i never said a literal human would destroy ai

say that to john connor

XD

PLEASE

25th december 2019

THE EARTH

crap its agent smith

did you watch the video where someone trapped a selfdriving car?

no

i just want a tesla

yeah its all over the place

why not a ferrari

https://vimeo.com/208642358

gas cars

plus tesla is more advanced

and is considered the worlds safest car and has much evidence to prove

it shows basically that there is still a way to go

but lets hope the ai wont ever work like feard and just like an extension for us like a calculator but instead we input some goal and it outputs whatever like a houseplan, a programm etc

you seem to imply that if AI was programmed with some goals like "human well being" etc it might kill us off
im saying thats bit impossible, @negatic

ferraris are faster and why would u need an ai to drive your car

i wanted to imply that if it WAS NOT then maybe

thats impossible too

why?

an ai specifically trained to perform music for example, will not automatically for whatever reason start to kill off humans, it was never programmed with the goal of "human well being" but that doesnt mean its going to do the opposite for the luls

i agreee with your example

but we are not at strong ai yet

u probably heard the paperclip example

or the one solving some math problem as goal

and killing humans off in the process

yeah ...its kinda a bad one

in those cases you simply gave the goa solve this problem or produce as many paperclips as possible

why is it bad?

the most realistic threats AI has currently is becoming a threat to a lot of jobs
that could lead to some big unemployement issues

https://youtu.be/721JQZw6Spg

agreed

old skool

anyway the examples remind me of an error in a programm causing an infite loop

because its so stupid..
so in that example afaik an AGI programmed to maximise paperclips suddenly goes *boom*

to infinity