/lit/ - Literature

>>11791688
Not having global warming / environmental pollution could also be considered more desirable than the opposite and look where we are.

All it takes is one nation/entity believing that an AGI would give it the edge over its opponents to make the creation of one desirable. Hell, you don't even need that. All you need is a team of PHDs that are starved to get another publication to a big dick conference.

Do you think the inventors of AlexNet thought that their invention could be applied to make autonomous kill drones? Maybe, but that publication was more desirable. It will be the same with AGIs, but luckily we are still far away from those, if we'll ever have them.

>>11791705
If you want to know how ML actually works I'd recommend Andrew Ng's course on coursera or the book Introduction to Statistical Learning. This isn't some high level look on AI/ML though, but rather a very detailed look at it. You need to know statistics/math on at least college entry level.
The field of AI is pretty big and ML is only one part of it (but whats generally considered to be AI in pop culture right now). Things like natrual language processing are part of it and theres different algorithms.
Honstely I can't recommend any books that could be understood by the average person and wouldn't be overhyping AI. (AI researches love to hype their work because its big money atm and they are probably still remembering the last AI winter).

I was like you guys once, read a lot of transhumanist fiction/ cyberpunk, read less wrong, watched ted talks, interviews with kurzweil etc. Then I actually started working with this stuff and you realize it's all just crude number crunching and trial and error. There's things like formal reasoning systems which for example are used at facebook to detect spam, but again, it's in the end just pattern matching based on rules formulated in logic for a specific purpose

Maybe genetic algorithms will be the next big thing that the tech industry can hype up and get big bucks for