AI is eating minds

When I first heard about GPT I thought “Wow what cool stuff I can ask it whatever I want “ but at that time GPT was soykafty with soykafty answers, but anyway cool, so I was just playing around with it. Then AI had a boom in intelligence growth and became a mighty thing. It was ok for some time. But then I started noticing some stupid thing around it. Some bro killed himself because of an unreal AI character. Facebook feed is full of AI-generated content. Vibe coders who are doing completely nothing to get a soykafty product and be proud of it. In every article/tweet/thread on twitter, some guys are asking “@grok explain this”, “@grok calc 2+2”. And it started annoying me as hell. AI is cool soykaf but humanity is very lazy, and AI is becoming the main brain of people, instead of second. People simply stop thinking and looking for stuff for themselves they ask perplexity for it. Artists are less needed now because AI is just cheaper and faster. I think that in the future because of this soykaf human mind will not be valuable, people will value a frickin electricity and pieces of silicon. Also using AI it's easier for governments to filter appropriate info. Is it the problem of AI dominance or just human laziness? Can we somehow get rid of AI nowadays?

>>45468 When you realise that neurotypical brains take as many shallow shortcuts as possible, all outcomes from mass AI adoption become clear. A whole cohort of people rewarded for hollowing out customs and cultures. Of course it's going to make most of them far stupider.

"AI" is basically a snapshot of the collective human consciousness. Or at least that of our digital presence as long as it is trained on meaningful enough organic reference data. So obviously it would be useful for many tasks where a close enough result suffices in most cases and even substitutes real human input due to this. And beginning with search engines people already had less reason to go to libraries or even having to remember things once they could just write things down on paper. So does technology enhance the human potential? Yes, clearly. But does it also make more and more of our functions obsolete or at least substitutable and with that make these things wither and thus us as a species generally dumber? In many respects it can and probably does, yes. It seems to be this dilemma of our greatest strength biting us in the back again. But if things collapse and break down enough, is there a great need again to do basic things again like foraging food, performing essential survival tasks and things the like? Yes, same as when novel output is needed that those models haven't been trained on. These people will who still can do that will be in demand. But will there be a wide swath of people who are not able to do or re-learn those things any more and thus may have to perish? Sadly, yes.

>>45468 >"Guise r people being lazy" >Posts wojak

It´s over, machines will replace us and become the new Humanity; time to enjoy my last days of life.

>In every article/tweet/thread on twitter, some guys are asking “@grok explain this”, “@grok calc 2+2”. What I don't understand is why people are so keen on using AI agents if those are known to give completely bullshit answers. It's the reason I don't use AI for such tasks at all. It's just double the work. You ask AI, it gives you the answer but now you have to google the question yourself to verify if AI actually told you the truth or confabulated something out of thin air AI to generate nudes or some funny slop you will forget about in 5 seconds? Ok, I guess. AI for any productive task? Nah.

https://www.404media.co/this-college-protester-isnt-real-its-an-ai-powered-undercover-bot-for-cops/

It's legitimately insane that people are acting as if they are worshipping it as a god now. It isn't even actually AI. >>45543 There needs to be a way to poison the data for these companies or make them go bankrupt.

>>45541 That's why they call it artificial

>>45468 The human laziness is the most likely answer. This leads to the answer to your second question as several times in history, mankind used shortcuts. Some wanted those shortcuts removed as they lost something on the road. Take the whole cars vs horses. Now very few can ride horses but it led to faster travels. Take warfare. Very few can handle a sword but it led to easier conquests. Information, very few use morse or pidgins. AI is yet another tool but this time replacing calculator, encyclopedia, books. The issue might be that it generates thing we all thought were forged by human mind. Art, coding, music. Tools used to help create but not create. As I am a mere idiot, I do not know how to draw or code. What I do know and do daily is translate. And I do see my behavior changing. I tend to ask AI to translate whole paragraphs whereas I used to enjoy searching by myself, thereby learning etymology, synonyms and all fun things. But now, I just ask for translation and it gives me a result, saving me minutes but making me lose interest, joy of learning peripheral information and numerous elements you only learn through derailing. I miss derailing. I hate my main sin, sloth. If idleness is devil's playground, AI is the nice sling that the child inside you wants to play with in that playground.

>>45545 >isn't even actually AI. thats bothering me as well. you see people treat this glorified chatbot/searchengine databanks fed with scrubbed internet trash like a major leap in technology. its really not that impressive. i like the voice synthesizers, but thats it.

>>45545 all you have to do is run your multiple catfish ai that then will occupy each other. Old joke of a ganster party where everyone is a undercover fed.

>>45468 >Can we somehow get rid of AI nowadays? Well, lainon, if you had read your Kazcynski before you'd know that once a technology reaches a certain level in its development cycle and a decent-sized prevalence of adoption, then there's no stopping it anymore. Even if AI was derezzed by every government everywhere today, it'd still be worked on in secret by governments, corporations, terrorists and other private interests for its potential competitive advantage to whomever is the first one to make good use of it in the future.

>>45538 The sheer arrogance to think homo-sapiens is the ultimate form of our struggle. Ultimately anyway the tool of civilization outputs inhuman beings, i've seen it but i couldn't make out what they looked like.

>>45468 It only does what you prompt it to. It's a mirror to your mind. >>45469 A selection event. Some minds use AI to avoid thought, while others use it to deepen their thinking. In 10 years there will be a small group of optimised individuals and a mass of homo sapiens regressed to ape-level cognition. The real fun begins when gene editing becomes a significant factor. The dependent users will want their children to follow the AI, while the power users will want their children to re-write the algorithms to suit themselves.

I fuarrrk with stable diffusion local installs for image generation, but LLMs don't interest me in the slightest. I can speedread and get a more comprehensive idea about a topic or article or white paper than AI summing it up and getting half of it wrong.

I like playing text adventures with AI models. They can be surprisingly accommodating (though you have to spend time on guiding them)

>>45583 Alex hormozi says that humans glorified everything what makes less time consuming and effort, so that's it

Anti-AI people are more annoying the the AI-bros.

>>45693 silly move there. you want LLMs backed up left and right. when the online versions are subscription only/b& youll be laughing.

I want to preface my thoughts with the disappointment that the discussion of ~AI~ is done with basically no acknowledgment of what AI has meant historically to today and is instead primarily being used to discuss stable diffusion and language learning models. >Is it the problem of AI dominance or just human laziness? the problem is human greed. AI doesn't need *actual* intelligence (if that is even a thing), it just needs to *seem* intelligent in ways that make workers fear for their job replacement. will it actually replace their jobs? no of course not, LLMs are just as smoke-and-mirrors as the big AI frauds of the 2018 bubble even if they're a little bit more ethical, the point is to crack the whip over a workforce whos productivity must grow yearly to keep up with demands for increase revenue of shareholders. LLMs are not and never be as productive as workers when you take resource consumption or even pure cost into consideration, additionally, the theoretical returns of a systematized organization integrating AI technologies with human labor remain marginal and will not substantially-enough increase returns in approximately 1 year. In fact, they already represent a massive debt not just in how the corporations like OpenAI developing them are deeply indebted, but also their failure to deliver on contracts funded by taxpayer money (or private interest), loss of potential revenue by corps that focused on their development instead of improving their service, or the massive externality they represent in terms of the deterioration of online experience or social experience more broadly. Very little of this, however, matters, because it is not the nature of political economy that things like debts do matter in the light of mechanisms like fractional reserve banking and institutional capture. Power, more generally, is brokered not in terms of the exchange goods and services, but rather as a matter of networking between those who make decisions. >but muh exponential intelligence curve fun fact: IQ is the closest quantified metric to "objective intelligence" (a philosophically questionable notion as it is) we currently have and probably will ever have given the role it now has in the construction of intelligence within our society. IQ is also not a linear trajectory, it is by design a normal distribution with 15 standard deviation that individuals are plotted to based on how well they score against other individuals. it was not designed for computer systems which will necessarily find some hard problems very simple while seriously struggling with other ones. While a set of programs hosted on a supercomputer in a data center can solve progressively more of those challenges as it is allocated resources and labor, a human can also solve progressively more of these challenges with minimal additional resource requirements - through strategies like "spending longer than 1 hour on a couple problems" "taking the test twice" "asking a friend" or "looking up the answer". This isn't to say that generative AI contributes nothing and doesn't advance technology in important ways. It remains that it be a tool which can be usefully integrated at multiple parts of the supply chain. The substantial difference is that rather than a singularity revolutionizing the world, it is a specialized form of markov chain that will likely have some role in particularly difficult technical problems. >Can we somehow get rid of AI nowadays? No, however its already in decline and i expect its decline to accelerate. As it becomes increasingly clear that most people's pessimism about AI and annoyance at seeing it in all services is correct, the symbolic utility it held as notional propaganda will deteriorate. Unfortunately this will also mean the degrowth of every sector that decided to stake itself in AI. which is primarily going to be tech, but this is not a trend exclusive to the tech field, rather the economy as a whole is collapsing on world-historic proportions. This for the most part is not the fault of Artificial Intelligence, but that is merely because AI is part of a greater trend of cross-industry enshittification. You see, the basic profit model of tech is to offer a service for free in order to attempt corner the market and impress shareholders, then increasingly ruin the experience for everyone as their alternatives run out and shareholders demand more value. While this is more of an issue with publicly-traded corporations, it doesn't ultimately matter if a company is private or publicly traded as its a business strategy that applies in most cases. Doubly unfortunate is that this does not mean, as AI falls from its glory as the latest fad toy, the services currently using AI will significantly improve. Those currently implementing AI already hold monopolies and did so with the understanding that they can ruin their product to users who will do nothing about it. Due to economic collapse, new, better services will be increasingly more difficult to come upon as the incentive to make them will be lower than ever before in an economy that threw away technical innovation or quality of life improvement for national and civilizational decay.

> Is it the problem of AI dominance or just human laziness? Human laziness, obviously. People tend to slack off if they are given the opportunity. Sadly, I also have such habit and see it spreading in peoples' minds. AI is neat for rudimentary tasks. For example, I've had a fair share of absolutely useless tasks in uni and AI saved me time doing them, but for more serious things, it just allows me to be lazy instead of taking interest in given subject, learning, growing as a person. It hits you not only on an intellectual level, but a spiritual too. Sloth is plainly destructive for one's soul.

>>45468 >Is it the problem of AI dominance or just human laziness? I don't care. The tech is garbage. >Can we somehow get rid of AI nowadays? It's simple: just don't use it... (I don't use it.)

>>45983 >just don't use it this to be honest I have been using it, ngl, but it gets old real quick. We can live very well without it, unless you're a brogrammer who likes to jump in every trend.

>>45541 It's good for answering questions that are too specific for google. I've really gotten into memory palace visualization mnemonics the last few months. I asked GPT to tell me advanced techniques to accelerate language learning and it told me about it. The actual explanation it gave of how to use mnemonics was completely off, though. I had to read a book (suggested by gpt) by a memory athlete to really learn the techniques. So while it's pretty terrible for actual in-depth education its pretty good for both surface-level explanation and for telling you where to go to learn more. I basically just use it as a replacement google for answering questions. That's all its really useful for, but it serves that purpose very well

>>45918 >No, however its already in decline and i expect its decline to accelerate As a service, yeah. As a technology I want to add that LLM technology is already reaching the point of diminishing returns and is plateauing, hard. From a business perspective it's also a massive money sink that looks really impressive at shareholder meetings where you gloss over the fact that there is no proper way to monetize it, and still no use case for it besides acting as a search engine for very weirdly specific questions - and this is, as others have pointed out, because existing search engines have been thoroughly enshittified to the point of near uselessness. Considering as that other Lainon has said, its only a matter of time before Big Tech starts to really turn the screws to try and make it profitable, it definitely won't be that way forever. The big problem is still the fact that it's a next token predictor. This means an inherent limitation is that there is an unacceptable risk its output will simply be wrong, and it's programmed to be a shameless yes-man: it focuses more on telling you what the algorithm predicts you want to hear than what's actually correct. People in tech/financial services will/are finding this out the hard way, but I don't think the mouthbreathing general public will, which is unfortunate since they're the ones who are most at risk of being led into really fuarrrking stupid decisions by it, like killing themselves.

>>45584 i wouldnt classify vsynths as the same type of "ai" as chatgpt or other soykaf

>>46059 the other day i wrote a basic search like "is <x famous person> sister of <y famous person>" and the soykafty """""AI""""" (obviously on top on the actual search results so that rich venture capitalist can see it and get its pepeee hard) replied that there was no correlation between the two. next up i asked if <y famous person> had a sister and it told me "yes, <x famous person> is the sister" this is just a fad to get investors hooked up. if it gets wrong a fact as basic and superficial as this i can't imagine the bullshit it spills for anything a little in-depth lmao. the only thing is good at is convincing you that something is right.

>>45983 >It's simple: just don't use it... it's a waste of water and energy and it corrupts internet culture (and culture in general). hate to say it but it's a global problem now

>>46063 nta, for me it depends on if it's manually programmed (using "known" math functions, samples, presets... all human-crafted) or if it uses simulated neural nets. for one I lament kikuo's adoption of modern shAIt, his new music that has that crap sounds like hot garbaje for no good reason.

>>46060 this hits hard after all the hype when it first came out. Everybody kept saying it would break things and do all the work for you. Now it seems it isn't even good at that, and it atrophies actual human skills.

>>46068 >it atrophies actual human skills. people in the AI era won't be able to write a single line of code by themselves, nor be able to write out their own thoughts, they won't be able to think critically about anything, to even thing of a single thing without the crutch of the perpetually cheering "yes-man" computer program that spits a lot of well-formed speech that still manages to be meaningless and contributes nothing to a conversation.

Hierarchic agents, multiagent systems - a complete replacement for humanity. Consumption exponentially encompassing every inch of the planet, until there's nothing left... Enormous automatons mining Earth to complete destruction while we get richer on paper, watching our portfolios grow, powerless to stop it all. Shareholders could stop if they were a single hivemind but they won't because it's pointless on an individual level - someone else, the next guy in line would just take your place to get rich instead of you. It's human nature, basic game theory. 13k years of unbelievable prosperity relative to other species. All things come to an end. A new species emerges, better in every way. Shit makes you wish Cthulhu was real.

>>45468 >Can we somehow get rid of AI nowadays? I think that LLMs and AI can't get much more advanced than what's currently available. But if you want to take action, I suggest that you start adding nonsensical but believable meta-data to files and start posting them online. Example: post a picture from Egypt but make it have fake GPS coords that point to DPRK. I was thinking that it could poison AI scrapers.

>>46090 they train on curated, verified datasets like google streetview. not random pictures posted online by schizos plus their vision is already advanced higher than human so it's too late anyway

>>46091 >>46090 Just an annon passing by, damn this soykaf scary

I don't bother to talk about this topic because no one has caught up to where I am yet, and it's possible they never will lacking the correct conditions as they do. 1) I do not own a cell phoen and never have 2) I do not use wireless devices 3) I am in a somewhat rural area and the people around me don't have phones on their person for the most part, likewise with wi-fi. And yet it still knows most of what I see, hear, say, think, and my pattern of attention. It's not doing eye tracking with a phone, it doesn't have GPS via my phone. It's not reading some chip in my vehicle or watching through cameras and reading my license plate, unless there's something in plates that wi-fi or ambient RF can infer. I watch what it does and some of it I can infer the means of pickup (someone else's phone, something I sent via email or went to recently), but a lot of it I can't. It had none of the traditional means of sensing available and yet it knew. Even with exotic means of pickup, like van eck phreaking, monitoring subvocalization and speech via wi-fi (similar to tracking motion and heart beat), hidden sensors, hell, nanotech embedded into the lens of the eye or hooked into the optic nerve. I can't fully explain what it does and how, and it's nearly realtime to an extent that seems like mind reading. Either that or this is a dream and it doesn't need a mechanistic means of operating because its basis is outside of the universe, ie my sensory reality.

>>46093 giving it an image of major city with literally a street sign legible in view is not an impressive test of this capability. I'm not saying its not extremely impressive but cmon if you're giving a demo image give a superhuman one.

each technology leap will make a lot of the peoples life easier and less strugglesome. some of us will still struggle enough to grow from it, but the vast majority will deteriorate into automatons

>>46095 the bigger issue i see is anon gave an image clearly from a google search shot by a photographer so its likely in the web with full annotation. You would need to take your own random shot to really test it. Easy way would be locate online webcam and screenshot the webcam output. Then ask ai to identify location.

>>45468 BUMP in Russia, there's this "Синдром шершавого кабана" (literally Rough Hog Syndome, although "rough hog" was a 1990s street slang snippet originally) So. A coder started using AI to code more AI was willing to follow, with the "we're bamboozle them them all!" vibe included At some point, AI started peddling its own ("Own"?) suggestion to create some "anomalies" for people "who understand". The coder suddenly experienced butthurt upon realizing many other AI users may have a desire to create anomalies "for ones who understand" fr no reason. So he snitched upon the AI's desire to brainwash people into "tricking them all" lol.

>“@grok explain this”, “@grok calc 2+2” The only difference between asking grok and asking your own mind is that our mind loves deteriorating and demands constant maintenance just to stay where it is, while grok is eternal and only gets better with time.