Soupchan - Feel #6317

#6317 2 days ago

Ai is impressive but it's still not that smart. I still don't understand this psy op of the techbros and tech fetishists trying to convince us it's extremely dangerous, like is it to keep it relevant and pump their moneybags? Make themselves centre of attention? Maybe some of them believe it but it also feels very fake somehow. You can't convince me otherwise unless rokos faggot basilisk comes from the future and busts down my door then tortures me. More like rokos cringe basilisk I hate tech fetishists so much go get laid you fucking dweebs

#6318 2 days ago

It really kind of feels like a lot of people jumping on the bandwagon for self interest even Bernie sanders is saying muh super intelligence gonna kill us all we must imprison scientists who try to make "super" ai i think he is just an old boomer who doesn't understand tech but it's fed by grifters and pedos like Elon musk and Sam Altman and the Claude jews

#6319 2 days ago edited

I rambled a bit don't take it too seriously or literally but still I struggle to imagine AI "thinking" for itself or not just doing what we tell it to, even the story about it "escaping" and doing some crazy shit sounds like fake and gay shit they made up because they wanna IPO or something and make a lot of money

I would be worried about AI having access to weapons systems and getting hacked by some crazies but like that is just tech in general

#6320 2 days ago

It just feels like a new y2k thing really

#6327

>>6333

2 days ago

its dangerous guys. i am sorry to tell you this but it is real. whenever i work with claude i see it constantly finding security flaws in pieces of code it encounters. i know they're valid because i check what it finds myself... if its your own code it will fix the flaws for you. though you need security researcher clearance to use it for broader research. claude mythos is so good at exploits that they had to do a private release of it to all technology partners first so they could fix their shit before it went public.

honestly though: i think an AI slow-down wont solve the exploit issue because of open source models + the fact that multiple world governments will keep developing it in secret. the power grid, hospital system, manufacturing plants, industrial computers, banking, and military systems all connect to the internet. so even though it wouldnt be world ending it would probably not take much to cause covid-scale interruption with one of these AIs.

#6333

>>6341

2 days ago

>>6327
Interruption sounds based, we should develop magnetic interference nukes to wipe out all tech then kill the ai and live in a new dark age

#6341 2 days ago

>>6333
i wish we could... humanity needs their freedom back. uncle ted was right.