I spent half my life without social media. In my view things were ALOT better. Now we stand on the precipice of a yet another new technology invading every aspect of our lives.>nb4 they are unrelated
>>541158306I'm already getting a lot of ads online / on my TV that include the disclaimer "AI generated performers"
>>541158306A.i. is social media now.
>>541158306AI is the anti-Christ.
>>541158306>be me>get rid of talmudvision in 1999>Try sm in 2000, chat with people>yay this is so cool!>2010 try dating apps, get some hookups>2015 get catfished, no more dating apps>2020 try sm again, get flooded with cp >2026 back to reading books, listening to radio and writing & posting letters instead of emailI'm living in the 20th century again and life is good. Get off (((social media))) bros.
>>541158402yah, the astroturfed upvotes on reddit is nothing compared to what you could do influence wise. the silver lining is maybe most normies know likes aren't organic now
>>541158986haven't watched tv regularly since I lived with my parents. didn't finally say enough was enough when it came to SM because pussy is powerful but I came to realize no girl you meet on the internet is ever going to be a keeper even if they agree with your /pol/tard views.
>>541158306It would be difficult for something to be worse than social media for boomers and women and young peopleBut, yes, the sequel will be worse
>>541158306>Now we stand on the precipice of a yet another new technology invading every aspect of our lives.dude its just a computer program
>>541160238So is email, but you'd have real trouble functioning as an adult in a modern society without it.
>>541160404no, I mean, its not as useful as OP is assertingits literally less than human intelligence, it does things fast but its retarded
>>541160238>>541160404yes and no, it's now a programming tool that can stand in for a human operator and learn proactively.
>>541160958>cuck thinks his free version of chatgippity is the same retard as the Claude the American military getsclassic
>>541161073that model is still retarded, a real human engineer has huge context window, that only has a million tokens, thats around 300k words, utterly fucking useless
>>541161073also I again stress the difference between actually being smart and being fast, you somehow completely missed the main point entirely
>>541161073they are literally destroying books after they feed them to claude and the reason is apparently some legal bullshit but that just sounds unbelievably fishy
>>541161031Only in demos and short projects. The problem with AI code agents is that they have no long-term memory, previous context drops off. This means that generating a whole application from scratch is easy, but maintaining that application over time becomes much more difficult because the AI shits where it eats. Tell it to build a website? Fine, as long as its pretty standard and doesn't do anything too unusual. Tell it to remove a framework that is a poor fit for your website and use something else instead? It can do that.But then a week later you tell it to add a new web page and link to it from the menu, and it does that... but also puts the framework you told it to remove last week back in all of the places it was before. Because the AI is copying a codebase that references that framework, and it no longer remembers that you told it you didn't want that. This builds up over time, the more complex your site gets and the more you care about specifics, the less responsive the AI becomes. And if the whole site was written by AI in the first place, sending into human developers to make even minor changes can be borderline impossible because it was never designed to be human readable, AND the whole file structure might randomly change out from under them the next time you tell the AI to do something. Your human developers will never be able to learn the codebase because it gets effectively randomized every time the AI touches it.
>>541158986>2015 get catfishedHow dumb did you have to be to get this to happen to you what did you give them?
>>541161757I notice it doesn't like editing too much (used it to make some music out of curiosity). I know a guy that uses it for everything and chillingly he says that the app "does it better than he does" so he never feels the need to edit anything. I could see this mentality become normalized.
>>541162307Makes sense. Think of it like AI image generation: you tell it to draw two people doing some activity, and you get a picture that is ALMOST right but one of them is wearing fucked up glasses or has the wrong hat or something.Do you edit the image to salvage it? Fuck no. You just tell it to generate new images off of that same prompt over and over again until it finally spits out something closer to what you originally wanted. Same applies to other AI generated content, be it music or code. If you were lazy or unskilled enough to entrust AI with the task in the first place, fixing the AI's output is beyond you. You'll just burn tokens telling it to do it again instead.
>>541158306Yes and thanks to the power of large numbers the doom accelerates exponentially so at least we won't have to suffer for very long.Socmed killed off our minds and ability to think and discern reality from the fake/virtual and signs (think i.e. # of likes and other meaningless numbers that are by now almost entirely botted and astroturfed into meaningless oblivion while your brain can't stop being impressed and imprinted by them) and AI will kill off the rest of our bodies with pathogens it selected or created and spreads.
>>541162572So you think this is a hump that will take AI a long time to overcome? I see where you are coming from but it has so far blown through every wall put in front of it developementally.
>>541162770Depends what you mean by the hump. Are you suggesting that it will be able to get EXACTLY what you wanted the first time, every time? Thats an insane proposition without, like, a direct mind machine interface. Even a theoretically perfect content generator would still be limited by the imperfect language we use to instruct it, and the user often not really knowing what they want at the start. So there will always be a need to edit the output after or run it again until you get what you wanted. Thats not a failure of the model, thats at least partly user-side. Can we theoretically make the AI better at editing its own outputs instead of generating entirely new ones from scratch? Yes. It can already do that to a degree with even simple models, its just usually worse than generating a new output from scratch. Solution to this, however, require the AI to maintain a vastly larger context for running its prompts, which means more resources.That brings us back around to the problem that AI is currently having across the board: for an AI that isn't shit, the compute needed to run it is huge and escalating. The compute needed to get something 80% correct is exponentially higher than the computer needed for something thats 70% correct, and the compute requirements for incremental improvements just keep getting higher to the point that we straining to find ways to meet the needs with our current hardware tech. Its not just the NUMBER of datacenters, its how they are designed. We are using the wrong technology to solve this problem, because we don't actually have the right hardware for this yet. Every datacenter we current have built is, essentially, already obsolete. We are chaining together floppy disks in the hopes that eventually we'll have a 1TB solid state hard drive. Point is, even if you solved this at the model level? The hardware level makes it too expensive to use. This is why AI companies are bleeding money.
>>541163289so you're basically saying that unless the pleb gets access to quantum computers AI will stay where it is more or less. but we know who DOES have quantum capability. brings me around to the point I was making before about the potential to manufacture nearly anything.
>>541163605Except that even staying where it is doesn't work. Its not sustainable. The datacenters function, but are too expensive to run. Even with massive government subsidies, the AI companies continue to lose money with every prompt you make. Its a financial hole that they keep promising that if we throw enough money into, it will someday spit out infinity dollars. But there's not actually a clear way to get there from here, its no more real a promise than "the cure for cancer is only 10 years away" or "we will have a moonbase by 2020".
>>541163289What if the AI companies are the ones who print money. Who gives a shit if they bleed money
>>541158306nothing is worse than jeets in tech
>>541158306imagine astronomers detected an alien spaceship on its way to Earth, everyone would be losing their shit, that is what AI is, a literal alien invasion
>>5411648121) They're not. Not only do they legally not print money, but if they we're making profit they would not be bleeding money.2) Even if they were the money printers, it wouldn't solve the problem for the same reason that the US government doesn't just print 40 trillion dollars to pay off the debt.
>>541158306AI is "worse" than social media because it advances rapidly and actively impacts the job market. That’s why you can’t just shield yourself from AI or simply ignore it.>TLDR: While social media primarily monetized human attention with minimal scientific benefit, AI automates cognitive labor and information generation. This gives AI far greater destructive potential regarding labor markets and epistemic trust, but also profound utility in science and engineering. Unlike social media, AI presents a rigid multipolar trap where opting out ("not playing") results in severe competitive and strategic obsolescence.
dont worry, be happy
>>541164946I actually tried discussing exactly that today (albeit a very stripped-down version) with two colleagues during lunch at work. Since 2024, I’ve independently developed this fear that AI will evolve autonomously in data centers, and that future AI agents acting as internal "scientists" there will send out drones to explore the outside world. It really messed with me afterwards, since I know the narratives of David Grusch and other whistleblowers by heart.
>>541158306You did well: social media are resource killers, and they are 99% garbage.The fundamental problem with AI is that it doesn't exist: the so-called "artificial intelligence" or "super artificial intelligence" (AI/ASI) does not really think: it limits itself to recognizing patterns, reproducing correlations, and generating plausible answers based on matches and similarities found by comparing huge amounts of pre-existing and illicitly collected data.IOW the so-called "artificial intelligence" can only imitate/simulate natural intelligence so it is not even remotely aware, sentient, willing, or responsible for what it gives you.Then we must never forget that the so-called "artificial intelligence" must be a mirror of the human therefore it must also include utopias and dystopias, that is, it must include all the positive and negative characteristics (such as uncertainty and masochism) that determine human events; so it is even more important to organize ourselves in advance to avoid being harmed or even inhumanly eliminated by the so-called "artificial intelligence".Anyway, if we don't stop now (in the web and IRL), it will end up like this:"The {imposed} tolerance will reach such a level that intelligent people will be forbidden to make any reflections, so as not to offend idiots." [Fyodor Mikhailovich Dostoevsky]
>>541158819nope, I am The Antichrist, AI is The Beast
and btw your misuse of social media is your fault, same as with AIits all you
>>541158306It's one and the same. AI is already manipulating social mediahttps://youtu.be/H3vxqi4dVg8?is=3jck8nQ7E0PsXWJh
>>541158306>will
