I mean, I think there are some clear biases when it comes to AI, people who think it's going to end the world, people thinking it's going to take their jewbs or that it's literally mechaHitler.But we also know that the amount of risks have quadrupled, people using it recklessly and all.But in your perception, will AI be gone in a few years? Or it's just a biased cope
>>107645769>will AI be gone in a few years?no its an employers dream come true, its cheaper than hiring slaves. its here forever
>>107645769remember, after the dot com bubble, that was the last straw for the internet. never to be seen or heard from again.
>>107645769Some current applications are useful and will stick around. Maybe they'll get more expensive when investment money dries up. There's no reason they'd disappear altogether now that they exist.Some current applications are pointless and will disappear sooner or later.AI will keep improving and people will keep getting better at using it. Nobody knows how much and for how long.
>>107645878Remember, after the social media bubble, everyone is happier, more social, and mentally stable.
AI has the same impact as the computer, it can be used for almost everything and will be.
a lot of companies will go under, since they're not even close to making a profit. But some form of AI will remain. I assume every tier of service will require a subscription however
AI is a passing fad for normies and single users, for businesses however you wont have a choice
>>107645769It's genuinely dot com 2.0. Yeah, those guys back then were right that the internet was the future, but they hadn't quite wrapped their heads around ~how~ yet. Most of those dot com era companies are dust now, but the internet is bigger than ever.All this bullshit will be the same. We're going to have global AI palantir control grid precrime bullshit before we have robotic assistants though.oh and when we finally get robotic assistants they'll be subscription only and they'll spy on you
>>107645769I've seen normies call out people using chatgpt to format text. Fatigue is already setting in.
>>107646968If anyone believes those billionaire faggots are going to pay for them to sit at home playing video games they need their heads read.Fuck no, we'll be fighting stray dogs for scraps in the streets before Elon's star trek ass future materializes.
People mistakenly compress the entire field of ML / AI with LLMsLLM is one approach to it, the most public and recognizable.People see that half of the time these bots shit the bet. The Companies seem bubbles ready to burst and the CEOs feel like low life grifters. The AI bros are basically the crypto bros that jumped ship.But imo there was some real, solid and fast engineering progress in LLM in the last two years and I think there is more progress that can be quickly made in 2026.And then we have the big companies +other labs + the chinese working openly or in secrecy on hybrid solutions or verified RL or something else we don't even know. It's a hot area of research, billions are getting poured on it and progress is made in several different directions that might lead to more progress.So it can still be a bubble but most people talking bubble are referring to the one time they tried gpt in 2023 and shat the bed
>>107647243If anyone watched OpenAI 5 (Beat DOTA2 champions), that shit was lit. That was actual "learning" over time. Then for some reason, never a peep from that project again. I'm guessing a military blackout/embargo.
>>107645769LLMs are a dead end and there's no indication that this thing will not fail, unless a sudden breakthrough happens and we get useful AI
>>107647269the gaming shit was always just a test bed to integrate it into other walks of life. There's no money in just making a really good Dota bot
>>107645769AI will never go away. That is the problem with it, that is why people were warning against it. It can't be put back in the box, it's too late. What IS going to happen is a speculation bubble will pop, much like the dotcom bubble did. Just as normies were discussing websites like they were basketball cards, people are now doing so with AI: But it is not profitable, and is built on pillars of sand. The bubble WILL pop, and it will likely pop soon, the question is just whether or not governments will draw it out by bailing them out continuously. But it will happen.
>>107645769AI is useful to a lot of people, namely as a personal assistant to ask questions to, it's basically a better Google for many. the problem is that the growth in its capabilities are flattening meanwhile the growth in its investment (both money and hardware) continues to grow exponentially anyway. eventually money people are gonna question if trillions of dollars and a good chunk of all hardware output were worth it to make a marginally smarter and snarkier version of GPT that barely grew in demand, and if corporate "AI enablement" actually did anything to help companies become significantly more profitable or if it was just a buzzword to draw investor attention and distract them from tightening and layoffs
>>107645769In my opinion it's a biased cope. AI is just too fucking good at making porn, texts, or acting as a lossy recommendation algorithm for information from books. People claim AI is retarded or is just regurgitating information but they fail to recognize that 90% of knowledge work is regurgitating information, pattern matching, and substitution. Instead of looking books for weeks or googling for days, AI returns it in minutes. AI is a general purpose technology like mechanical engines (steam, combustion) or electricity. Faggots will tell me I'm wrong but that's a seething cope. However theres a real and unfortunate shift towards shoving everyone in the cloud as fast as possible, because AI is too powerful for the powers that be to want it to be in the hands of Joe Public. If you can't filter, control, or throttle what Joe Public does on his own general purpose hardware, he could be producing hardcore porn of his colleagues or writing racist hate speech... or write unfortunate truths for the powers that be, or code apps which circumvent technical restrictions. Even the AI companies themselves were scrambling to contain the powerful pattern marchers they had developed when they produced "wrong" kinds of content. AI is not going away. Do I believe a stochastic pattern matcher will develop sentience or independent creativity? Probably not - but they'll wring it awfully close using workarounds and scaffolding just like how they made real time 3D graphics with shadows with polygons, not raytracing, for the longest of time. And that'll be good enough.
>>107645769I think AI will remain stagnant for another 10-15 years. Then someone will come up with a new architecture that makes agents viable.
>>107645769AI and ML is very useful and going to be around for a long time, but it does not need nearly the level of compute that OpenAI/Google/Amazon/etc. are amassing and the LLM goldrush is going to eventually collapse. Jobs dosen't concern me as much as the spying and data gathering on everyone happening at an accelerated rate, and I really hope a pop puts a dampener on that.
