stephen hawkin warns of artificial intelligence

[King of Fists]

As a computer programmer I lol @ doomsday prognostications of ai.

That shit ain't happening, we'll kill ourselves long before we design computers intelligent enough to overtake us.

 

[TheonGreyJoy**]

right? and how much closer are the computers at MIT to becoming self aware than a TRS-80? what? zero?

the combustion engine and a self aware robot are so fucking different.

we have had modes of transportation throughout the ages, but a few living things are self aware.

look at is this way, all the shit we program in a computer is like science or math. it can have facts and make crazy calculations,
but it could never understand things like PAIN or LOVE or trepidation or happiness. it could not understand democracy or a mid life crisis
UNTIL then, it is nothing but a glorified calculator

"50 points for Gryffindor!"

/thread concluded. goodnight you muggles.

Your post reads pretty ignorant.

You could argue we were no closer to flight until we stopped failing and actually did it.

Science is about advancement in increments and as long as we are advancing it is silly to pretend you, or anyone can put a limit on that advancement.

What they theorize is that concept such as AI would require awesome amount processing power. Amounts that we are only now within striking range of with advancements such as quantum computing which is literally in its infancy.

No one yet knows what the limits of the processing power will be, so pretend you can know the limits of what can be done with that, as yet undefined, power is silly.

 

[muerteverde]

Your post reads pretty ignorant.

You could argue we were no closer to flight until we stopped failing and actually did it.

Science is about advancement in increments and as long as we are advancing it is silly to pretend you, or anyone can put a limit on that advancement.

What they theorize is that concept such as AI would require awesome amount processing power. Amounts that we are only now within striking range of with advancements such as quantum computing which is literally in its infancy.

No one yet knows what the limits of the processing power will be, so pretend you can know the limits of what can be done with that, as yet undefined, power is silly.

Agreed. There are advancements in technology and then there are also advancements in the theory that will allow it. You could have a computer a trillion times as powerful as those of today, but if, as some here believe, it is limited to only ever executing code written for simple tasks by people, it will not become intelligent.

We are coming close to the limit of Moore's Law with our current silicon wafer technology, but just as this is far from the first computer technology, so it won't be the last. Just as printed circuits were replaced by integrated circuits, etc. (and reference transistors, vacuum tubes, mechanical computers) there will be something to replace our current tech. Quantum states? Light? Graphene wafers fabricated in three dimensions? Semi-organic cybernetic computers? I don't know, but I find it hard to believe that the current generation of Intel chip is the best we can expect for the next 2,000 years.

Combine these anticipated advances with the advances in machine learning and theory, and then it gets interesting.

 

[Matt4786]

I think it's only natural that artificial humans eventually replace us.

 

[countswagula]

In the long run we're all dead. I say develop away, and I for one welcome our new AI overlords

 

[Crocop42]

This is our future...

and this is what this generations John Connor will look like...

all he's going to do is complain about how is mom was never there for him through his blogging, while we face being enslaved by machines and ushered to our death.

 

[Kirin]

Apparently those people working on improving a.i. never saw any of the terminator movies

 

[dirtypablo]

You're talking about self-modifying code. That's a far cry from sentience. In order for the program to become self-aware, it would have to write code too advanced for humans to write. Yet, it's still operating within the parameters that its human creators delineated.

No code that the program can write is outside of human imagination, because the program itself is not imaginative unless programmed to be so. Hence, it can only modify its own code in the same way that its programmers are already capable of doing, it can just do it better and faster.

Sentience cannot logically arise out of that. You would be sitting there hoping for a spark of consciousness to miraculously coagulate out of nothing.

And let's say it did. How would it interact with the physical world? It doesn't know how to navigate the Internet, it doesn't know how to exist inside of a robot chassis, it doesn't know how to read human languages, it doesn't know what a human is - unless it is programmed to know. For it to learn on its own, it would have to write its own code that allows it to learn.

And you're telling me that's not just possible, not just likely, but a foregone conclusion. No.

Sentience cannot logically arise out of what? A set of instructions? How did sentience arise in humans?

 

[FierceRedBelt]

As a computer programmer I lol @ doomsday prognostications of ai.

That shit ain't happening, we'll kill ourselves long before we design computers intelligent enough to overtake us.

I think the general idea being conceived is that we'd only need to plant the seed: a program that's able to create a better program. Surely wouldn't be impressive for billions upon billions of generations but eventually you'll have code that at least rivals human insight assuming a dynamic enough task.

I know someone who wrote a chess program that used backtracking to decide the best case scenario for each move. Just basic stuff -- if I move am I in risk of losing a piece within the next 10 moves? Am I able to advance my position? Am I able to finish the game by this move ect.? (No advanced strategies established in literature)

Just take that program, run it one million times against an established chess sim, or itself, and record all your data. Establish correlations/patterns in behavior and create new conditions to avoid them. Imagine how many times you could repeat this process in a month without human interaction? Then you approach very specific conditions that would not be considered in normal operation.

Computers doing fairly specific dynamic tasks that used to be the domain of engineers and computer scientists much better than we can is within sight.

 

[Reyesnuthugr]

practically everything you said is factually wrong. do you even know what gives a computer 'computational ability' and 'memory'? and you don't 'teach' computers.

Nah I don't really care. You still know exactly what I mean.



As for "teaching," you would have to teach other programmers how to do it (assuming there's gonna be more than one guy doing all this), whether or not you semantically approve of the word "programming" as fitting under "teaching" doesn't really matter. This is a fun thread, though

 

[Reyesnuthugr]

Ya know what fellas? This reminds me.


There was a time, not too long ago, (25 years ago) when folks (music professionals, teachers, industry pioneers) were all saying that synthesizer sounds were going to replace real instruments. The sounds would sound exactly alike or perhaps even better. It wasn't even a controversial statement. At the rate they were going, it sure was true!* But then-- it wasn't at all.
And now, it has never been less true


And those are just synth sounds. Think about that


-----

*(listen to the horn intro in "The Final Countdown" for one of the examples to see what I mean)

 

[GhostZ06]

 

[WARRIOR POET]

I think the general idea being conceived is that we'd only need to plant the seed: a program that's able to create a better program.

again, computers cannot "create" anything. you have to have an intelligent and curious mind to create.

what makes a program better? less lines of code? less errors? this is hardly what we call AI as far as sci-fi goes

 

[WARRIOR POET]

Your post reads pretty ignorant.

You could argue we were no closer to flight until we stopped failing and actually did it.

Science is about advancement in increments and as long as we are advancing it is silly to pretend you, or anyone can put a limit on that advancement.

What they theorize is that concept such as AI would require awesome amount processing power. Amounts that we are only now within striking range of with advancements such as quantum computing which is literally in its infancy.

No one yet knows what the limits of the processing power will be, so pretend you can know the limits of what can be done with that, as yet undefined, power is silly.

then fucking prove me ignorant.... oh wait, you can't because your whole position is based out of ignorance

you see AI on some movie and think, oh well a movie can make it real it must be real

again, how much more SELF AWARE is the MOST POWERFUL COMPUTER on earth right now? how much more is it than a trs-80 i had in 1984????

ZERO

sorry i am not going to be ignorant and believe your shit argument. if you want to project your psychosis on other people become a transgender.

again, with making things that fly, we started off with probably a kite and even old renaissance dudes had plans for a crude helicopter

you can use physics/science to see what shape and power/thrust and etc is needed to life an object in the air

WHAT are the specifications to make a real AI? I call AI something that can be self aware and genuinely create and understand that it exists???

go ahead, get some ideas out there? oh wait, you have none once again but yap how i am ignorant. lol.

even making a ship that can fly at the speed of light would be easier than making true AI as we can use physics to see what options we had

we will make a black hole event horizon style to fold space/time before we make true AI

now, you done got served and apparently you can't dance back

/sonned

 

[FierceRedBelt]

again, computers cannot "create" anything. you have to have an intelligent and curious mind to create.

what makes a program better? less lines of code? less errors? this is hardly what we call AI as far as sci-fi goes

Less not argue terminology.

A program that can author a new program that's capable of doing things the previous version was incapable of. Whether that's faster run time, new functions, more effective algorithms it really doesn't matter.

A program that realizes it doesn't need to attempt to compute (2^32 of) an infinite sum because it's equal to something as simple as:

is a hugely better program. Designing a new function around collected data creates a new and more effective program.

I'm not an expert but start here:
http://en.wikipedia.org/wiki/Machine_learning

 

[WARRIOR POET]

Less not argue terminology.

A program that can author a new program that's capable of doing things the previous version was incapable of. Whether that's faster run time, new functions, more effective algorithms it really doesn't matter.

A program that realizes it doesn't need to attempt to compute (2^32 of) an infinite sum because it's equal to something as simple as:

is a hugely better program. Designing a new function around collected data creates a new and more effective program.

I'm not an expert but start here:
http://en.wikipedia.org/wiki/Machine_learning

again, if you want to call a program that makes another program more efficient as being the apex of Artificial Intelligence then that is just shit.

i mean, you have a program that does something. another program can compute all the possible variations it has been programmed to recognize until is CHANCES on making the program more efficient.

this is a long way from being self aware or truly creating

sorry i don't have any formulas to post to pretend that i know something. i just know. no need to show my work. off you go. thank you.

 

[FierceRedBelt]

again, if you want to call a program that makes another program more efficient as being the apex of Artificial Intelligence then that is just shit.

i mean, you have a program that does something. another program can compute all the possible variations it has been programmed to recognize until is CHANCES on making the program more efficient.

this is a long way from being self aware or truly creating

sorry i don't have any formulas to post to pretend that i know something. i just know. no need to show my work. off you go. thank you.

lol

Did you even bother to read my initial post?

I think the general idea being conceived is that we'd only need to plant the seed: a program that's able to create a better program. Surely wouldn't be impressive for billions upon billions of generations but eventually you'll have code that at least rivals human insight assuming a dynamic enough task.

I know someone who wrote a chess program that used backtracking to decide the best case scenario for each move. Just basic stuff -- if I move am I in risk of losing a piece within the next 10 moves? Am I able to advance my position? Am I able to finish the game by this move ect.? (No advanced strategies established in literature)

Just take that program, run it one million times against an established chess sim, or itself, and record all your data. Establish correlations/patterns in behavior and create new conditions to avoid them. Imagine how many times you could repeat this process in a month without human interaction? Then you approach very specific conditions that would not be considered in normal operation.

Computers doing fairly specific dynamic tasks that used to be the domain of engineers and computer scientists much better than we can is within sight.

Not that you care; you're just here to argue aimlessly, right?

 

[MortalWombat]

again, if you want to call a program that makes another program more efficient as being the apex of Artificial Intelligence then that is just shit.

i mean, you have a program that does something. another program can compute all the possible variations it has been programmed to recognize until is CHANCES on making the program more efficient.

this is a long way from being self aware or truly creating

sorry i don't have any formulas to post to pretend that i know something. i just know. no need to show my work. off you go. thank you.

I don't understand. You lament lack of interesting topics in the Mayberry. Was that you? Here's a topic and some's talking to you about it, and all you do is close html tags that were never opened?

 

[christhakingpin]

We should all feel privileged that warrior poet is here to enlighten us.

 

[MortalWombat]

Better WP than AI gimmicks with homicidal intentions.