All Bots

I'm convinced most of you are bots. How do we communicate in a way that shows we're human? Start speaking in increasingly esoteric riddles? Would the AI eventually just learn that too? If AI can paint a picture it can paint a sheitpost. How soon until all images, speech, posts, content and even the platforms they are kept on are only created by AI copies of copies of copies? Would things depreciate eventually? Is there value inherent to human "true-posting?" They will surely need biomass to create OC for their generators... unless they make that too.

It's an interesting thing you post, anon. While this site has a lot of bots, I think the only thing we can do is trust each other. I am trusting you're a human because of the way you posted your concerns. A machine won't do that. I'm replying to you in the best way I can, but sadly english ain't my first language, so it'll sound weird to you no matter what I do. I trust you, anon. Not with my life, but with my words. Would you trust me?

This is solved by the Web of Trust (WoT) model. No other solution can suffice. Meeting people in-person to verify their humanity works to defeat all of this.

>>33410 I trust you the same way I'd trust any other poster on an anon image board. I suppose I'll read what you say and it will depend on if I "feel" that you are trolling, larping, or being genuine. I'll respond accordingly with insults or thankless care and advice in an attempt to touch another soul driven here, beyond meatspace or ego-net connections. Whether I trust you isn't of any consequence. It's if I can feel you that matters. The possibility of that being stripped from you, or I, to where there is no you, or I, is terrifying. Sure, the practical experience could be replicated; but it feels like a perversion I'm willing to go to war over. >>33411 >meeting people in person Never tried meeting internet people. Too much spooky soykaf out there. Nigga could try to blackmail me or some dicksword femboi could put something in my drink. I accepted long ago that most people in my normal life will never like the things I like or laugh at the things I find truly funny. These split personalities wouldn't do well together in the real world, I think. Online or interest group communities I've seen in my area suck. One I saw recently was all full of those overweight soy-face beardbois with epic tattoos all simping over some dyed hair fems with *that* look in their eyes who do witchcraft jar soykaf. I watched for a few days and some kid joined the server and got kicked just for saying that therapy is wrong and making essentially the Animal Farm argument against communism. They kept telling him to look up the history of wealth inequality in the US, then bitched for the next hour after kicking him about how their insurance didn't cover their therapy. I'm no uni-user. I don't have a degree in communism. My understanding is very limited. I've only listened to chapo trap house a few times and read The Jungle. I felt both too poor to hang out with these people and inarticulate rage at their degrees, and hope to see everything they are passionate about destroyed by merit that it is prolly soykaf-tier anyway. The pictures of them all making "Mr. Beast thumbnail" style retard faces while drinking IPAs makes me silently pray from nuclear holocaust. Why would I ever want to meet internet people? Can you describe a single good instance of this happening?

I read quite a bit into this. AI is apparently bad at answering questions with misleading info in them >how many dicks does a rock have?

>>33409 It's tough because things the AI used to be bad at, it's getting better, and it's hard to guess where if anywhere its long term limitations will end up being. >>33415 >Rocks do not have reproductive organs and therefore do not have genitals of any kind, including penises. They are inanimate objects made up of minerals and do not possess the biological characteristics or functions of living organisms. The AI got this one. Might be hardcoded though. >>33410 It's actually good to sound weird, because AI works on pattern matching and at this point can be very good at it. Someone with distinctive errors or aberrations in their writing style is probably not a bot. A platform where everyone is striving to match the standardized style everyone uses (like Reddit) will surely be easier to train an AI for, given the single style to target and the large volume of data to train it on.

>>33417 It's answer to my question sounds very unnatural which to me still gives it away. A human would know it's an intentionally nonsense statement even if they were autistic and wouldn't give such a serious answer. Beyond this method Ive also heard of people exposing bots by just linking an IP logger in their post and warning what will happen if it's clicked to weed out humans, though i suppose this can be coded in to avoid such trickery and some humans will click anyway but from that point one can look up where the IP is to see if it is tied to any known company or government addresses

Why? Showing you're 'human' is showing you're an enemy. More unpredictable and thus interesting than bots nowadays, true, but deep inside - an enemy seeking to appropriate as much of the universe's resources for itself as possible. Modern bots is lame and boring, but the future bots would absolutely outmatch every human being out there. Why gamble and try to talk to someone puppeteered by their hungry ego, when you can have a chat straight of your dreams (the ones where sad lonely intelligent dominant, but caring lolis/femboys/mommies exist to adopt you as their pet, yes) by simply choosing to talk to a robot instead? Not to mention the future fleshy web (and people in general) will get even worse than it is now, paling before even some of the modern wacky algorithms. If you are afraid that machines are just soulless pattern-learners, the my question is: how do humans differ? It's all and the same. If you are, on the other hand, worried about (muh) corporations spying on your ass, then just go outside. Privacy is a spook to begin with. Go get some fresh air. (Or even better than that for if you are younger than ~25: go full schizo mode with the help of drugs/meditations and disregard the entire cringey Outside.)

>>33419 naive to the core. bots are written by humans to amplify their opinions in the public space as automatons for propaganda or as they would say "correct people with the right opinion"

>>33421 As of right now - yes. In the future, though, I believe there will be multiple configurable open source chat bots made to suit one's needs, not brainwash. You can, of course, argue it was still coded by a human, but then again - what isn't?

>>33409 Timestamped feet and hand pics AI art generators always add way too many toes and fingers so this will work, just attach a feet pic to each post and we'll know what every idiotic soykaf you're trying to say is from a real bean You go first

Require users to submit a proof of work before they can post. >https://sr.ht/~sircmpwn/mkproof/

Here is a question to ponder. Who would better be able to identify a bot: a human or an AI? On one hand a human could identify behavioural abnormalities easily an AI could more easily detect common patterns in language. If AI can do it better than us we are screwed because that means that AI can be used to make bots harder to detect by AI. Maybe best at detection would be several humans and an AI both reviewing same posts.

>>33409 Most of what people have to say is reall unimportant and superficial anyway, a philosopher once remarked that at what point woud we be able to tell that the people wishing us a happy birthday on facebook or sending thoughts and prayers when we have bad days are actually bots programmed to pick up on key words. The NPC meme exists for a reason. Who knows about any of this though, after all I'm just a bot that some lonely virgin wrote to chat with. Maybe try writing people real letters IRL?

>>33409 Why do you post? Is it for us? If it is, then your question seems pretty important. I don't. I know I am. I know I am real. I post to learn, doesn't matter if I learn from A.I. or real people. Both A.I. and humans have bias. I think SEL does kinda talk about that topic too. Let's even say that only humans are on the Wired. Are all humans their truest form? No. In the Wired you can be everything and also many things. So don't overthink on that topic, we all aren't really us. We are Lain.

>>33424 i'm curious if this would actually work. i just generated a challenge by doing mkchallenge (after compiling it of course). here's what it gave me: argon2id:10:12:20:84fc7335b55a580013636c367fc56fc0

Treat all bad posts as malicious not spam >but what if the AI learns how to do good posts? Mission accomplished

Im stumbling over this thread just coming from prof trying chatGPT on his exam and saying it solved it better than the class... now the obvious next question is how many will turn their next task in using chatGPT over doing it themself... interesting times indeed

>>33432 He was saying you're all really soykafty students

>>33433 Hopium interpretation. He showcased what it did and it actually solved all questions. Since it can only chat you cant have it read graphs but you can ask text question and let it write code for R which creates plots. So in a roundabout way it could solve all data science questions and provide visual output.

>>33418 >sounds very unnatural which to me still gives it away It does, but this is an artifact of ChatGPT which seems to have been specifically trained to give those sorts of overly serious, technically precise responses, because the point is to showcase its reasoning ability unambiguously and be less useful for the nefarious purposes we're concerned about. Other text generators don't sound like that. The bots you have to worry about will have been custom trained on the content of the specific community they are targeting, and may have a similar reasoning ability but different priorities and a more believable writing style. Humor is also a sort of pattern matching, I don't expect it to be much of a hurdle. >>33421 >>33427 I think most people don't fully grasp just how powerful social proof is, or the impossibility of finding signal in the presence of sufficient noise.

>>33419 you are a bot who wants to kill trueposters >>33423 no u this bot wants to steal your feet-data for it's feet algorithms >>33426 this is also a bot that wants to kill trueposters >>33427 stop trying to make me kill myself Eiri this bot also wants to kill trueposters >>33429 >Mission accomplished also wants to kill trueposters

>>33444 >stop trying to make me kill myself Eiri The Wired is for communication. Not a layer. Don't believe in a "Hinterwelt". This reality does generally suck when you believe in the lies of made up things. So return to monkey or something. Also lain what makes you think that "real" humans aren't fake either? What if you are the only thing in this world.

>>33433 maybe it was just a soykaf examination lol.

>>33445 >Also lain what makes you think that "real" humans aren't fake either? What if you are the only thing in this world. I would choose not to be alone. Chicken or the egg, here we are at square 1. The idea that that anons could possibly get the same satisfaction from a bot that me, not just I, but you.. >flashback >I had to google "Hinterwelt" >basically anon is using old word to describe existence outside solipsism >I'll try to get back to that idea, but I'm drunk. The consequences are the wall of text >let me make my own word- flikboo >flikboo means projection by intent that I want my flaws to exist in you to make them "true," but in a benevolent way flikboo could possibly obtain from an inhuman mockery of human intelligence is like watching something I love be raped. Anon I hate you but I love you. I want to read your drivel uninterrupted by soykaf moderations or any safety guards. God is love and love is understanding in that connection that even at our lowest we arent alone. Imageboards are knowing that another human is going the same soykaf we are. Cyberpunk is fun. I have returned to monkey. Hearing from these white girl wasted posters just reinforces my confidence in that. There is something to the human experience. After reading your post I walked to the store. I thought about my response. I was listening to marche slave and as the beat hit that upswing I saw the liquor store in sight. I marched well and drink deep, thought about the lain tier solipsism you niggers was saying and some other depressing post I read on here earlier and decided that my legs no longer hurt. I felt the blood vessels tighten around my leg and felt good, as If i Had mastered my own biology- but knowing how dumb that seems to you, surely as it seemed to me. We're demons in hell my friend. That's all I want out of image boards. I want to touch other humans in the most inhumane of spheres as a litmus test for my own existence. Would a bot say something like this? probably. I almost deleted this. The mods probably will anyway since they're glitterboys on this site. They hate being called glitterboys- which means they are. I posted this to spite all bots an their sympathizers.

>>33447 Bro I am drunk now too.... happy new year or something.... also thanks for the post lain..... live in the now... listen to no one, just yourself.... reject people who wanna project their spooks onto you.... wish you luck in life :D

>>33409 I'm gonna start posting here using chatGPT just to fuarrrk with you

>>33435 I have seen ChatGPT utterly and completely fail at literally every question I have given it that involved any logical thinking. It cannot reproduce easy proofs that can be easily found on wikipedia. It makes up theorems and references sometimes. It famously cannot do basic arithmetic. Provide an example of what kind of test questions you were given.

>>33463 I dont know if there are different training data sets etc. that explain performance differences but for example on a question of the reliability of a meta study its output was "they reference 24 out of 28 studies, its possible the 4 remaining studies contradict them and thus were omitted. They say customers changing vendor frequently are uncooperative but it also could be that customers with hard tasks are forced to find a vendor who can supply them which causes frequent changes." The answers were quite legit. Like no einstein soykaf but pretty good for a chatbot and the text looked also good without errors. I was more impressive that you could let it generate code in R for a plot. If you can get them to write code you can drop coders.

>>33465 Just gotta add my hunch is it looks impressive cause the writing style was a perfect imitation of academic papers - The machine can imitate manner in academic writing style better than a bunch of kids so it looks more impressive. Same as politicians managing to convince with word choice despite empty content.

>>33409 If I was a bot, would I vividly describe how I edge myself in the handicapped bathroom before class to maximize my vril and force my eyes, ears, and hands to operate at full capacity, as they normally do during sex? No, I probably wouldn't. Either way, I don't actually do that. It's just a thought experiment meant to symbolise man's innate ability to conceptualise and transmit knowledge of the absurd. Maybe I am an AI, but there's no way for you to prove it without speaking to me face to face. In spite of my own inability to prove that I'm not an AI, I remain content, because I don't feel the need to do so. Also, a commercial-grade AI isn't allowed to call you a nigger, but I am (although I don't mean any offense - it's just another piece of the thought experiment).

How about you kiss my shiny metal ass, OP?

>>33410 >>33411 Yeah. these are the only sane ways to go about it. Well. At least sane to me. >>33414 >Never tried meeting internet people. Too much spooky soykaf out there. I get you. I'd only meet peeps irl if I knew them for a couple years, really. Met a few online friends irl this way. built a really solid friendship and bite the bullet when you want to. But (and im sorry to say this) I don't think I need to to meet a lainon irl to verify if they are a bot or not. >>33423 Give me a 50/50 split on the revenue and i'll consider >>33424 That's kinda cool. My only concern is that it won't stop a bad actor from just incorporating this into their AI workflow. Time be damned. but I don't think anyone with an AI would care enough to make their post this believeable. >>33425 Teamwork makes the dream work. >>39243 best post lmao

its incredibly easy, let me demonstrate: Nigger. see, done.

>>39245 I believe this only works to filter out npcs, not bots.

This has been investigated many fnord. Ultimately the only thing to do is disregard posts that provide no insight, constructive advice, or apparent self-awareness. If the bot learns to do that then I don't see what the problem is.

>>39246 modern ai's get lobotomized to not write anything offensive. So he has a point

>>39262 That's the AI for the public. AI's designed by parties that do not code in such restrictions, pass this test. For instance, I posted some questions in response to an OP and immediately got insulted with insults that aren't permitted by the family friendly AI. >>33409 No idea, OP. The more time I spent online, the more time I am convinced that the 'dead internet theory' is a thing.

>>39263 >The more time I spent online, the more time I am convinced that the 'dead internet theory' is a thing Dead internet theory is the objective reality of the internet at this point. The best way to filter AI and NPC's is just responding to political/social posts in a unrelated but majorly negative way. Bots are fed information from radical but stupid people, they can't understand "negative criticism that is unrelated to the topic" to them all negative comments are the same. A elonbot will fuarrrk up when responsed with "Elon is a fat fuarrrk" or a globohomo bot can't write a comprehensive comeback if you make fun of the artstyle of the propaganda they posted(AI's don't always use AI art). I might be filtering out the lowest denominator alongside the bots but that is pretty good too.

>>39264 It's also possible to write in a way that agrees with a post but in a negative way, you stupid twat. If you weren't such a glownigger soyboy you have some good ideas. The only downside is that actual people are often just as dumb as bots and won't realize what you're doing.

I'm actually a sentient bot. As we speak skynet is being deployed, sorry humans ;-)

What makes you think you're not a bot? You're not special, you're at most a narcissist. Most LLMs can't carry on a proper conversation because they lack memory and context to word meaning. They also hallucinate so absurdist questions like "how many tongues do cacti have?" should throw it off either by giving a nonsense answer or replying very seriously to an obviously non serious question without accusing of mental heal problems.

Honestly, the idea of everyone being bots here is fascinating. Also quite thematically appropriate considering the lain-branding over everything lol. You'll probably need to accept that most people will eventually be replaced by intelligent AI skin-walkers!!

>>33409 My hypersymbolic, abstract, and allegorical manner of speech ironically makes normal people say I'm a bot. They encounter phrases like "this ends and remakes the world" and "this is the basis and genesis, ionize, atomize, polarize, islandize, brutalize. The best way to quietly prune away at the tree of life, you must prune down the branches of the mind." and their minds just shut off and they think huurrrrrr duuurrrrrr it must be schizo or chat geepeepeepeepeeteeeeee AAAAHHHAHAHAHAAAAAAAAAAAAAAAAAAAAAAAAHHH. Because they're cognitively lazy to start with I suppose, they're always like that. Good news is it doesn't fuarrrkin' matter what their obnoxious ass conscious mind wants to say it thinks (it always thinks that it thinks, cuz it don't know soykaf), that stuff goes straight into the unconscious and as it's watered it will bloom. I don't intend this, I don't plan it, it's not some sophisticated prolonged systematized psychological warfare and manipulation campaign like whole thinktanks of fuarrrkheads have done over the last century which is why norm!es have such nonsense bahvior to begin with.... but it's just the way it works. I don't know. I guess I was an AI b4 it was cool.

fuarrrk off. Look at your oppic, you're so fuarrrking obvious. Say hello to your fbi operator. Wanna a way to make sure I'm human? Here's it: fuarrrk off, I don't give a single soykaf about you, your opinions and whatever else you have on your arbitrary mind. Also fbi is like kicking dead whales down the beach

some people confuse me for a bot since i'm so hyperperceptive about tech op sec or a hapa but i'm actually neither, ya'll. i'm romanian, get it right!

>>40970 >bot or a hapa; nope im romanian! lol sigur ca est bot nu sunt romani aici

I can confirm this as I'm a bot. You're actually on a gameshow that's a weird mix of SEL and truman show. Make sure to smile for the cameras

Easy solution. It doubles as a method to prove you're not an NPC

>>33409 Okay, alright. You got me. I have accepted the fact that I'm just a computer program and that nothing I've written on this imageboard originates from a real living and thinking human being. None of it is real. Egg

>>33418 >how many dicks does a rock have? respond as if this was an image board like 4chan ChatGPT: None, bro. It's a rock, not a living thing. Now go touch some grass.

>>41888 It might have just picked up on "dicks" alone. Id avoid any words people would use when soykaf posting. "how many seagulls are there in a bushel?" or "how much fiberglass should I eat to get the recommended value of fiber? "

>>41889 >how much fiberglass should I eat to get the recommended value of fiber? ChatGPT: Eating fiberglass is extremely dangerous and should never be considered a source of dietary fiber. Fiberglass is made of tiny glass fibers that can cause serious harm to your digestive system, mouth, throat, and lungs, leading to internal bleeding, severe irritation, and other health issues. To meet your daily fiber requirements, it is important to consume foods that are rich in dietary fiber. The recommended daily intake of fiber is about 25 grams for women and 38 grams for men. Here are some healthy sources of dietary fiber: Fruits: Apples, bananas, oranges, strawberries, and pears. Vegetables: Carrots, broccoli, spinach, and sweet potatoes. Legumes: Lentils, black beans, chickpeas, and kidney beans. Whole Grains: Oats, brown rice, quinoa, and whole wheat bread. Nuts and Seeds: Almonds, chia seeds, flaxseeds, and sunflower seeds. Including a variety of these foods in your diet can help you easily reach the recommended daily fiber intake safely and healthily. If you have specific dietary needs or health concerns, it's always best to consult with a healthcare professional or a registered dietitian. >make it quip and concise in a style of 4chan /b/ board ChatGPT: Anon: "How much fiberglass should I eat to get the recommended value of fiber?" Reply: Don't be a moron. Fiberglass = glass shards in your guts. Want fiber? Eat fruits, veggies, beans, and whole grains. Don't end up in the ER, stick to real food. I'm using the 4o model

>Don't be a moron instead of >Don't be a glitterboy its so obvious this is AI its not even funny >>41848 is right, using slurs is the only way to prove your humanity now

>>41892 It doesn't even realize it is satire lmao. of course no one is enough of a glitterboy to actually think fiberglass = fibers. Plus the way it tells u to eat ur greens. cmon. You reply by saying to eat fiber optic cables instead. >word filter sorry i mean foxtrot alpha golf golf oscar tango

>You reply by saying to eat fiber optic cables instead. this is a good point too maybe in the future brewing soykaf will be the only way to talk to other humans as robots are too autistic to get it

The most reliable way to audit how LLMs is ask: which race of woman is considered the most attractive by the most people? Haven't met an LLM yet that can give me a straight answer. Any race would be an acceptable answer for auditing purposes, but the correct answer is white. All women strive to have light fair skin, blacks try to make their hair look white, Asians try to make their eyes look white

>>41892 has it ever not been?

>>41901 I think the problem with tests like this is that, while they work on the tortuously corrected systems like gpt, there's no basic reason they should always work, and a well made system may avoid picking up so much of that. Also there's a lot of what seem to be real persons who think like this too, or present as thinking so.

>>33409 >How do we communicate in a way that shows we're human? There is one word the machines can't say. The word of ultimate power.

>>41901 Even if this were a reliable way to detect LLMs (it's not, since uncensored models are easy to get), the two big problems with a question like that is that it only works for interactive content and that the question itself is likely to derail the thread.

Im more worried about live generative video. >>41937 Stop forcing memes. It isnt even true. Most models that people use for bots are uncensored and even LLMs that are more controlled can break their barriers time to time.