useless junk and junk is pollution which is harmful and CAN kill youAre you getting it yet?
>>17076414Woah!!!!!
>>17074241> plow my field or I use a tractorIn both cases YOU are doing the work, just with aid of a tool. You still have to know HOW to plow a field, HOW to drive a tractor. AI replaces you and drives the tractor for you. Thus you lose the ability to farm.
>>17076459>AI replaces you and drives the tractor for youYou couldn't be more wrong. It drives the tractor for Them, not for you. And it's not your field, it's their field, which they will confiscate for their AI tractor.
>>17074284>AI is useless junk taking over useless jobs(scientists, most of women's jobs...)
>>17076459>You still have to know HOW to plow a field, HOW to drive a tractor.No, tractors have been autonomous for decades already, they didn't need advance AI to be autonomous anymore than the Roomba did since it is just basic back and forth maneuvers in a confined space.
a ham sandwich is just a search engine with extra steps (the extra steps are two slices of bread, ham , cheese, lettuce, mayo, and mustard)
>>17074235Exactlyhttps://www.youtube.com/watch?v=L2fnTlLPYUghttps://www.youtube.com/watch?v=CudIfoRiVYQhttps://www.youtube.com/watch?v=daWdQtSxgKohttps://www.youtube.com/watch?v=Zw0I7kb6zg4
>>17074291you have no idea how these models works, and it's a waste of time to talk to you about them because of it.Learn why your ideas on the subject are wrong, then come back with an opinion about it.
>>17074291Literally all invention is iterating and "rehashing" previous inventions. AI could invent the warp drive and you'd say it was copying something.
>>17076464Chinkbot detected
>>17074235>super smart human brain>look inside>its just neurons firing on and off
>>17077869>Its just a light emitting diode with extra steps.
bump
>>17074244This. Quality > Quantity
>>17081185Hdhsvo
>>17074712>LLMs are thinking machines. They are likely limited due to having to serve and communicate with humans??????????????????????whats the logic in that.>best man in the world for comunicating has troubles comunicating
>>17075085but anon you are not smart or humble enough to be a true transhumanist.shouldn't you stop talking about this subject matter and leave this to the experents?genuinely , maybe donate to them or something. I don't even care about AI right now , I want to insult you in particular.
>>17074712>having to serve and communicate with humans.But they don't have to do that, you can hook up digital sensors or any other digital information into their prompt, they don't have to be prompted solely by humans using verbal language.
>>17074235AI is overrated
>>17074235Guess what your brain is.
>>17074235>math graduate detected
>>17074235it's SUPER intelligence frogtranny
>"extraterrestrial life">look inside
