when do you think it will be socially acceptable and safe for NPCs to finally accept AI? what would need to happen for them to suddenly do a 180?
>>107718779Younger generations already do. AI simply is the future in a lot of industries. It probably won't fully replace people like many low IQ normies say, but it will be integrated into many peoples workloads and those that can't adapt will be left behind. It's very similar to when engineering moved from manual dafts to autoCAD. It didn't matter how talented of an engineer you were, if you couldn't adapt to the software your lack of efficiency compared to someone who could negated any value you had and you were let go. Give it 5-10 years and people will be posting screenshots of post by AI detractors and laughing at them.
>>107718942Cool post but CAD fucking sucks and I'd much rather still be drafting by hand. The last people who did manual drafting in my office retired around 2008 or so and I still miss working with them.
>>107718942would
>>107718961Gtfo with that geriatric opinion
>>107718779They overplayed their hand. Instead of a slow drip feed of shit they tried to force it down everyone's collective throat violently and all at once. They'll continue to implement it into everything usecase be damned until the normies relent, lay down and grudgingly accept it as the new normal.
>>107718779FUCK THIS GOD DAMN CAPTCHA I HATE IT!
>>107718779to finally accpept? as if they aren't using it daily
>>107718779>what would need to happen for them to suddenly do a 180Cute robomaids that jerk you off
>>107719190how does it feel knowing the government just printed money from nothing, stealing your purchasing power, and are making it unobtainable to buy a new personal computer while it still refuses to parse a 4chan thread
>>107718779Probably seeing some actual useful features. Why companies are spending trillions of dollars just to replace some basic algorithms that could be written in a day with incredibly inefficient AI I don't know. You don't need a billion dollar data center to automatically group tabs in a browser.It is pretty telling how it's just an investor money sink currently with how they have spend like the last 4 years scrambling trying to figure out actual use cases for AI before giving up and adding yet another AI powered summarization feature and ChatGTP chat popup nobody will ever use.
>>107718779It doesn't matter.AI and robots will be forced upon the disgusting subhumans whether they like it or not.At that stage they can either accept their position as a slave to a bot, or kill themselves.
>>107718779what does ai actually do? ok i ask chatgpt instead of googling nowadays. but can it actually do something? its coding ability is dogshit for something more than a toy snippet you could just copy off github.
>>107719643well, it's still in its infancy, but give a couple more years, and the ability to learn by itself, and it could run millions of tests per hour, discover new technologies, new cures and such
>>107719846>but give a couple more yearsAnd a few more trillions and we'll have AGI in no time. Two more weeks, trust the plan.
>>107719846this technology, by design, cannot learn "by itself"couple more years is more than idealistic lol but you're free to prove me wrong by showing me the breakthroughs that enable your imaginary future
>>107719846>it's still in its infancyinitial chatgpt gave me code snippets of the same exact quality as the version they have up right now. it seems to me, that transformer is just good at fetching kinda relative information, but there's no actual "ai" there.>the ability to learnthat requires completely new hardware. our brains work like they do because you can't do intelligence the way ml industry does it right now.
>>107719846So we're in the "just N more years and X billion more dollars" part of this now
>>107719643Z.ai does my work presentations for me now.Pretty fast, what used to take to take half an hours takes 5 minutes of simple editorial work now.Also AI works with spreadsheets, so grunt of the standard corporate office work is already covered.Also, when your boss asks your for a data review of your project, or some project management tools like Gant charts, it's pretty easy to save some time and ask AI to do it, especially if it is something not so trivial such as everyday bullshit office work.
>>107719911>Also AI works with spreadsheets, so grunt of the standard corporate office work is already covered.no it doesn't, no one uses this because surprise surprise, a guessing machine that is known to be bad with calculus is not trustworthy when the expected output is supposed to be precise and correct
>>107719911so you'll get fired and presentations will be just slop?because here in software, i still haven't seen a model which would've been able to code something useful. i can ask it for a bunch of snippets, and possibly even talk to it in order to indirectly vibecode the code myself, but that's harder than actually writing the code.i mean i can see how you can generate low quality slop, but if your job was to produce low quality slop it wasn't a real job anyway.also that's not what sanakan looks like.
>>107719923Depends on the model. Latest Gemini 3 Pro does hallucinate a lot less.Also, I sometimes wonder how the AI manages to understand the data holistically and sometimes even recall data points or oddities that I would have missed (for examples once I though it hallucinated a lot but later found hidden sheets in a spreadsheet shared by my colleagues, which actually made a lot of sense for what I was inferring in terms data analysis).
>>107719952>prompt in broken englishSSSSSSAAAAAARRRRRRRS
>>107719952>does hallucinate a lot lessyes but can it do an actual task for you? to me ai still looks like a lot of gacha where you hand pick slop you could then work with.you can counter that slop is good enough like your hoplite example for some journal illustration, but that just speaks to the job's expected output being crap in the first place. and if everyone has access to nano banana, you still need a human editorialising the result in order to compete against all the jeet pooroduct mills.i believed we'll have AGI shortly after llms can reliably call tools, but they're still useless and can't figure out anything for themselves. you have to handhold the model for it to do anything. and frankly professionals in their field will probably find doing the task directly faster, certainly is this way for code.
>>107719951well, most models used to suck for any coding purpose until the release of Gemini 3 Pro with its near infinite context window.have you tried GLM-4.7? It's pretty cutting edge and free (being chinese and all).>>107719987>yes but can it do an actual task for you?with each new model the complexity of the task it can do increases. at first it was simple proof-reading or writing text for memos, being the simplest in terms of complexity.later it helped design frameworks or propose what data visualization would be more appropriate for specific data.with latest models it gets pretty cool actually (you can check what the latest model can do in this user showcase video https://www.youtube.com/watch?v=KaWQ2Ua9CW8)
>>107719987Speaking about slop. Why can't Japanese artists draw literally anything that doesn't copy and paste the same face and facial structure with overly designed clothes or hair to try and compensate? I know they're a bunch of conformist insects but Jesus Christ, man.>dragon character looks like cute animu girl>robot character looks like cute animu girl>elf character looks like cute animu girl (with long ears maybe)>alien character looks like cute animu girl>warrior character looks like cute animu girl>post-human character looks like cute animu girl
>>107719971the fact that you don't really need to correctly spell words for AI to understand you is neat though.Sometimes it even gets scary how AI can understand what you wanted after you've accidentally sent your prompt halfway while typing it.
>>107720020>well, most models used to suck for any coding purposei dunno, i got really good snippets for my shaders out of chatgpt on release. haven't seen it code anything more complex.> GLM-4.7why? i have gemini claude chatgpt grok fo free. they're all the same.>with each new model the complexity of the task it can do increasesnot seeing that. what can it actually **do**? because i can fetch information through it, but as soon as i want it to, for example, come up with an algorithm for me, it shits the bed hard. creating slop text isn't a real task, task means it correctly solves some sort of a problem.>helpedthat's basically just googling.>>107720024well japs arr rook the same, what do you want.
>>107720037>the fact that you don't really need to correctly spell words for AI to understand you is neat thoughNo, it's like paying for internet access for africans and jeets. Nothing good will come out of this. It'll only benefit """people""" that will actively ruin things for everyone else.
