Will Computers Outsmart Humans In 40 years?

According to the author of this site, it will take 40 years for computers to outsmart humans. The author attributes the distant estimations to exponential thinking. His personal theory states that since the rate of technology is advancing from thousands of years to short bursts in decades, that computers will become more intelligent then humans in the near future. I think it's a good rough estimate because we are obviously dawning on new technological eras. Even people who are not in the field of technology realize that computers will soon become more powerful - as they do every year.

bit.csc.lsu.edu...

Similarly, you can see the same exponential growth in virtually everything that has to do with technology: transportation (from legs, to horses, to carriages, to sail ships, to steam trains, to steam cars, to gas cars, to airplanes, to helicopters, to spaceships, to "smart" cars - as you see every step taking less and less time), dentistry (from knocking out that aching tooth, to drilling, to anesthetics, to prosthetics, to preventive dental hygienists, to X-rays, to ultrasound whitening), archeology (from "hey, here's an old cup" to satellite site detection - I skipped the intermediate steps for brevity), clothing (from animal furs to "smart" clothes that can transmit a person's vital signals such as heart rate and temperature), window-making (from a hole on the wall to high-tech windows featuring glass with power-adjusted transparencies), etc.

What are other opinions on this guys estimation? Is there some serious flaw in the logics of his arguement? Any input is welcome i'd love to hear other people's thoughts on this.


[edit on 30-10-2005 by Linux]

I would say in 20-40 years there a good chance computers will be able to process as much information as the Human brain does. Now if they will be able to outsmart humans thats a whole other story. The ability to do trillions of caculations a second doesn't mean that a machine will be Self aware or able to learn. If you cant learn from your mistakes for example all the brain power in the world dont amount to much.

In theory when or if computers rival humans in levels of brain power there is nothing stopping them from advancing far further perhaps millions of times more powerful then any human brain. Add that with a machine that could learn and is maybe self aware and things could get scary.

My opinion is that a computer can only be as smart as the smartest human. Or the person who programs it. That being said, I must concur that unless AI produces instinct or emotion, it will never best a human.

Originally posted by Rouschkateer
My opinion is that a computer can only be as smart as the smartest human. Or the person who programs it.

You have to remember that it wont just be one person programing it . Look at Deep Blue none of its programers by themselves could have come close to beating Kasparov in a game of chess.

They were able to combine the best chess tactics of world class players the world over including Kasparov and poured into Deep blues programing was able to win.

Its kind of an arbitary theory imo, I mean can you even quantify the processing power of each of us? Do we become sentient because we have reached a capacity to form a conscious in our brain or is it really a gift from our universal creator? Perhaps it is as simple as the instructions you give it, in which case all our computers need to become sentient is a better program.

Binary computers will never give rise to a self-aware artificial entity. We need a whole new class of computer technology that mimicks the human brain with artificial materials.

As for them "Outsmarting" us I highly doubt it. We are getting close to the technology needed to start modifying our own brains so by the time AI does become possible we will still most likely be far far smarter then they are. I'd say that in 40 years that the measuring stick for AI will go like this.

Cat

Dog

Chimp

Pre-Modified Human

Transhuman

Posthuman

IMO We will be somewhere between Transhuman and Posthuman when AI first gets to the level of a 20th century human brain.

Originally posted by picard_is_actually_a_grey
Its kind of an arbitary theory imo, I mean can you even quantify the processing power of each of us?

We can only estimate the processing power of the average human brain as of yet.

This is based on factoring the capability of the brain's 100 billion neurons, each with over 1,000 connections to other neurons, with each connection capable of performing about 200 calculations per second. This puts the human brain's probable processing power at around 100 teraflops, roughly 100 trillion calculations per second

www.wired.com...

[edit on 30-10-2005 by ShadowXIX]

Well i can allreasy say the computer i am useing is smarter then me i dont kwno all the codes and stuff it takes to run this forum or the internet so it out smarted me there

40 years? I know quite a few people who are "outsmarted" by Pong consoles...

Originally posted by ShadowXIX
You have to remember that it wont just be one person programing it . Look at Deep Blue none of its programers by themselves could have come close to beating Kasparov in a game of chess.

As a programmer, I might add something here. Beating Kasparov in a game of chess is a fairly simple problem. Look at the current state of the board. If you move this piece, what possible pieces can he move? Go through all of the possible options until you find a winning combination. The problem lies in doing it efficiently and quickly. It's easier for a computer to do it than a human, because the comp can process many more scenarios in a second and can store and refer to many more reference scenarios in a much quicker manner, but it's still rather slow.

Then again, I don't know what they were trying to do with the AI for Deep Blue, and I doubt they were going for such a simple approach as what I stated above.

Originally posted by ShadowXIX

We can only estimate the processing power of the average human brain as of yet.

This is based on factoring the capability of the brain's 100 billion neurons, each with over 1,000 connections to other neurons, with each connection capable of performing about 200 calculations per second. This puts the human brain's probable processing power at around 100 teraflops, roughly 100 trillion calculations per second

www.wired.com...

Well if a Human is only working with around 100 teraflops then we have been beaten.

Blue Gene surpasses its previous record

Oct 31, 2005

IBM's Blue Gene/L supercomputer broke its own record, to be judged the world's fastest computer. The Blue Gene/L did a whopping 280.6 teraflops.

Physically yes Blue Gene is more powerfull but pound for pound we still outperform it.

Software is the last major hurdle as it seems we will have the hardware capability within 20-30 years.(We need different hardware for Quantum Computing)

[edit on 31-10-2005 by sardion2000]

Numbers wise computers have us beaten already, but AI basically will be limited to the intelligence of the developers, which it has been my experience is a hit-or-miss thing. Being outsmarted isint that big a deal really; there will always be someone smarter than you, as well as someone who isint as smart as you. The thing that I think would send up flags would be if computers step past calculations and logical things to having anything resembling emotions. Being smart isint the pinnacle of humanity I would say, sure it may make for some impressive feats, but theres alot more that I would classify as more important.

Originally posted by MCory1


As a programmer, I might add something here. Beating Kasparov in a game of chess is a fairly simple problem. Look at the current state of the board. If you move this piece, what possible pieces can he move?

I dont know if I could call that a simple problem since there is a heck of alot of possiblities in a Chess Game. The number of possible games of chess as 10^(10^(50)). The number of possible board positions is atleast 10^120. Even so just knowing all possible moves Kasparov could make means very little without any form of strategy.

mathworld.wolfram.com...


But I only used Deepblue as a example showing how that a Computer can be made by a whole bunch of programers that alone would have no chance of beating Kasparov in a game of chess.

In response to Rouschkateer saying"computer can only be as smart as the smartest human"

Originally posted by lost_shaman

Oct 31, 2005

IBM's Blue Gene/L supercomputer broke its own record, to be judged the world's fastest computer. The Blue Gene/L did a whopping 280.6 teraflops.

WOW good find thats pretty impressive. Last time I checked the best super computers were doing around 30 something teraflops.




Look at the size of the thing though, It takes all that to out do 3 pounds of human brain.

Sardion is right on. Machines can't outsmart us because they can't think. They can learn, they can imitate it to a near-perfect impression, but there's a certain something they'll always be missing, even if we never notice it. AI could even get to the point to be legally considered alive, but that's another point. What we need to start looking for is modifications. Cyborg's, enhancements. We can utilize their strengths and combine them with our weaknesses for more powerful (quasi-)human existance.

I am very surprised this thread has gotten this far and no one has mention Ray Kurzweil and his book "The Age of spiritual machines". Definately an interesting read for those interested in the subject.

As far as binary computers not being superior to the human brain, thats correct..

but scientists have been working with things called Neural Nets...

this site has some basic information on them
www.faqs.org...

also take a look at Moore's law and the law of accelerating returns.

Originally posted by Amorymeltzer

What we need to start looking for is modifications. Cyborg's, enhancements. We can utilize their strengths and combine them with our weaknesses for more powerful (quasi-)human existance.

I agree with that people will integrate the technology to improve ourselves not only our computers.

We are not very far from the day now when amputee's with prosthetics will be out preforming people who have normal limbs.

Originally posted by ShadowXIX
We can only estimate the processing power of the average human brain as of yet.

This is based on factoring the capability of the brain's 100 billion neurons, each with over 1,000 connections to other neurons, with each connection capable of performing about 200 calculations per second. This puts the human brain's probable processing power at around 100 teraflops, roughly 100 trillion calculations per second

Er I think your math is off... 100 billion x 1,000 x 200 = 20,000,000 billion calculations per sec or 20 petaflops not 100 teraflops. SO... there wont be any computer to outdo the human brain FLOPS wise for 20-40 years... But remember computers are good at just crunching number, just deciding to pick up a pen is more complex for your brain then for a computer to find 379 to the 51th power. (Just an example)

[edit on 31-10-2005 by beyondSciFi]

"Computers are only as smart as humans make them."

the human brain is incredibly adept at paralleled calculations, but only mediocre at extensive sequential calculations.

a computer has the edge in sequential calculations, but in the realm of paralled calculations, a human brain is much more powerful.

and yes 20 million billion calculations is correct...

100 billion neurons, estimated 1000 connections between each neuron and its neighbors, then you get to 100 trillion connections, with each capable of simultaneous calculation, and each can make about 200 calculations each second, which compared to a pc, is very slow.. it is the fact that we have so many connections that we have more calculation power.

the above information was taken from the book "The Age Of Spiritual Machines" by Ray Kurzweil.