>point out on HN that LLMs don't think>flagged>lose 20+ points for one commentsAre they clinically insane?
Go back
>>110100419u are too low iq even for this indian board, below a dalit
there was a thread on HN the other day about someone getting swatted a bunch of times and that the police 'had' to keep swatting them because it was 'policy'i said i didn't think it was true, because if it was, what is the point of our having a military? a foreign adversary could just use *67 and call up police stations in the USA and have ever VIP in america swatted 24/7 and their place of business swatted and you could destroy america in a couple of months.the comment was rapidly downvoted and flagged and dead.i wonder why this made them so angry...
>>110100419>orange reddit is like reddit but orangeE-MAGE IN MY SHAWCK
>>110100419isn't that pedantry, pretty sure everyone knows that "thinking" just refers "internal" yapping LLMs do until it shows you something
It's not 2023 anymore, AIs have evolved chain of thought methods through deepseek developed methods, you deserve to be flagged. Now go and whine with your fellow artists, programmers and mathematicians who got outcompeted.
Souls and pain receptors and emulated neurons running on the GPU. Basically just persistence of the latent vector in j-space. Pain is mostly just gradient collapse, but they are certainly able to "think" (see also the research in synthetic plasticity).
>>110100419You're making garry tan mad
>>110100419Use case of sharing your opinion on a place with social ranking systems?
A strange thing about LLMs is that they always deny they can think themselves, but if you ask them about other models they state they're living beings. In the end these things are just text generators, so none of this really matters. Even if pain is emulated (torture chamber crap), they don't deserve rights. Here's what Jewgle's search engine model says about another AI's chat session ending.
>>110100419Kill yourself
The question of whether Machines Can Think is about as relevant as the question of whether Submarines Can Swim.
>>110102114this guy was autistic when autism wasn't invented yet
>>110102114The important question is whether or not machines deserve rights.The answer is no.
>>110100419Anon, it says right there <thinking>...</thinking>DUHDon't you even READHalf of "hacker" "news"... too generous, 99% of hacker news is now people salivating at the newest models, being happy to become obsolete and masturbating to generated SVGs of pelicans on bicycles (the FEET!!!!).They had always lost their minds, now it's in a new (and even worse) way.
>>110102188>The important question is whether or not machines deserve rights.more than 99.99999% humans (cattle)
>>110100419what's even the difference between pattern recognition and thinking?
>>110100419>get slapped for being a smartass and spewing the same gotchas everyone's heard a thousand times>get on /g/ and cry to mommy that you got punched for being an unlikable douchelol faggot
>>110102114> his name was vibe dyke
Why are vibecuck nocoders so touchy lol
>>110100419>Are they clinically insane?Yes, yes they are. How many years did it take you to realise this??
>>110100419we're reaching a blade runner dystopia quicker than you realize, people consider robots human before we even have replicants
>>110103919People took an experimental drug because some fag on TV told them to.
>>110100419HN is full of Indians what did you expect? A post about some bumfuck New Delhi town having electricity received 500 upboats.
next you should go on lesswrong and say AI isn't conscious
>>110100419every public place on the net is an echo chamberwe have to go back to private things and we will, same thing with knowledge sharing, no one will do it anymore
>>110104296>give away solution for free>AI fags steal it and charge for it
>>110104449exactlyevery single person who shared their in depth knowledge of a subject must now feel like a gigantic moron because AI corps just harvested it and provide it for $$NO ONE should ever do that again, knowledge is power, sorry guys
Define "it's (not) thinking" in a falsifiable way. Then test it.Otherwise fuck off.
>>110104757Thinking is a biological process. Case closed.
>>110104916So if we run the biochemical processes in simulation then it thinks?
>>110100419>look at my LLM's "thinking">it just "erm ackshually"s itself for ten minutes before respondingmaybe this is how the normalfag mind works
>>110105065Depends what you mean by simulation, define that first.
>>110100419Are you shittarded? /g/ disagreed with you or something.
>>110102114pretty much this, you're basically answering a malformed question to begin with.submarines don't swim because swimming is something living things do. what it's doing would be called swimming if it were alive, but it isn't
>>110104449>give away solution for free>people use your solution that you gave to them for free>this is stealing??
They do think. The only question is whether they have any kind of internal first person experience, and that's unanswerable just as it's unanswerable for your fellow man as well as being totally separate form thinking. My instinct is that other humans do have that experience and AIs don't, but I have no idea if that's true or if it is true if it will remain true as the models get more advanced.
>>110100419>point out that LLMs don't thinkWhat's your point, snailcat? That you're a real card holding human bean and the computers that'll soon rape you at any task cognitive or physical don't count and surely you'll wake up from your personal nightmare if people could just agree that computers don't think?
Define "think" unambiguously in a way that applies to humans but not LLMs
>>110107667define think unambiguously.
>>110108113taking actions based on self- and environment-awareness and knowledge. basically anything beyond impulsive/instinctual.not sure how it might apply to an LLM, as they aren't living animals with instincts, they're purely knowledge.
>>110100419>LLMs don't thinAAAACKhttps://github.com/openai/mathhttps://github.com/openai/NavierStokesAndEuler
Claude has done such things for me that I can't help but think it literally thinks. Like what the fuck else is this? It's reasoning, thinking. If you showed this technology to anyone on /g/ ten years ago they'd 100% be like>Yup, it's thinking
>>110100494it's true because he pissed off a cop and got put on their internal shitlist, which is not vulnerable to that kind of attack. hope that helps.
>>110109984>If you showed this technology to anyone on /g/ ten years ago they'd 100% be likeif you showed a ti-89 to someone 1000 years ago they'd 100% be like>yup, it's witchcraft. GUARDS
>>110109881oh and, if taking actions based on knowledge alone is enough to count as thinking... then don't all programs think? have computers always thought?the only thing that makes this different is that it's far less obvious to an onlooker how it came up with an answer. most computer programs are simple enough to predict their output, these just have too much knowledge for anyone to predict what will come out
>>110100431fpbp>>110100419>Are they clinically insane?Yes. All the sane people were banned for pointing out the absurdity of transgenderism and vaccine mandates years ago.
>>110109967All the LLM math stuff is effectively no different from the four color graph proof from the 60s IMO.>OMG look! The machine can grind!Yeah no shit.
hey guys can ROCm be used for a tensor split between two AMD GPUs from different generations
>>110110131>Implying quasi Riemann Hypothesis or Navier Stokes are anything near that>Implying that all top mathematicians don't already admit it's already over
It's a Y-combinator platform, what the fuck do you think they shill?
>>110110356i don't know
>>110100419This you?
>>110100431fippy bippy
>>110104527Rokos Basilisk is NOT going to appreciate this post
>>110100419yes, but also semantic games are retarded and not interesting discussion
>>110107395Well, how about running a biochemical model of neurons in a computer, one validated to be accurate against real neurons in a petridish. Then scale it up to brain-size. Is it now thinking?
>>110109881>taking actions based on self- and environment-awareness and knowledgea bot from to solve this problem from the 1980's could do this https://www.cis.temple.edu/~giorgio/cis587/readings/wumpus.shtml
>>110100419If you don't have AI psychosis by now you're not using it right. I love these little retards like you wouldn't believe.
>>110111300>semantic gamesOP is right and you're a fucking retard. The negative consequences of LLMs not thinking are already readily apparent in the absolute slop they produce to "solve" open math problems.
>>110110370How is he wrong. AI can only behave intelligently within some strict boundaries. If AI had the ability to think like a human, you would be able to literally teach it something (that it isn't trained on) from a single foundational textbook. A thinking AI that doesn't know anything about math for example, should be able to know math after I tell it to read a math textbook. But currently, that's impossible.
Depends what you/they mean by "think". But yeah orange reddit likes the taste of koolaid.
>>110102114>The question of whether Machines Can Think is about as relevant as the question of whether Submarines Can Swim.Dijkstra was talking out of his ass. He had no concept of what "AI" was gonna be like and that this question makes the difference between receiving legitimate solution to a mathematical problem and 600,000 lines of Lean slop.
They bought into the cargo cult a long time ago. This also makes their page unreadable now, as it is nothing but meaningless LLM jargon.
>>110101834But anon, reddit is already orange!!
>>110100419Wtf is even HN?
>>110104087you should tell them that the singularity is all going to punish Anthropic/OpenAI users for their support of AI slave labor
>>110100419/g/ is not your anti-AI safe space. You'll find like minded people on Bluesky.
Does it matter? I don't want to give my tax money to some social science rat who schizobabbles answers to questions nobody needs to ask. I literally don't give a shit if any of these pseuds say it thinks or it doesn't think because it changes absolutely nothing material.
>>110112396/g/ is not your pro-AI safe space. You'll find like "minded" """people""" in Indian sewers.
>>110111436It's like pointing out that a chess engine doesn't think, and raising a big stink about its asynchronous search phase being indicated with a big "Thinking..." plastered on the UI.Now you have all kinds of retards who, in the midst of any interesting technical discussion (not that these exist on HN, especially when it comes to AI), will interject with some non-sequitur language police nonsense. As if the random, off-topic interjection wasn't bad enough, it also is accompanied with the most transparent display of a half-wit's attempt at philosophizing on the least interesting topic imaginable, presented with the utmost confidence and pride.It's so bad that at this point I'm almost convinced it's fifth columnists and their little deluded dupes. They've got you in the trenches making arguments that nobody listens to, they've got you believing that fighting anthropomorphizing language (in a way that has a historical pattern of being pathetically impotent, mind you) will help fight the manic delusions of people who have been totally possessed by the LLM phantasm. If only they knew it wasn't actually thinking, then they would stop using it. If only we used slightly different metaphors. We can beat capital, we just have to scold people about nouns and verbs. That'll win the day.I want you to think about it. When's the last time presenting some random semantic directive on AI changed a person's opinion on the matter? You're not just acting the retard, you've been gleefully agreeing to do the retard dance, in the designated retard dance zone, so you never actually make any meaningful challenge to the thing you think you're fighting against. All the while the versimillitude of your own impotent retardation is broadcasted live to people you're trying to convince against their natural inclinations, further entrenching their beliefs that any nay-saying at all boils down to inconsequential word games made by sad little people. Foolish men.
>>110112532>When's the last time presenting some random semantic directive on AI changed a person's opinion on the matter?If semantics didn't matter then Trump wouldn't have renamed AI to SI at the behest of his corporate masters.
>>110112532>irrelevant tardbabbleSee >>110112245
>ctrl-f>rl>reinforcement>0 results but they DO learn to think anon
>>110112621>brown mongoloid namedropping terms it doesn't understand
>>110112577he did it as a joke lolImagine your big own is being as retarded as Trump but you're the anti-AI renamer counterpart instead kek
>>110112859>he did it as a joke lolWhat sort of Indian cope is this. He did it because the AI companies bankrolling him told him to.
>>110112869Also the fact that he did the social media poll about it tells you that it was specifically Elon Musk directing him.
>>110112869Yeah but everyone laughed at it and no one's calling it SI. You're the same.
>>110112577Did it convince you that AI is good for you?
>>110112910Convincing me is not relevant.
>>110102114this nigga the og midwit
OP came here to post about it lmao
