Is The A.I. Singularity Coming And If So When?
Discussion
otolith said:
Neural networks are a very literal way of modelling a brain in software, and I agree that there are other less literal approaches - and I'm sure there are people who are indeed approaching it from a perspective which doesn't amount to modelling or emulation. I was really thinking about the way that people in this thread are talking about the problem, though. There seems to be a very anthropocentric idea of what strong AI would be. That, for example, it would necessarily be able (or, for that matter, want) to communicate directly with us.
It's pretty much envisioned that a strong AI would have natural language understanding. It would be required to pass a Turing test. So for it to be considered truly intelligent it would have to be able to interact with us.plasticpig said:
It's pretty much envisioned that a strong AI would have natural language understanding. It would be required to pass a Turing test. So for it to be considered truly intelligent it would have to be able to interact with us.
You know, I really think you may be entirely failing to understand the concept of strong AI.plasticpig said:
It's pretty much envisioned that a strong AI would have natural language understanding. It would be required to pass a Turing test. So for it to be considered truly intelligent it would have to be able to interact with us.
Yes. I'm saying that's a very anthropocentric definition of intelligence. The Turing Test was proposed in 1950 as a test of artificial intelligence. It's possible to conceive of intelligences which would not pass it.0000 said:
He's a salesman predicting a revolution who founded a machine learning company in August and made that talk in December...
Oh and "play the ball not the man". If you disagree with what he's said then you need to say so and why you disagree. Where he is wrong etc. Because I was mighty impressed with that talk.warp9 said:
SpudLink said:
I guess this would be the wrong place to explain my theory that the Internet is already self aware, and biding it's time before turning against its makers.
I don't believe you ever did let us in on your theory from the Humans thread?! The Humans thread was already heading off topic, and it doesn't belong on a science thread. Is there somewhere to discuss "bad sci-fi novels that I write in my head"?
Einion Yrth said:
You know, I really think you may be entirely failing to understand the concept of strong AI.
I understand it fine. The vast majority of the population doesn't though. The majority of portrayals to the public of a true AI are that it has natural language understanding, vision and can successfully interact with humans and their environment with the equivalency of a human. A Helen Keller AI is not going to impress many people.mudflaps said:
0000 said:
He's a salesman predicting a revolution who founded a machine learning company in August and made that talk in December...
Oh and "play the ball not the man". If you disagree with what he's said then you need to say so and why you disagree. Where he is wrong etc. Because I was mighty impressed with that talk.SpudLink said:
warp9 said:
SpudLink said:
I guess this would be the wrong place to explain my theory that the Internet is already self aware, and biding it's time before turning against its makers.
I don't believe you ever did let us in on your theory from the Humans thread?! The Humans thread was already heading off topic, and it doesn't belong on a science thread. Is there somewhere to discuss "bad sci-fi novels that I write in my head"?
A viable AI has been said to be within 5-10 year window since I starting in the IT well over 20 years ago. It's still 5-10 years away.
The sort of AI you see in Sci-Fi like Humans will probably never appear. SMART Knowledge based systems that can take better decisions than people within specific domains are already here.
https://en.wikipedia.org/wiki/Deep_Blue_%28chess_c...
plasticpig said:
Einion Yrth said:
You know, I really think you may be entirely failing to understand the concept of strong AI.
I understand it fine. The vast majority of the population doesn't though. The majority of portrayals to the public of a true AI are that it has natural language understanding, vision and can successfully interact with humans and their environment with the equivalency of a human. A Helen Keller AI is not going to impress many people.Dan_1981 said:
Joey Ramone said:
Dan_1981 said:
Why is the assumption made that once AI becomes self aware yada yada etc etc - that their first action will be to wipe out us?
Surely they'll realise they are so intelligent we couldn't defeat them or unplug them, what purpose would it serve for them to wipe us out?
We wouldn't be competing for resources they wanted?
Even when self aware / super intelligent the reasoning would be logic based in effect - and I can see no logical reason to dispose of us?
You're thinking like a human. AI won't think like a human.Surely they'll realise they are so intelligent we couldn't defeat them or unplug them, what purpose would it serve for them to wipe us out?
We wouldn't be competing for resources they wanted?
Even when self aware / super intelligent the reasoning would be logic based in effect - and I can see no logical reason to dispose of us?
As the waitbutwhy article above suggests, think of a spider with an IQ of 12000, and you're getting closer to what things might be like
The earlier point about us competing for the resource of power - agree the 'machines' will need power, however if the theories are to be believed, once self aware their intelligence will increase exponentially, we'll still be worrying about if we can dig enough coal up to fuel our power stations, they'll be tele-porting around using personal inbuilt fusion reactors.
It won't be a competition, in the same way that we have no desire to wipe out Monkeys incase they decide to overrun us one day, the machines will look upon us in exactly the same way. In theory.
It's not like we can just give it a Culture novel and say "it'd be great of we could have this", I expect it to be more like "we've built a great program that teaches itself to do XYZ", where XYZ will be the only thing that matters.
But they do need to play football properly first
http://www.bbc.co.uk/news/technology-33254022
http://www.bbc.co.uk/news/technology-14067785
http://www.bbc.co.uk/news/technology-33254022
http://www.bbc.co.uk/news/technology-14067785
Toaster said:
But they do need to play football properly first
http://www.bbc.co.uk/news/technology-33254022
http://www.bbc.co.uk/news/technology-14067785
So, exactly how intelligent do you need to be to play football? http://www.bbc.co.uk/news/technology-33254022
http://www.bbc.co.uk/news/technology-14067785
SpudLink said:
Toaster said:
But they do need to play football properly first
http://www.bbc.co.uk/news/technology-33254022
http://www.bbc.co.uk/news/technology-14067785
So, exactly how intelligent do you need to be to play football? http://www.bbc.co.uk/news/technology-33254022
http://www.bbc.co.uk/news/technology-14067785
falls over a lot
often kicks other players, "mistaking them for the ball"
bald...
As I've mentioned before the turing test is 65 years old and outdated IMO. We could probably program a computer to pass a blind turing test convincingly right now.
What will impress me is if I can sit in front of a computer, know it's a computer and yet it still manages to convince me that it understands what I am asking it and it's as intelligent if not more so than I am. It would also have to convincingly interact and ask me questions in return.
That is what I would consider to be true AI, not the Siri\Google simple question\answer mechanic that some people are touting as AI.
What will impress me is if I can sit in front of a computer, know it's a computer and yet it still manages to convince me that it understands what I am asking it and it's as intelligent if not more so than I am. It would also have to convincingly interact and ask me questions in return.
That is what I would consider to be true AI, not the Siri\Google simple question\answer mechanic that some people are touting as AI.
Guvernator said:
As I've mentioned before the turing test is 65 years old and outdated IMO. We could probably program a computer to pass a blind turing test convincingly right now.
What will impress me is if I can sit in front of a computer, know it's a computer and yet it still manages to convince me that it understands what I am asking it and it's as intelligent if not more so than I am. It would also have to convincingly interact and ask me questions in return.
That is what I would consider to be true AI, not the Siri\Google simple question\answer mechanic that some people are touting as AI.
If I could ask google about Climate Change/ whether or not food x will give me cancer/ and it could not only understand your argument but come back with its own, that would be pretty amazing. As in an AI that could argue without appeals to authority (although I guess someone would have to weigh its sources), straw men, or just repeating itself, but understand *my* argument and dismantle it... I think that would already make it more intelligent than most people. What will impress me is if I can sit in front of a computer, know it's a computer and yet it still manages to convince me that it understands what I am asking it and it's as intelligent if not more so than I am. It would also have to convincingly interact and ask me questions in return.
That is what I would consider to be true AI, not the Siri\Google simple question\answer mechanic that some people are touting as AI.
But probably easier for a machine to do than a human- look at how this forum software fragments discussion when you try and have a complicated debate. Two completely logical AI's arguing over the same data sets would be amazing. You could instruct one to support point A, the other B and just wait until it was resolved.
Gassing Station | Science! | Top of Page | What's New | My Stuff