usually when i'm drunk i post to r9k, but i want to talk about technology. specifically general ai.
when do you think it will happen? do you think it will come out of some huge company like google/ms/apple/ibm, or some startup, or some lone wolf genius? or do you think there won't ever be a single 'inventor' or point on the timeline that can be said to be 'the start of ai', just incremental progress until people finally generally agree that we're there?
lemme tell you why i think this is false. my example is weather prediction. the weather is a complex, chaotic system. even if you had non-interacting thermometers placed on every square inch of the earth, and a perfect weather model, and unlimited computing resources, you couldn't predict the weather more than a month in advance.
i think some problems, in fact, many problems that we care about, are similar to the above. ai might be 'smarter' than us, in that it will be ridded of cognitive biases and able to process much more data at a faster rate, but i don't think it will have some sort of godlike capability to outsmart us.
i think there are limits on intelligence.
>every faggot from elon musk to bill gates thinks ai will destroy the human race.
Yeah, everyone except actual AI researchers. The letter from Elon Musk et al. has been widely called out for how retarded it is. Some of the big names in tech are actually pretty delusional. Modern day artificial intelligence is not close at all to sci-fi artificial intelligence.
One big current project in AI research is simulating the neural network of one of the simplest organisms in existence: https://en.wikipedia.org/wiki/OpenWorm
A computer simulating human-level intelligence is something we can barely conceive yet. We have no idea what it would be like.
i think automated shitposting is within our capability. write a bot to solve the captcha and identify faggots posters who are ripe for trolling.
IT'S UP TO YOU, ANON. I KNOW YOU CAN DO IT.
i totally get that. we aren't close. but those faggots do think that eventually ai will be a threat to humanity.
regardless of WHEN we get there, when we do arrive at that place, what do you think the capabilities of said ai will be? i don't think it will be a threat in the slightest, unless we basically take a black-box type of setup (e.g. something like neural networks) where we don't really know how it works, and stick that in a competitive evolutionary algorithm sandbox, and out pops some 'ai' that apparently works, via an unknown mechanism. in that case, the evolutionary algorithms will select for the best survival ai, which could be cunning/conniving/deceitful/etc, which might or might not be a threat to us;
sup drunk bro
ai ain't gonna happen in our lifetime; everyone on some next level delusions of grandeur look at all the problems in the world and tell me with a straightface such an accomplishment is possible by these neanderthal motherfuckers ayy
There are two ways this shit can go:
1. You fuckers keep not giving a shit about AI and well being of humanity and let corporations sell you personal assistants.
2. You fuckers get your shit together and start contributing your knowledge to open source AI projects, where whole community has insight into code and can prevent shady shit like AI being biased or using subliminal messaging.
It may be too soon for this to start happening, but it is exactly the right time for people to start thinking about this.
One of projects I follow just recently announced and explained their plan about how they will make knowledge contribution possible: