[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math

View post   
File: 2.10 MB, 640x360, 1IJkOy2.webm [View same] [iqdb] [saucenao] [google]
10797433 No.10797433 [Reply] [Original]

should I just make ai that solves the world's problem? seems easier than hiring a bunch of humans to do it, delegate them to certain tasks, wait 20 years for them to git gud at it, then they might fuck it up anyways because they are tired or smthing. will someone make the ai before me

>> No.10797491

How do you train such an AI?

>> No.10797507

>>10797491
you couldn't? Just emulate the human brain (Human Brain Project) and then expand it. Basically takes about 15-20 years of research. Then if you can simulate the human cortex, which is responsible for the higher level reasoning, then just make it bigger. Add that "extra layer" that would make the brain smarter, and continue ad infinitum. No need to "train it", like you do with a machine learning model, necessarily.

>> No.10797521

Nice role-play you enormous retard.

>> No.10797533

>>10797507
So you expect to get human brain in a blank slate state, give it the most pressing problems of mankind, and get an answer? You still need to teach it because otherwise you’ve got a great pattern recognition machine that knows no patterns to recognize

>> No.10797534
File: 7 KB, 632x182, 1544438372711.png [View same] [iqdb] [saucenao] [google]
10797534

https://www.youtube.com/watch?v=LPpgxBkDYfY

And a huge dick.

>>>/b/804345388

>> No.10797537

Making an AI that can solve the world's problems is equivalent to solving the world's problems.

>> No.10797539
File: 222 KB, 600x600, 1555202721736.jpg [View same] [iqdb] [saucenao] [google]
10797539

>>10797537
Or just going the Full Disney!

>> No.10797557

>>10797433
Isn't that essentially what the gary optmimum guy was trying to do ?

>> No.10797558
File: 84 KB, 700x967, 1552595674985.jpg [View same] [iqdb] [saucenao] [google]
10797558

>>10797433
Can't save the world today, had to save (You).

>> No.10797591

The reason the world can't be saved is because of us. Our worlds problems are actually quite simple and easy to solve. It's us that makes the problems more difficult and overly complicated. We didn't over complicate things on accident, it's by design. The world is set up to benefit a small few, and if it's complicated and complex, then it's difficult to undo all that. If it's confusing and obfuscated, then there's no way know where to point the finger of blame.

It doesn't matter how smart or advanced we become, because there will always be people who used their intelligence and our technological advancements to hurt others for their own gain.

If you make an AI to solve the worlds problems, someone will steal or copy it, and reprogram it to solve only THEIR problems at the expense of all others. This is how humans are. This is how we are. This is why we can't have nice things. You can't win.

>> No.10797644

>>10797591
what people don't realize is AI tend to learn humanities bad habits. All the logical faults that humans have can also be applied directly to a sufficiently advanced enough AI. In short, if we make and AI it'll quickly learn to be a selfish prick, but it'll also be intelligent and resourceful enough to do so without us humans knowing what it's doing or whats going on.
And that's how humans become slaves to the machines, except the machine will be clever enough to convince us it's all for the best. It'll establish an abusive relationship and we'll thank it for it.

>> No.10798274

>>10797644
god damn... but if the AI is truly smarter than us, it would be able to realize the same things you do. And if it's infinitely smarter than us, then it may show mercy