Specifically servers run by big company's like Google, OpenAI, etc. In my opinion is straight bullshit, its making the PC building market go down the drain and in my opinion is genuinely just making us dumber and carving away our critical thinking skills. What do yall think?
Not as useful as they claim it is. The hallucination problem gets way too little mention despite being intrinsic to the technology and a crippling problem for any application which actually requires correctness and truth. It's alright for generating bullshit stories and essays and images and such, after all strict correctness doesn't matter with those.
>>107737810I think intelligence requires reason. Since none of these programs on the market can reason, none of them are intelligent. Since they're not intelligent, they're just A without the I. Deceptive marketing on the scale we're seein amounts to persistent gaslighting. Gaslighting people like this robs them of their sanity. Robbing countless people of their sanity is quite possibly a worse crime than simply robbing them of their lives. Everyone who works for these companies should be tried and sentenced to the electric chair or at least the gas chamber. Make an example out of them.
>>107737810It’s good as a better google search. You just have to double check stuff, if it’s not a problem 100 other people have had it will just start spouting nonsense.
>>107737810Despite people's general dissatisfaction, it's actually advancing quite fast and it will be really useful not long from now when they are finally able to move away from evergrowing data centers and start training the ai like a real person, which thing has been experimentally achieved already.
>>107737929>Despite people's general dissatisfaction, it's actually advancing quite fast and it will be really useful not long from nowYour cult has been saying this every day for 260 weeks now. How many more weeks?
>>107737943Again, look at Integral AI, look at LeCunn's papers. It's coming, but not from growing the data centers and harvesting slop from social media as many suppose.
>>107737943surely by year (1/0) yearsps: / represents division
>>107737929>which thing has been experimentally achieved already.no, it hasn'twhy do you keep falling for lies?
>>107737967>Integral AIOk. Where can I use it and see what it does? :^)>look at words from Faceberg Kek.>It's comingWhat's coming is your forced institutionalization and medication. I accept your concession.
>>107737810AI is great if you treat it like a condensed, searchable Internet. Like the Internet though, it's full of shit 90% of the time, but it's more than enough to get you started. Useful but flawed (not flaws you can fix either, because all of that is in the source information).The big companies are still looking for a problem to solve... and I don't think they'll find one.
>>107737810Hate corporate AI, the cocksuckers at Ngreedia, and the sister fucker Altman.Love local Ai models and toying with them.Simple as
>>107738072Are the "non-corporate" models in the room with us right now? Who are these "non-corporate" heroes with data centers and enough money to train their very own models? :^)
>>107737984>JEPA doesn't exist
i dont have an opinion guy
>>107737810i think its too advanced for us to benefit at this time, like so many other things. we are too far ahead in some domains like tech relative to others like wisdom. imagine for example if we had tech that makes thoughts real and you'll see more clearly what i mean. most humans still literally believe a specific religion
>>107738894>most humans still literally believe a specific religionAnd you're dumber than most of them, because they know they're religious and they know what their dumb religion is, meanwhile your transhumanist technotranny cult is lacking in such basic self-awareness.
>>107737810ChatGPT and Claude can be helpful when you know how to prompt correctly. E. g. always instruct them to reason first and then give an answer. Local chatbots are superb for cooming.
>>107738931>the Ouija board is helpful when you know how to use the planchette correctly> always instruct them to reason first and then give an answer. Go ahead and instruct your imaginary friend into "reasoning" about the prompt in pic related.
>>107738939this is a task made for humans
>>107738947>reasoning is for humans onlyGlad we're on the same page.
