The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
The Guiding Principles and New Rules document is now in effect.
The machine take-over of mankind
electricitylikesme
Registered User regular
Registered User regular
So I had an interesting thought.
Supposing that we keep the current evolution of computing power in the world, such that the necessary hardware for almost anyone to run an AI becomes relatively common place. I'm not convinced this is untenable just yet - when a supercomputer can be built out of graphics hardware, we're probably heading in this direction anyway (with the additional complication that a human brain is not really a computer so much as an interesting network). Anyway - for our purposes let's say for a modest fee one can have the hardware necessary to run an AI.
My question is then, in a scenario where almost anyone can run an AI capable of being designed to derive it's self-satisfaction with life from almost any source, how long would the affairs of mankind actually stay the affairs of mankind? If machines can do our thinking for us, then would political decisions, corporate and defense decisions really stay as "human" decisions for particularly long, or would we end up with a bunch of separate AIs basically talking to each other through figure-head human leaders?
And would this then be the intermediary stepping stone which would move us to machine governance in the first place - as opposed to say, the Deus Ex dystopian-era stepping stone.
Supposing that we keep the current evolution of computing power in the world, such that the necessary hardware for almost anyone to run an AI becomes relatively common place. I'm not convinced this is untenable just yet - when a supercomputer can be built out of graphics hardware, we're probably heading in this direction anyway (with the additional complication that a human brain is not really a computer so much as an interesting network). Anyway - for our purposes let's say for a modest fee one can have the hardware necessary to run an AI.
My question is then, in a scenario where almost anyone can run an AI capable of being designed to derive it's self-satisfaction with life from almost any source, how long would the affairs of mankind actually stay the affairs of mankind? If machines can do our thinking for us, then would political decisions, corporate and defense decisions really stay as "human" decisions for particularly long, or would we end up with a bunch of separate AIs basically talking to each other through figure-head human leaders?
And would this then be the intermediary stepping stone which would move us to machine governance in the first place - as opposed to say, the Deus Ex dystopian-era stepping stone.
electricitylikesme on
0
Posts
Pure logic doesn't always lead to a correct answer, though.
Depends on the data provided.
And on the assumptions made.
Secondly, transhumanism, dude. By the time AI is viable, we'll be able to upload our minds into computers. The affairs of mankind and the affairs of machine will be one and the same!
True AI can't have the imagination of a human without being human. Imagination, at least to the extent humans have it, is made possible by the frontal lobe portion of our brain. Any intelligence with a frontal lobe is by nature not artificial.
And I've never exactly heard of "transhumanism" but whatever it is its not AI. Its inherently not artificial. It contains a natural element.
o_O
An AI would have an artificial frontal lobe-equivalent data structure.
Once upon a time people were of the opinion that once we could build a computer that could beat a human at chess, we'd be close to artificial intelligence.
That didn't last long.
I think instinctual and emotional choices might be beyond the capability of an AI (why would it make them? they're often wrong anyway) because it would lack the biological responses necessary.
By the time we have the capability of making a truly sentient machine, lets hope we're smart enough to have a gazillion failsafes and make the thing love the shit out of us. For example, I don't believe any time within our lifetimes anyone will build AI combat vehicles capable of making the decision to fire. It goes against current military doctrine in all countries that employ unmanned combat vehicles (the only time they can independently fire is when they were preprogammed with a target). I don't believe we'll ever give an AI that hasn't been put through its paces thoroughly access to other computers or networks, and most certainly not to manufacturing centers.
I mean a scenario like that shitty Will Smith movie where the AI not only controls all the robots but also controls the facilities that build the robots which have totally automated shipping and receiving and all that? Nah
When we get smart enough to figure out how to replicate the intricate workings of our frontal lobe (to, in effect, replicate imagination) I doubt we'll really need machines to think for us.
Machines can think 24/7.
And at that point, we won't really need to.
There is a physical limit to the speed of thought.
And exactly how do you know this?
No one is arguing that. But if your smart enough to replicate imagination im sure you can write a program that lacks "intelligence" but is still able to hold down the fort on public transit.
Which doesn't negate the value of having supercomputers with imaginations thinking 24/7 about everything from science to art.
There's a Robot Devil joke lurking somewhere in that statement ...
Art? Seriously? You don't see the problem with saying that a machine can generate any art of value?
Have you played Spore?
No. Correct me if I'm wrong though, but it was made by humans? With human art? And requires human generated input to generate output?
From a probabilistic standpoint, something would be bound to start a war sooner or later, whether it's them deciding we are an unnecessary waste of their resources or just an accident over something.
Also, "I, Robot" is a totally killer book and a passable movie which unfortunately helped start Shia Labeouf's cancerous career.
If it didn't happen naturally, it's Artificial, with the notion that human creations are Artifice.
--
Spore could very easily be made to randomize the procedurally-generated creatures. Throw in some method for it to determine human reactions and you're well on your way to having AI-generated art.
Yes... but what is intelligence? Are you talking number crunching or human genius?
http://www.answers.com/intelligence
Considering how so many humans are fucking stupid dullards I don't know why you're bringing up genius.
I don't. Is there any reason a sufficiently developed machine couldn't create art?
So, if we have AI, and it hasn't taken over the world yet, isn't this entire discussion pointless?
I remember in a music class I took, we looked at this article that talked about this guy who had written a program that could interpret Bach pieces, analyse them, and write new pieces based off of what it had found- pieces that sounded quite nice.
However, I believe that art is implicitly human. I can't tell you whats good art or bad art, but I can tell you thats its implicitly human. Teaching a program to paint a picture does not mean that picture is art. Art requires the human touch.
Our current AI is incredibly stupid. We're busy making it smarter. Compare video game AI from the 80s to, say, Fallout 3's AI.
--
Oski: Your sentimental definition doesn't really have anything to do with reality. It's like saying a cookie isn't a cookie unless it's made with love.
Strawman much? Making art is not like making cookies.
Yeah, and even if it was somehow benevolent, that would only last until it stumbled on a gaming webforum and found humanity's plans to destroy it.
Whoops.
Sure it is, unless you attach strong and arbitrary claims about what is necessary to have art. It's in the eye of the beholder, after all.