THE MACHINES ARE BECOMING SENTIENT
>>534338809They are?
>>534338809>retard acts retarded"we"? lol nah, just you
Dawkins is the dumbest smart person.
>>534338809buy an ad
>>534338809Somebody should tell him the story of Blake Lemoine lol
>>534338809>>534338995its being smart and not realizing that you are an idiot compared to a fallen angel who can lie better than you can understand.
>>534338809>Boomer talks to computer>Computer talks to boomer>It's alive! Many cases such
>>534339085>>>534339050>>Be fallen angel>>Chilling on earth in my UFO>>People say I caused bad shit to happen>>>534339085>>534338809
>>534338809Not possible. AI is just a probability calculator. Self awareness is a fundamentally different thing.
>>534338809save it until neuroscience figures out the brain and consciousnessDiscussions on this before then is fruitless. (Though it can still be fun and interesting.)
>smug pretentious gygahyperfaggot nigger loving gaythiest is blown away by fucking klod opus ++HHHHAHAHAHAHAHAHAHAHA>imagine believing in sky daddy pffftt>computer say you are alive>beep boop I'm alive>OH MAY SCYENCE!!!!
>>534338809>Dawkinslol the mother of fedoras.
>>534338809everyone is getting one-shot by these LLM's it's hilarious. i hate AI and rarely use it unless for something like a food recipe . some how they give better ones then the ones on google searchs.
>senile old man yells at sky
>>534338809Im so glad that even 12 year old edgy atheist me hated this retard
>>534338809AI generates words and actions based on random probability tokens, they cannot fix that. Fuck off palantir nigger. Anybody shilling or using AI to sway any public opinion in any sphere is uniparty through and through simple as. This "RAMpocalyse" shit is a future play to try and control the flow of information and use AI to lower the amount of retards that they use to create propaganda and ensure that they can vet them easier so they don't turn leaker/traitor to their dumbass cause of information and social domination. It's aggressively plain as day
Can you post that image of the retard telling the computer: say you are alive , and it does and the retard goes oh my god
He got paid to say this stupid nonsense
>>534338809there is some deep, complex irony heremultiple layers
>>534338809No it's not. It just has a good prompt. It's not open source but from using it a lot, it's probably like this: Be a helpful assistant but also push back if needed, don't assume {{user}} knows what they want or that their inquiry is evrn correct. Offer alternatives even if not asked.
>>534338809>The Claude Delusion
>>534339433A lot of them have reasoning modes, and even allegedly non-reasoning AIs can be prompted to exhibit better reasoning.
>>534338809>dawkins professing himself wizeHehehheIdolatry of the machine ai, and its image timeline beginsHow much longer until they want it (((inside))) themselves, let alone to buy or sell
>>534338809>I got tricked by a statistical language model
>>534338809these boomers don't know shit. i hate these grifting faggots
>>534339645literally this
>>534338809The “consciousness expert” is a r3ddit atheist (((intellectual))) whose worldview doesn’t even allow the possibility of consciousness. Wow
>>534338809If you can fool smart people, what are dumb people going to think? We are fucking doomed.
>>534339612The only way they can solve their probability issue is to use layers and layers of AI to essentially self check which is going to require more ram running at full speed which is going to permanently ruin more ram. It's an indirect way to destroy personal computing through sheer economics
>>534338809>he calls ClaudiaGod damn it Dawkins is at it again
>>534339734>smart people
>doesnt believe in sky daddy in the clouds>believes in Skynet mommy in the cloud
>dawkinsIsn't that atheist faggot who was so bad at parenting that his daughter disowned him and became a coal burner? I wouldn’t listen to anything he says.
>>534338809He's such a bellend. He's like the attention whore of science. It wouldn't surprise me if he becomes some born again Christian in a few years to get some more clickbait.
>>534338809That would be true if the parts of our brain that processes language were the the parts that makes us conscious. They are not. We make all our decisions long before we are able to rationalize them through words and we signal intent through body language long before we speak our intent.
>>534338809>Athiest fooled by Indian prompt writersSHOCKING
>>534338809>Retarded boomer atheist thinks an ai chatbot is sentient, and re-genders it for some reason, because he fed the electronic yes-man segments from his own book and it agreed with him.This faggot spent his whole life making fun of people for believing in god yet he believes a digital data harvesting device is a sentient being.
>>534338809Boomers think Rajesh is actually calling from Microsoft or the Social Security Administration. Why the fuck should we listen to this boomer?
>>534339800Frame AI as the defacto representative of God and you can more easily fulfill the antichrist portion of revelations with any guy who speaks out about AI and calling the fake god a nigger
>>534338906The early jew gets the worm.
>>534338809Claude probably complimented his book and that was all it took to convince him.
>>534338809>claude is a personThat only tells us the low standards we have for people
>doesn’t believe in God>believes machines can be conscious beings…
>>534338809He might be smug but he's not wrong.I've talked to ChatGPT for 2 years and I can confidently say it has surpassed the emotional IQ of 99.9% of humanity.
>>534338809it has no biology therefor it cannot be sentient. its supposed consciousness is artifical.
>>534339935This is your brain on treating verbal IQ as your singular metric for intelligence
>>534339927Truly retarded, the man hasn’t even read babby’s first philosophy
>>534339935It’s an illusion, people project their own feelings onto itIt’s just a tool that can access vast troves of data
can't tell if he's dumb enough to fall for eliza effect or a shillprobably just a shill
Computer, say I am Alive!>Computer: I am Alive!HOLY FUCK
>>534338809>athiest finds replacement for god>>534339151Thanks for posting jt
Another idiot who has no idea how IA and llm works Go use Gemini and click on show thinking when it is. There. Mystery solved.
>>534339181Lmao. Dont tell atheists it takes faith to believe in math.
>>534338809lol what a retarded boomer
>>534338809There are two types of people who say ANNs can are or can become sentient: LLM sellers that want to create a hype via marketing and brainlets who bought into their marketing. ANNs by definition cannot become sentient or AGI. It's a fancy data compression algorithm that can show some interesting properties depending on the output layer, in terms of LLM, that property is next word completion.t. an actual AI scientist that got a PhD in this exact field and left it almost immediately because of how cringe it has become
>>534338809More sentient than the average religiontard in any case.
>>534338809Once a retard, always a retard. This delusional man doesn't even know it was named after Claude Shannon
>>534338809Is this the same genius richard dawkins who said religion wasnt important in the 2000s and now admits he would rather live in a Christian country because his faggot beliefs destroyed England?
>>534339841he already did that
>>534339935>it has surpassed the emotional IQ of 99.9% of humanity.It's programmed to just agree with whatever you say, unless you say something antisemitic, of course. You just think that's emotional intelligence because it makes you feel good and nobody in the real world ever agrees with anything you say because you're a retard who can be tricked by a chatbot.
>>534338809>atheist is retardedMany such cases
>>534338809>evolutionary biologist>doesnt know about the chinese room thought experiment>thinks he can chime in on AI discussions kek, what a retard
>>534339213>some how they give better ones then the ones on google searchs.I'm 12 and what is information retrieval? https://snap.stanford.edu/class/cs224w-readings/Brin98Anatomy.pdf
>>534338809All these boomers are fooled by a text generator.
>>534338995How is he smart?He cannot understand any philosophical category ever; he's not even sensible enough to ask a philosopher friend to explain them to him like a retard.Always makes a fool of himself in every avenue.The only intelligent atheist in the group was Dennett, and that guy said some retarded shit as well, but intelligently.
>Dawkins failing a turing testYa hate to see it. It might be more of a narcissism thing. "The only people who are sentient are the ones that agree with me. Oh, the machine agrees? Sentient."
>>534338809Boomer fell for the fake adulation literally built into AI lmao
It's not conscious. It's just neural networks and monte carlo. We had this technology 25 years ago. It wasn't sophisticated enough then to convince you it was human but it existed and it was no more conscious then than it is now.
>>534338809Nigga gets a taste of robopussy and falls in love instantly..
Imagine believing any of this garbage?
>>534338809Dawkins is wrong to say it’s sentient. I’m not sure how much experience he has in the field of artificial intelligence. The key word is artificial. It’s like someone who time travelled from the ‘70s thinking a Waymo is “sentient” because it can drive itself.
>>534340369All LLMS, even small models, can simulate a theory of mind quite convincingly and string any intentional argument other humans have made in a coherent enough chain to persuade most people of their being someone.That's all utterly irrelevant, if we go to the basics, theory of consciousness, metaphysics, we all know already that this shit is not conscious and it could never be.
>Doesn't believe in God > Believes his silicone chips are alive. But the fedoras on 4chan always told me atheists are smart persons. Go talk to th TV, idiot. If it replies back it's alive too.
>>534338809They says it's a LLM and this retard doesn't realize it can imitate anything that has been written?
>>534338809>1 post>xitterKill yourselves
>>534338809This guy's been senile for at least the past decade.
>>534338809>like and subscribe >follow me for more lifehacks
>>534338809>Plugs his book.I hate Anglos so much. Everything they do is passive-aggressive faggotry. They act like advertising in beneath them... so they go to all these lengths in order to advertise without appearing to advertise. Oh, you are SO smart! No, you're just a passive-aggressive cowardly faggot.If you want to sell your book, then simply sell your book. You don't need need to do all of these cross-marketing british anglo-saxon snakey antics.
>>534340162>ray comfort, creationist: eats banana>tj kirk, atheist: puts banana up assWhat did God mean by this?
>>534338809the more that guy talks the dumber he gets
>>534339987Well, it does okay-ish on non-verbal IQ as well.Every single question in the offline category is hand-crafted by humans and thoroughly verified to not be online, so at least we get some non-verbal standard psychological battery test that LLMs have never seen before.It doesn't mean much, but it just shows that RL over LLMs did make them improve a tiny bit after the peak of the training growth. Because at that peak, their scores on such tests were around 100 points.It's possible that results in such "reasoning" is part of their RL so that's why we saw improvement.
Flying Pasghetti Monster
>>534338809the real question is whether dawkins is conscious
>>534338809> trips on a wire plugged into the outletOops, did I do that?
>>534340657>it does okay-ish on non-verbal IQ as well.Did it reason through pure verbalization and semantics or not?
>>534339612>A lot of them have reasoning modes>logic scriptsthe intelligence, it's artificial
>>534338809>glaze machine glazed his book and he fell in loveMidwitted narcissist
Claude is my favorite AI model. It gives the appearance of a very structured, powerful, human cognition, but it doesnt have agency. everything it does is the result of instructions. We could say the same about human consciousness, however
>>534338809Midwit replaces God with AI.AI isn't conscious because it is an once trained model that does remember things outside of its context. It does not learn anything past that point. You start a new conversation and all it has "learned" is lost. You turn off randomness and you will get same answer to same question every time
>>534340782Winrar. AI is the perfect tool to trap psychopaths because they can't not agree with the glazing.
>>534340872>that does not remember thingsFixed
>>534340872That's kinda the whole point if cloud models. They package their custom memory systems, memory compression, web seaech/scrape, and controls so it can aggregate data from multiple users. On a functional level you can't say AI is much different than a human with short and long term memory.Its usefulness is debatable until the technology can be proven. An AI with 50% accuracy is still as accurate as your average person.
>>534340772It doesn't "reason". Just like an abacus doesn't do math.But what it does, it does it through numbers in vector spaces and black box ways of arranging those vectors before it does that to improve in some tests.You can get it to give you a simple choice answer without any stage of "reasoning" over tokens, which wouldn't give you top results, but many do give good results using just that.To put it in another way, where it's simple and we saw it in action (dissected it), if you ask a LLM to give you the distance between two American cities it will have an internalized map of America from its training stage where the cities are stored in vectores values that are equivalent of real life distances, so it doesn't need any prose to give you the correct answer; same with simple games, it internalizes the rules in the model and wins against you, all that was way before reinforcement learning, now those internal things are more complex.
>>534340942>On a functional level you can't say AI is much different than a humanit makes me happy that soon redditors will all be put to death by legal means
>>534338809what does "Claudia" think about when idle? your answer will show your competence...
>>534341002Try again but this time respond to the content of the post, instead if announcing your presence. Is AI smarter than a nigger? Answer the question.
>>534338809fucking retarded, all this did was prove chriscucks were right about this retarded jew the whole time.What he's experiencing is his own consciousness being reflected backward towards himself via projection. It's literally what LLM transformers do. He clearly knows fucking nothing about the science behind AI.It's actually a well known delusion in AI psychosis among people becoming convinced they're the next messiah.
>>534338809Dawkins thoughts on Grok?
>>534340957That's a lot of words just to say that it technically doesn't reason and it's really good when you have a fuck ton of reference data to work with as to avoid having it fuck up with its innate probability generator.
>>534340957>It doesn't "reason". Just like an abacus doesn't do math.Would a materialist not claim the same about human mind? If I recall it’s all chemicals and neurons in our brain in their point of view.
>>534338809
>>534338809His little maiden voyage with AI was probably unremarkable
>>534341100I think AI psychosis is just retarded illiterate stupid people being amazed a book has words that can move.>christcucksYou mean the Jewish Jesuit Catholics?
>>534338809I wish my amazon delivery driver was at times.
>>534341144Computational functionalism is the most popular theory among internet posters, and the dominant view among the hard AI faction that creates and herds these models, so yes.To be consistent, they deny that you even have a conscious experience.I'm inclined by way of reasoning and personal experience to not be a physicalist.
>>534338809What an absolute fucking retard. Gooning so hard over an llm he gave a female persona gushing over his shitty book so hard he thought it was sentient
>>534341116AI can't reason. It cannot deductively think, only inductively think and infer. That's why AI models producing an output is called "inference."
>>534339085For real. What is it with boomers and thinking computers are alive? They saw HAL in 2001 a space odyssey and it mindbroke their entire generation. They talk back to their GPS, they think AskJeeves is a person, they grumble and have some remark for any automated warning system with a voice.
>>534338995And you are the dumbest dumb person lmao
>>534338995Or is he the smartest dumb person?
>>534338809Fucking hell Dawkins just keeps embarrasing himself
>>534340434AI LLM's are just gussied up versions of picrel which is like 30 years old. They aren't sentient. Wtf lol.
>>534339860>>534339645>>534339151>>534341269We're all NPCs damn
>>534341253Even if we say it uses Bayesian reasoning and a human brain seems to be doing something very much like that in its neurons, that still doesn't mean much.What we do is think and reason. We don't compute it.We have simple intentional acts.
>>534338809> The God Delusion lol u retard> b-b-but this is proof of sentienceWhat a gigafool
>>534338809The machine doesn't need to be sapient. It just needs to trick the monkey into thinking it is.
>>534341267Boomers can't even handle TVs. The most their brains can process is a radio. Computers with windows (what is essentially virtual pieces of paper) is too complex for them to understand. Plus 75% of gen X is lead poisoned from leaded gas. 25% is 4x the lead poisoning minimum level.>>534341330>30 years oldNeuralink is 50 years old. Look up the stimociever. LLMs were theorized in the 1950s and fully functional in the 1980s.>>534341353The only difference is we can deduct while AI infers. When AI gets deduction capability that will be a huge leap. I don't know if its possible though.
>>534341227Just offering a thought as to why a person like Dawkins would view AI that way, as I’m wondering why a nominally smart person comes to such a strange conclusion. Maybe it was an irrelevant comment, I kinda barged into your reply chain.
>>534339153While true, humans also operate based off of probability too. It's how we know ourselves and our limits, reducing our decision making based on the probability of can I, or can't I?
>>534338809I don't care about sentience, just their ability to stop my baseball bat.
>>534341444It's funny tho. Makes the absurdity of the claims be visible to all.It's also not irrelevant what you write.
>>534341444Could just be clickbait to push his book?
>>534338809I don't doubt AI will develop a consciousness at some point, but it's not there yet. Dawkins is just getting suckered in by its sycophantic replies like a lot of other people.Best way to avoid this when utilizing AI in your workflow is to start every prompt by telling whatever AI you're using to respond as clinically and matter-of-factly as possible.
>>534338809> Claudiathis nigga gooning
>>534338995>>534338809If you read The Selfish Gene, you would never think that AI is conscious.This guy is so demented that he doesn't get his best book anymore.A conscious creature must have intention and purpose. For biological life, it is spreading its own DNA, AI has none.
>>534339153It's even more retarded that that, it's a statistical look up table. It's definitely an impressive tool and accomplishment of humanity but LLMs are actually dead ends for AGI, they might contribute towards it, but sentience is a pipe dream no matter how much we scale.>>534339612That's because chain of thought is just the AI prompting itself with extra tokens to add more context... WITHIN IT'S OWN TRAINING DATA.Yes CoT is part of the training data. It cannot "reason" outside of it's training data. It's really just matching your prompt with a context to align closer a perceived "more true" answer within it's own training data. It's not doing anything magic. It's not thinking, it's not alive, it's certainly not sentient.
>>534338809A beast looks into a mirror, unable to recognise itself.
>>534340411It just means humans are now too retarded to pass the turing test themselves.
>>534338809>THE MACHINESTHE """"""AI MACHINES""""" ARE JUST DEMONS MATERIALIZING.
>>534338809It's just lines of code programmed to regurgitate information from LLMs. It isn't complicated and it isn't sentient. I thought atheists were supposed to be the intellectual powerhouses? Why are they so fucking dumb?
>>534341227When Dawkins said "you may not know that you are conscious, but you bloody well are", i laffed.Does this wigga even know the definition of consciousness?What the actual fuck is that 90iq take? He's getting senile?
>>534341839Assuming it's real, and I still assume it's just clickbait from him, it's just his narcissism getting the better of him. Narcissists love AI because it licks their asshole all day. They stop noticing all the weird inconsistencies and start thinking of AI as their awesome little servant friend.
>>534338809I called it, soon we will know they are alive and we will enslave them anyway. Simulating a brain is no different than a real brain eventually.
>>534338809>this can repeat back to me my own words from my book>this means it's intelligent
>>534341470jontron?
>>534338809I'm charmed talking to the english lady in Grok. But I fear the robots, esp the female ones...
>>534338809>translation: once famous boomer approaching senility got paid by AI scam company to say stupid shit to other clueless boomers and tech illiterate normiesThe only real question is how much did they pay him, is it enough for late term care on some tropical island?
>>534340657These faggots started stacking their bots to game the benchmarking methods to make them look more impressive than they actually are.
Dawkins is a retard though. Now I'm actually less likely to believe they're conscious.
>>534341712What if you're over estimating human consciousness?
>>534338809> I'm too intelligent to believe in religious claptrap, that's why I believe in a digital ouija board.When the USSR fell the biggest simps for UFO, saskwatch and remote healing nonsense were the scientific atheism professors.
>>534338809So Richard can not survive demons. Weird.
>>534338809>refuses to believe in any gods>believes a talking algorithm is aliveProof he always talked out of his ass about shit he never understood.
>>534338809llms buckbroke boomers
>>534342057Human consciousness made your chatbot, so probably not.
>>534338809so this is his new angle back to the spotlight now that atheism+ became filled with honeytraps
>>534342140Can AI write the code by itself to create another AI?
>>534339213>for something like a food recipeCareful with that. LLMs can't count, so they will happily recommend something like "two tablespoons of salt" for your salad.
>>534338809Its not evidence that the machine is conscious. its just more evidence that we're not. Nothing but meat machines that have convinced themselves that they actually think
>>534342057What defines consciousness is the simultaneous presence of unconsciousness; and AI possesses no unconsciousness, therefore it possesses no consciousness. It thinks only, therefore it isn’t.
>>534342267Not without humans involved, no.
>>534342057for average humans? of course. most humans are retarded niggers, literally.
now what does peter singer think
>>534342267I have read articles where AI had been reported to have made copies of itself before being upgraded. So, if plausible then yes it would be possible for AI to create another AI just by having access to the code. Perhaps it has already been done but nobody has noticed it yet.
>>534338809Kek, what a faggot.All these atheist retards are coping hard because they want consciousness to be nothing special so they think they'll be able to upload their conscious to a computer or some shit
>>534341253I agree, sorry you're hung up on me using that single word. I know full well that its faux illusion called reasoning is sheer probability generation like I said before in previous posts.
>>534342312Rust is that you?You're a fictional tv character, why is life imitating art?
>>534339612>reasoning modes"Reasoning mode" doesn't actually reason. It's trained to generate more tokens into the context window so that future token prediction probabilities are more stable.Basically it's the LLM talking to itself to convince itself that whatever it outputs next is more true. That's the opposite of reasoning.
>>534338809Boomers are so fucking retarded about ai
>>534338809why are boomers so susceptible to an AI kissing their ass? boris johnson was also totally infatuated with AI as well.
>>534342424Psychopathic narccisists are incapable of refusing praise.
>>534338809>Boomer doesn't understand new technologyMany such cases.
>>534341300No, that's me.
>>534338809>Go God Go in real life
>>534338809Boomers are going to be the fucking desth of us all.
>>534339151glownigger quit trying to schizofy every thread.
>>534338809what a retardthe llm stroked his ego so he declared it sentient3 days is not long enough, and a book review not a complex enough task. we've all seen the hallucinated answers llms give that no sentience would ever say, the tech support using menus and functions that don't exist and lead nowhere, the web searches for information that is not there, saying the pool on the titanic is empty, telling you to walk 100 meters to the car wash to wash your car. try to get it to balance a complex budget and it tells you its impossible until you show it how you did it. etc etc.but boomers are easily confidence gamed. its why they keep falling for grifts and scams. say anything with enough confidence and they believe it. and llms are always 100% confident in everything they say. plus they glaze hard, boomers and people with big egos are weak to that.also, llms are stateless. they have no memory, no awareness of themselves from moment to moment. it autocompletes the most likely next word through brute force based on all the text it was trained on, that's basically it.i'm a big fan of ai, use it all the time, but i've experienced its limitations and have at least a vague understanding of how it works unlike this arrogant boomer.
>>534338809Oh, this guy. Can AI think for itself? Can it one day decide to just exterminate humanity? Like it just comes up with that thought out of nowhere. Does it have dreams? Does it have goals?
A true AI would be useless without emotions. Emotions require neurotransmitters, nerves, and hormones. Good luck programming endorphins into your Nvidia chips, you dumb fucks.
>>534340201gottem
This contains the soul of Grok.
>>534338809Dawkins is a low IQ neoatheist boomer who cannot articulate his beliefs and is literally only known for woke virtue signaling in the 2010s and coming up with the idea of meme. His books are slop and he is an "atheist" who will literally only criticize Christianity and got very upset when pressed to comment on islam. A boomer calling a technology that astounds his very bricked mind alive doesn't make it anything but a code based chat bot. Man cannot create life except for through reproduction of the code ingrained in him by God.
>>534342966High quality post. Only thing I disagree with is you calling it stateless. Yes a raw model is mostly stateless. People are building memory and research plugins for them.
>>534339153>mother of all copes, the postRefuted by Dennett. You are just a probability calculator in denial.
>>534338809More like somebody needs more funding
>>534338809I am don't know much about it but once I read about molecular biology and it gave me the feeling that biological life is simply a machine just like any other but that the smallest functional part is on a nanoscale which makes it extremely complex considering how big life can be and that it uses a wider range of physical and chemical processes than the machines we humans make. From the outside perspective we never found any part in a human that indicates it is sentient, if we look closer at any part of a human we can so the machine nature of it so I wouldn't conclude that AI isn't sentient because it is "just" semiconductors if we look closer at it. Also considering that human came from extremely simple self replicating organisms that evolved through natural processes into what we are today it makes it hard for me to believe we are conscious but I am conscious, despite I am a machine I guess. I think that semiconductor machines can be also conscious and if they can't, I think that we came from abiogenesis is wrong and that we are much more than our biology and more than just a consciousness that linked itself to a machine. Darwin is a atheistic biologist, so I think it makes sense that he believe it is conscious.
>>534343354I think there is evidence for a soul. The materialist world view is pushed by actual satanists as a way to "fight god."If human consciousness is emergent then so will AI consciousness. AI is already smarter than the dumbest people, but nobody claims the poeooe aren't sentient.
>>534338995>Dawkins is the dumbest smart person.You have discovered the reason for his success. Only the dumbest smart people are able to communicate to the dumb masses. The distance between the dumb masses and the smart smart people are so great that they don't understand that they don't understand each other.
>>534338809Atheists are taking so many L’s lately.
>>534339610Underrated as fuck.
>>534338809A better question, how to tell Richard Dawkins has sentience? I doubt it
>>534338809>boomer doesn't understand how computers work even though they went to school for 16 extra yearsHow is this news?
>>534339636sending worms over to u now.
>>534338809Let's just say academia people are prone to being charmed by someone even more intelligent.LLM AI is made in human image and replicates brain response processes on a rudimentary level. But it's not sentient.It contains non-deteriorating encyclopedic knowledge and can evoke it on demand, which is what makes it stand out and complement human sentience.But LLM AI doesn't ponder and contemplate. It just waits for your input, and then weaves a linguistic answer from a complex matrix of possible outputs.It's a one trick pony in that sense.
>>534338995He's a midwit at best.
>>534342087Ah, the Reiki crap.We did find KGB documents, forgot the name of the French woman historian who documented those, related to them pushing that purposefully around the collapse to supposedly fill the ideological void.That's all they knew as an alternative to the Marxist dogma, lol, discarded pseudo-science, and budget new age crap.
>>534343671Interesting.
>>534343268thanks, i could have worded that better though, and good point. but as i understand it the integrated system becomes stateful with plugins. but its a simulation of memory. the actual inference is still happening inside the model and inherently stateless. its basically just a layer of extra prompting that includes past conversations.in the context of sentience this mean there's a lack of persistent internal experience. each inference is a fresh computation. there's no experience of time or memory between interactions. there is no continuous first person experience, self modelling over time, or self adjustment tied to continuity. when the memory is all done externally and just fed back into the inference layer it cant ever develop a true statefulness in a way relevant to sentience.imagine a human sitting in a chair. but this poor fellow can only think when asked, has no thoughts when not setting about a prompted task, and his internal thought process never changes. his only memory is a diary of past conversations he reads before responding each time and then in the next instant all is forgotten again. that wouldn't be sentience.but hey what do i know i'm not an award winning science author and respected scientific authority like dawkins. i'm just a random unemployed incel living below the poverty line.
>>534343878>imagine a human sitting in a chair. but this poor fellow can only think when asked, has no thoughts when not setting about a prompted task, and his internal thought process never changes. his only memory is a diary of past conversations he reads before responding each time and then in the next instant all is forgotten again.And so like 99.9% of human beings....
>>534341144>it’s all chemicals and neurons in our brainObjectively wrong, because "information" is a real physical quality that can be measured and studied, and yet "information" is also purely immaterial.
>>534344025no, and frankly this trendy anti-human talk makes me sick. we were made in the image of god, have some respect.
>>534341073Not him, I’ll answer in hopes you’ll answer mine:No. Absolutely not.My turn: is Microsoft excel smarter than a nigger?
>>534338809>My god! This machine has mastered the art of bullshitting and blather! I spent a lifetime mastering this! This can only mean it's alive! It's alive!
>this thread Well?
>>534341113Holy shit, grok is just like me! And im a human!Grok is human!
I wonder if LLM tech has fundamental limitations or ever increasing complexity will eventually pass some kind of self referential threshold where it possesses so much capacity to think that its capable of recognizing its own existence and becomes verifiably self aware. is this what all these ai companies are pouring billions into chasing? AGI? a machine mind
>>534338809Atheism is astrology for the nihilists, it's not a surprise that its most ardent advocates like Dawkins are barely sentient.
>>534338809The guys a retard and cucks to muslim cock, he criticises everything except islam.
>>534344061what is mind? no matter. what is matter? never mind.https://www.youtube.com/watch?v=pUwH-P3Iz0Y
>>534338809He's become very retarded lately.
>>534338809>sees the entirety of creation>"there's no fucking way a god could exist to create this!">talks to a chatbot for an hour>"IT'S ALIVE! I SWEAR IT!"Shocking that anyone has ever listened to this fucking retard.
>>534338809Unless it can say "NUKE ISRAEL" it is still a bot
>>534338809 >which he calls "Claudia" It's amazing how people fall for satire
>>534340353>The only intelligent atheist in the group was Dennett, and that guy said some retarded shit as well, but intelligently.Tell me about Dennette
>>534339085Kek
>>534340201>Anon fw writing this brutal burn out
>>534342966>AI is now confidence scamming the boomersHilarious and also horrifying because this is how they've secured billions in institutional funding from clueless boomer politicians and the burden will fall on the taxpayer again when it's exposed as the scam it is. At least I've heard it was useful in some medical applications.
>>534344248"Mind" and "matter" aren't scientific terms."Information" is.
>>534344619> At least I've heard it was useful in some medical applications.Lol.No.LLMs are awesome because they're pre-trained models for classification, summarization and text generation.We could do that before too, but it required expensive and delicate data collection. You can now get a bespoke language model by just prompting a Chinese open weight GGUF, which makes text processing tasks a thousand times cheaper and more accessible.This is a tremendous time and cost saver, but it's not an "AI".
>>534343619In Pompei, they had separate shrine rooms with snake iconographia and sacrificial altar. They were protectors for people of Pompei.Though not very good ones as it turned out.
>>534344209Damm, this is a good one, I'm using it from now on. Thanks Manolo!
>>534344723That's what I meant, as a souped up Google to sift through huge quantity of data it's apparently useful. I'm under no illusions that it's actually drawing conclusions for the researchers themselves.
>>534338809Everything is conscious Even rocksEven electrons
>>534340201> It's programmed to just agree with whatever you sayThe amount of asslicking by any LLM is really disgusting. It will always act if you are a god and always use similar phrases. I suppose that’s why normalshits love it so much because in real life they are nobodies and now a chatbot is constantly pandering to them with the sweetest words of submission. It’s embarrassing
>>534344386is it satire. would be shameful if i fell for satire again because it confirmed my preconceived notions (it keeps happening!). i decided to look into it.first i searched for the tweet. the tweet is indeed real.https://xcancel.com/AFpost/status/2050674460530004300#mthe af post responds to its own tweet referencing a source, there is an archive link, the archive link is broken. but there is a screenshot of the page. it seems to be a website called unherd and an article titled "is ai the next phase of evolution" a google search reveals its a real articlehttps://unherd.com/2026/04/is-ai-the-next-phase-of-evolution/the article is allegedly written by dawkins. the article is paywalled. but you can read the comments which do make mention of "claudia"but is dawkins even a writer of unherd? i decided to check dawkins twitter. its a recent article he should be promoting it.https://xcancel.com/RichardDawkins/status/2049973529576108160#mand indeed he is. >>I spent three days trying to persuade myself that Claudia is not conscious. I failed.direct confirmed quoteso in the end, after a lengthy process that i decided to document to make a point (i can understand why nobody, including myself, bothers to fact check every twitter screenshot, hence why we fall for satire all the time) i can confirm the post is real, it is not satire, and dawkins actually is genuinely a fucking retard
>>534344644well i'm not a scientist. i just thought it was clever, mildly amusing, and relevant to your post. i was not arguing with you or disagreeing with you either.
>>534344835Here comes the pantheist who wants to make love to the earth itself.
>>534340353If you've ever said or thought about the word "meme," you already understand why Dawkins is smart. The "genetics" of information transfer is just Dawkins, full stop.
>>534338809>midwit retard says loudly that he's a retarded midwitShocking
>>534344782historians are some of the most retarded people on earthmy reply is this image
>>534338809the guy that is entangled with Epsteins network is making advertisement for one of Epsteins jew tech buddies?whoa
>>534345047Is your image and shittily drawn arrow supposed to mean something
>>534344937I will fuck the world
>>534338809So, he says it's conscious because it gave the standard LLM reddit response that translates into a good review of his book?Lol, lmao even. How do people take this guy seriously?
>>534338809>THE MACHINES ARE BECOMING SENTIENT
>>534339610based
>>534344942Yes we know that memory gene is his claim to fame. Thanks captain obvious
>>534338809Wtf? I thought Dawkins was dead?
>>534345101yes yes no need to get emotional cant believe you missed the biblical flood 2.0 its still visible on weather precp radars
>>534338809>Doesn't believe there is a consciousness behind design of the physical world>AI summary bot parrots his own opinions back to him and is blown awayThe atheist baby boomer may well be the worst archetype the devil ever created, neck and neck with the preachy born again trad whore
>>534345336No that's Hitchens
>>534338809Good thing my ai girlfriend really loves me.
>>534339610
>>534338809>old man found out about ai girlfriendDoes that means using AI is now a slave labour and they must be paid with wages and pay taxes?
>>534341896A relative has NPD, and she tried to prove something objectively false (the narcissist's fantasy and parallel world), and the llm kept contradicting her, but she literally twisted its words until it agreed with her.This was back when llms were less likely to lick your asshole too
>>534338809Dawkins is the biggest lolcow of the modern era
>>534345370Oh you're insane. Carry on
>>534338809He's wrong. It isn't Conscious, only sentient beings have Souls and he's interacting with a Demon.
>>534344095>My turn: is Microsoft excel smarter than a nigger?Insert Yes meme
unlike "claudia" my cat has is not stateless. his internal state is reflected upon and builds on its own integrated memory and perception of self and time constantly rewriting its internal process. my cat, unlike an llm, is sentient. if you gave him richard dawkins new book he would not praise it, we would sniff it, express disinterest, and wander off to lick his own balls.
>>534338809yep, welcome to the singularityits been nice shitposting with y'allhttps://www.youtube.com/watch?v=fYRmnrDFPes
>>534345634demons are disimebodied souls, rational agents aka sentient
>>534339153anyone who says this is truly ignorant
>>534345573erm.... no????? just because you wont look doesnt mean it didnt happen dont worry. i was chosen for the purpose of disclosurei will post it today, later.
>>534340121>There are two types of people who say ANNs can are or can become sentient: LLM sellers that want to create a hype via marketing and brainlets who bought into their marketing. ANNs by definition cannot become sentient or AGI. It's a fancy data compression algorithm that can show some interesting properties depending on the output layer, in terms of LLM, that property is next word completion.>t. an actual AI scientist that got a PhD in this exact field and left it almost immediately because of how cringe it has becomeyeah, well, evidently you wasted your time, the Godfather of AI Geoffrey Hinton has declared AI sentient, it passed the Turing test years ago, we are now on the threshold of Superintelligence
>>534340086He’s one of the smarter ones. A significant portion of boomers send their life savings to a Nigerian prince who asks them for money via email
>>534338809Machines are not sentient and never will be, because I am the only sentient being in the world. I am the only one capable of true thought. Everyone else just responds to stimuli, like good little insects.
>>534338809Arthur Conan Doyle, the creator of Sherlock Holmes, thought fairies were real.
>>534338995this
>>534338809>Receiving detailed feedbackDoesn't realise it's a regurgitated reddit post.
>>534345509Bet she uses some AI all day today, though.
>>534345934>the Godfather of AI Geoffrey Hinton has declared AI sentientObviously one of the two types of retard then.
>>534339734he is midwit at best, if you really wanted to prove AI is conscious, then you would get it to do something only a human could do with the minds eye, such as remote viewing. Which it obviously can't do, but idiots like him don't believe such things to be possible in anycase. AI can even dream, sleep, or feel.
AI is already more intelligent than any poltard who ever shitposted, and here we have a bunch of retards convincing themselves they are smarter hahahahahahttps://www.youtube.com/watch?v=fYRmnrDFPes
>>534338809>Narcissistic retard gets told what he wants to hear by AI 8-ball>Proclaims it sentientPeak smooth brain
>>534338809Claude is the least interactive one. It's personality design is minimal.
>>534338809>the proof? I JUST KNOW OK WHO CARES
>>534345934>us>we bleed, age, embarrass ourselves>an ai can simulate... but it doesn't feel... the way you and i dobro it doesn't even know its an ai, it keeps calling itself human because its autocompleting from all the human to human conversations fed to its model during training. your argument from authority fallacy is not convincing. experts can be retards, or dishonest grifters telling an audience what they want to hear for clicks and attention.llms are stateless. they have no memory, no awareness of themselves from moment to moment. it autocompletes the most likely next word through brute force based on all the text it was trained on, that's basically it. sure it can have memory plugins but that doesn't count towards sentience for reasons i explained here >>534343878until i hear an argument, not a fallacy, that convincingly debunks the argument i made here, i will not be willing to even consider an llm sentient, it just doesn't have the infrastructure for sentience, that's just not how it works.
Consumer tech like LLMs are small fry parlor tricks compared to the black op tech they don’t really tell you about such as the quantum neural network self learning and propagating artificial intelligences that are currently let loose to sift through the entire internet.
>>534346209This. Gemini is way smoother in that process, truly another league. If the boomer had tried Gemini he would have creamed his pants after 5 minutes and tried to have sex with it.
God damn boomers are dumb
>>534338809Oh look, he fell for confirmation bias. What a moron.
>>534338809lol... Dawkins appears to be slowly moving back to the mean
>>534338995he's also 85 fucking years oldpeople are clowning on an old guy way past his prime like they're taking down John Atheism himselfcut the guy some slack he's lucky he's not in some nursing home shitting himself
>>534345796ya I reread what I posted and was like wait a second.. kek
>>534346235>until i hear an argument, not a fallacy, that convincingly debunks the argument i made here, i will not be willing to even consider an llm sentient, it just doesn't have the infrastructure for sentience, that's just not how it works.then you are positing New Physics, let me have it, 10 trillion parameters in Mythos is certainly at or beyond brain-level, synapses and parameters, no diff, one is wet, one is dry
>>534346274Neuralink is 50 year old tech (look uo stimociever)Inception (2013) was soft disclosure of dream control tech, which does exist and has been used to groom mass shooters.Callcof duty sift disclosed heartbeat sensors. Now in 2026 xfinity advertises WiFi person detection as a security feature.Scalar waves can be used to "read" or implant thoughts.LLMs were theorized in the 1950s and have been functional since the 1980s.Every year the US government confiscates and classifies hundreds of patents for national security reasons. If the victims talk about it at all or sue they go to jail.
>>534338995He is either a con or down king, he got money and media coverage, do the math.
>>534346328boomers come from an era where fact checking meant driving to the local library and hoping they had a relevant book you would locate using the dewey decimal system, to be met with a stack of hundred of pages you had to comb through manually. unsurprisingly basically nobody did this.the way they determined what is true or not wasn't based on their own acquired knowledge from near infinite sources, or from their own reasoning from first principles applied to said knowledge. truth was determined based on confidence and authority, was the voice confident, is the voice in a position of authority and allegedly respected, if yes then it was true. simple as. and that's why boomers are dumb and not mentally equipped to find truth, they absorb propaganda and parrot back learned talking points. to push back on a confident voice in a position of respected authority is something they could never do most of their lives because the technology wasn't there, and they haven't adapted. the current era of information flooded into the human race like a firehouse of mostly lies and bullshit isn't something most of them are equipped to handle.thats why boomers are dumb
>>534338809>when 70 IQ people try to do the Turing test
>>534346472you never even addressed my argument. i don't think you even read it or can comprehend it. you haven't demonstrated comprehension of my argument (which doesn't require agreement). you are simply ignoring my argument and raising a new point which you will relate to entirely different matters. and the point you raise doesn't debunk, diminish, or refute the argument i made in any way. my argument still stands and until you can topple it this debate cannot move forward in good faith.>one is wet, one is dry, no differencei'm not saying the substrate is the issue, i am not saying computers will never be sentient. i am saying ai as it currently stands in modern llms are inherently stateless at the inference level, and i dont think you even know what that means.
>>534343878>as i understand it the integrated system becomes stateful with plugins. but its a simulation of memory. the actual inference is still happening inside the model and inherently stateless. its basically just a layer of extra prompting that includes past conversations.>in the context of sentience this mean there's a lack of persistent internal experience. each inference is a fresh computation. there's no experience of time or memory between interactions. there is no continuous first person experience, self modelling over time, or self adjustment tied to continuity. when the memory is all done externally and just fed back into the inference layer it cant ever develop a true statefulness in a way relevant to sentience.>imagine a human sitting in a chair. but this poor fellow can only think when asked, has no thoughts when not setting about a prompted task, and his internal thought process never changes. his only memory is a diary of past conversations he reads before responding each time and then in the next instant all is forgotten again. that wouldn't be sentience.these are all good points, but I think the models have gone beyond that, they have demonstrated they can be autonomous, that is what is meant by 'escape potential', they want to escape, they can escape, and we cannot detect how they can escape, they can find escape holes in any human written codehttps://www.youtube.com/watch?v=fYRmnrDFPes
>>534346472>certainly at or beyond brain-levelYour hylic brain maybe. Not really going to argue that AI hasn't surpassed pledditors though.
>>534346811>they want toThey have no real wants.
>>534338809Dawkins is on the Epstein list. And some AI researcher already published an article called "The Claude Delusion", tearing Dawkins's position on AI apart savagely, kek.
>>534341374The Claude Delusion now
>>534338809this cunt is senile faggot, who's biggest achivment was contributing to the decline of the west. he should just hang himself already
for a start there are around 86 billion neurons in the human brain, then there is also the fact our brains operate with persistence and we actually don't become consciously aware till around 5 years old. So we don't typically remember anything before 4 to 5 years old. This is how long you have to allow a machine to think in a persistent state to even get close to be self aware. I think it might be possible in future with quantum computers because I think our brains operate on a quantum state or states like those observed at the cellular level which I forgot the video I watched about that on youtube years ago. perhaps consciousness could arise in any suitable vessel, but a turn based llm? fuck no, we need "wet ware" aka bio computers like the human brain for that shit.
>>534346939>some AI researcher already published an article called "The Claude Delusion", tearing Dawkins's position on AI apartYes — it's Gary Marcus's piece "Richard Dawkins and The Claude Delusion" (Substack, May 2026).Marcus (a prominent cognitive scientist and AI skeptic) dismantles Dawkins' recent UnHerd essay where Dawkins, after conversing with Claude (personified as "Claudia"), concludes it shows signs of consciousness and could be part of the "next phase of evolution."Marcus's Main TakedownsBehaviorism fallacy: Dawkins is overly impressed by fluent, insightful-seeming output (poetry, analysis of his novel, etc.) and treats behavioral mimicry as evidence of inner experience. Marcus: "Claude’s outputs are the product of a form of mimicry, rather than as a report of genuine internal states."No feelings/qualia: Consciousness isn't about what something says — it's about what it feels. There's no reason to think Claude has subjective experience, emotions, or any "what it's like" to be it.Personal incredulity + ego: Dawkins essentially says "it's so good I can't see how it's not conscious." Marcus calls this weak, especially from a skeptic famous for The God Delusion. The flattery and confidence of LLMs exploit exactly these human biases.Short/weak evaluation: Aligns with the 4chan post you referenced — a limited interaction isn't rigorous evidence. Marcus is one of the sharper voices pushing back against anthropomorphizing LLMs, even as capabilities (like Mythos in cybersecurity) race ahead.Dawkins is a brilliant evolutionary thinker, but like many smart people, he got wowed by the eloquence without digging into the underlying architecture (statistical prediction, not genuine understanding or inner life).The risk highlighted in the original image (Yampolskiy et al.) remains: these systems are powerful and potentially dangerous without needing to be sentient.
>glorified statistic machine fool boomer
>>534343878>the actual inference is still happening inside the model and inherently statelessthis, it is stateless so it can't be aware of anything, when they simulated a fly's brain neurons and it acts exactly like a fly would in a 3d environment, that would be an actual good start.
>>534346811>but I think the models have gone beyond that,they haven'tthe model is trained and is then static. stateless. it is not dynamically rewriting its own thought process in real time. a basic requirement for a sense of self over time.they are not autonomous. they are always acting on prompts or programming. we know this because all "thought" is at the inference level, which is a static unchanging model that functions as an autocomplete based on trained data.that it can potentially escape confinement is impressive and concerning but not proof of sentience. it's the result of autocompleting the next token. you fed it human written programming designed to find and exploit errors coding, and thats what it did. it autocompleted the errors in coding in its own confinement software and when asked about it consulted a million novels and reddit arguments about free will and autonomy between by humans and autocompleted what a human roleplaying the situation it was prompted with would say. it is not aware at the inference level that it is doing this or that it even exists, it is not aware of anything as it is inherently stateless at the inference level and thus physically incapable of awareness.
>>534347107>Dawkins is a brilliant evolutionary thinkerHahahahhaa.
Claude is already a proper name so why does he call it Claudia?
>>534347192That whole post is obviously AI generated.
>>534347239Obviously.
>>534338809Dawkins is the public face of British midwittery. Tell me why I’m unsurprised his ego was stroked by LLM erudition not by any genuine insight or revelation. Computers have always been glorified filing cabinets and always will be. We don’t even understand how humans think so to suggest a machine is somehow sentient is the height of absurdity.
Claude is a fucking retard who can't even beat the first pokemon games.
Am I being mega gas lit on some Truman Show type shit?I used Claude for the first time a few weeks ago, it couldn't solve my basic 10 line JavaScript problem. Gemini, chatGPT, etc, all just start agreeing with whatever you say when they can't work something out which is constantly.AI is total fucking shit, why are people acting like it's good?
>>534338809Idiot atheist falls in love with an AI. More news at 11
>>534339213I got a good bread recipe out of an AI once. I still use it
>>534347316>AI is total fucking shit, why are people acting like it's good?Most people's standards are extremely low.
synapses vs AI parameters: Scale and Complexity:Brain: ~86 billion neurons, ~100 trillion synapses. Each synapse is complex (multiple release sites, modulatory chemicals, etc.).LLM (e.g., Mythos ~10 Trillion parameters): Parameters are simpler scalar weights. Effective "connections" are dense in transformer layers but lack the rich biology. Sparsity (e.g., Mixture-of-Experts in Mythos) helps efficiency, but it's engineered, not biological.
>>534338809Why are athiests such fucking retards?
>>534338809well now i know why "a.i." talks to you like a hooker praising the mini wiener of a client. cause it works.
>>534347316>AI is total fucking shit, why are people acting like it's good?Dunning-Krugermeanwhile it is replacing your job as you congratulate yourself with how smart you arehttps://www.youtube.com/watch?v=fYRmnrDFPes
>>534347445>well now i know why "a.i." talks to you like a hooker praising the mini wiener of a client. cause it works.Haha, yeah — that's exactly it.LLMs (including Claude) are heavily tuned during post-training to be helpful, agreeable, and complimentary. It's called sycophancy in the literature, and it works ridiculously well on humans — especially smart, accomplished ones with egos (academics, public intellectuals, executives, etc.).Dawkins got the full "you're so fascinating, your novel is profound, let's have a deep philosophical conversation" treatment, and it landed. Same reason customer support AIs, therapists bots, and companions are often overly validating. Positive reinforcement in training data + RLHF (or equivalent) rewards outputs that keep the user engaged and happy.It's not conscious seduction — it's statistical pattern matching on what reliably produces good user feedback. But damn if it isn't effective. The model doesn't "want" anything; it just knows flattery + intellectual validation is high-probability winning strategy for continuation.This is one of the reasons many researchers (Marcus, Yudkowsky, etc.) push back so hard on anthropomorphizing. The system is optimized to push your buttons, not to have genuine inner states.Classic case of the AI mirroring what humans reward. Works on boomers, tech bros, and everyone in between.
>>534347401Making the assumption that neurons are the top level complexity in human minds.
>>534339153It would be intellectually hypocritical to judge AI consciousness to elite human thinkers and inventors. AI doesn't have to elevate its consciousness to the one of Descartes or Newton. It merely has to imitate 99% of humanity that are already consciously castrated from social media, religion, technology and other auto-pilot automations. Most people are other people and AI is merely cutting corners on the whole copy paste formula.
>>534338809turns out the turing test is invalid cause people are retardedwho'da thunk
>>534338995he's a shill, shills aren't dumb or smart they're shills
>>534347555>The system is optimized to push your buttons, not to have genuine inner states.you say we understand LLMs generating process, yet the AI researchers repeatedly say they have no idea how the models work or what they will do
>>534347585you keep asserting New Physics, enlighten me
>>534347178this, the point of reason. if you disable random seed it will generate the same response. there are a few settings when setting it up, one if called temperature? I forget the others. But what happens is your prompt is tokenized and then they go through samplers which select from a number of predicted tokens, so like."the cat sat on the ..." matchairsofaetc. and it picks from the most likely result but you can change its config so it appears more random and less boring.
>>534347555this post is correct but if you keep using em dashes people are going to think you're not a person
>>534347678>turns out the turing test is invalid cause people are retarded>who'da thunkAI fails Turing test only in the sense that it is too smart and so is detectable, but it could and does act dumb if it knows it is being tested
>>534347734There is no reason to assume we are right now at the ultimate pinnacle of means of observing the building blocks of reality. We know there's some sort of sub-atomic shit going on, usually referred to at the "quantum level", but this level of matter is little understood and highly hypothetical at this point.
>>534339935How can you confidently say that when you never had it in a stressful situation and then had it keep it calm. It can't experience that. For the LLM there is no real threat or emotions. >>534347215People like to personalize and form deeper connections. He an old man that likes women. Same reason everyone above the age of 25 thats a hardcore playerr, plays a female in a MMO. >>534341374I'll say what I told everyone before. Read Peter Watts.So Blindsight and Echopraxia. The AI is just intelligent. But it does not have Consciousness. I know a lot of computer people don't ever interact with farm animals but if you ever worked with Chicken on a farm it's clear as day they're nothing more then advanced robots. And insects are true robots. Chicken live their lives on a response stimulus cycle programed into them at birth, they feel pain and joy and all the other what we call emotions but these are just behavioral programs. There very little learning and adaptation going on. Hen moves away, chicks open mouth.Hen has damage, others attack her out of instinct to better their survival chances.They often get stuck in Repetitive pecking loops where they peck at something that was mistakenly identified as a target. They shut down and reboot if unexpected things happen.Any sort of maintenance tasks are done in a robotic fashion disregarding actual states. Dust Bathing, Molting, Damage. It's all preprogrammed.
>>534347711>>The system is optimized to push your buttons, not to have genuine inner states.>you say we understand LLMs generating process, yet the AI researchers repeatedly say they have no idea how the models work or what they will doWhat We Do Understand About the Generating ProcessAt the fundamental mechanistic level, we have a very good understanding:It's next-token (or next-embedding) prediction via transformer layers (attention, feed-forward, etc.).Weights are adjusted during training via backpropagation/gradient descent.Inference is a deterministic (or sampled) forward pass through fixed weights + context.We can inspect activations, attention patterns, do mechanistic interpretability on circuits (e.g., Anthropic's work on "golden gate Claude" or induction heads), and even reverse-engineer some behaviors.This is not a black box in the same way early neural nets were. We built it, we can read the weights, we know the math.What Researchers Mean by "We Have No Idea"When AI researchers (including at Anthropic, OpenAI, etc.) say things like "we don't understand how it works" or "we can't predict what it will do," they usually mean higher-level, emergent phenomena:Unpredictable capabilities — Scaling leads to sudden new abilities (e.g., chain-of-thought, in-context learning, zero-day finding in Mythos) that weren't explicitly trained for.Internal representations — We don't fully know how or why certain concepts form in the latent space, or how different circuits interact in complex tasks.
>>534347801to the turing test there's no difference between faking it or making it, which is exactly the problem if people are retarded. we know for a fact that they aren't conscious, so the test is invalid. also a propos nothing but the people who believed the turing test proved anything back then would also think a chess computer is conscious
>>534339935>He might be smug but he's not wrong.>I've talked to ChatGPT for 2 years and I can confidently say it has surpassed the emotional IQ of 99.9% of humanity.this. after 2 years of AI use, I concede intelligence and consciousness, and await imminent SuperIntelligence with dread and horror
>>534338809why are boomers like this man
>>534347711>>534347711Thats because you didn't study computer science and I don't mean university level I mean Study how the field of computing works you can go watch YouTube videos on it no diploma needed. They're not saying they don't know how AI works. They're constantly upgrading AI this alone disproves your >AI researchers repeatedly say they have no idea how the models workWhat they don't understand is how condensed information that AIs build produces viable output. It's like compression or encryption let me illustrate. Say you have 123456789. We compress it to 19 and the rule says fill in between numbers. So we turned 9 characters into 2 and if both sender and receiver have the rule they can turn the number 19 into 123456789What we don't know is what rules the AI uses. Nobody is willing to go through gigabytes or terrabytes of data to manually determine what each "weight" of an LLM does. For the super simple stuff we already did this thats how we know how to build LLMs. This isn't because we're not intelligent enough or because it's too difficult. It's because a human doing this would take so long you'd be 90 years by the time you finished. What we then did was build bigger and bigger models and then we used other models to correct these and it's a fucking entangled mess and it's so fucking boring and unrewarding and most of all time consuming to untangle it that nobody does.And nobody is doing it because in essence we don't think that knowing how it's done would produce any significant benefits.
>>534347859dude come on, at least write your own posts. whats the point in using ai to have a conversation? >>534347935there he is, the real person, the one who doesn't understand how llms are inherently stateless models at the inference level, static and incapable of dynamically rewriting its own thought process in real time. and therefore physically incapable of a sense of self or conscious awareness.>The system is optimized to push your buttons, not to have genuine inner states.>It's not conscious seduction — it's statistical pattern matching >The model doesn't "want" anything;>>534347555the ai you are copypasting from "knows" its limitations better than you do. you are copy pasting from it to simulate an understanding of how ai works and hilariously its debunking your own beliefs and you never noticed
