The Power of PCs

[ecoli]

I was curious, because I couldn't reason by myself, how come nobody could predict the power of personal computers. Science fiction authors of the 50's and 60's, especially, envisioned computers of the future to be giant monstrosities, but very intelligent. Computers are still non-sentinent and extremely small. How come they didn't accurately predict what computers would become?

[drochaid]

What form will computers take in 50 years time? WIll they be sentient?

 

No. I can't answer that accurately either.

[bascule]

I was curious, because I couldn't reason by myself, how come nobody could predict the power of personal computers.

 

Gordon Moore did...

 

Science fiction authors of the 50's and 60's, especially, envisioned computers of the future to be giant monstrosities, but very intelligent.

 

They also made other poor predictions, such as flying cars...

 

Computers are still non-sentinent and extremely small. How come they didn't accurately predict what computers would become?

 

They misunderstood the scope of the task of creating stron AI, and didn't understand how much computers could be miniturized. Plus they wanted the computers to be large and showy, not small and diminuitive.

[Pleiades]

Oh, but you’re so wrong, computers are powerful, they are just easy to ignore. Consider Google, behind that simple interface lie roughly 100,000 computers all linked together to form a super fast supercomputer. How can Google search through an index of billions of pages and return an answer almost as fast as human reflexes? I search for ‘Science Forums and Debate’ returns results in 0.27 seconds, that’s 270 milliseconds, human reflex time is 100-200 milliseconds. I know there all sorts of tricks used to do things so fast, but it still blows me away. http://en.wikipedia.org/wiki/Google_platform

 

Google isn’t usually refered to as a proper supercomputer but its performance is comparable to those in the top 10.If you look at the top 500 supercomputers list (http://www.top500.org/) you’ll see that the fastest is the DOE's IBM BlueGene/L system at 280.6 TFlop/s, that’s 280 Trillion floating point operations per second! (http://en.wikipedia.org/wiki/Blue_Gene). Although it’s comparing apples to oranges the human brain is sometimes estimated to have between 1 and 10 petaflop/s of power, (1 petaflop/s is 1000 teraflop/s)

 

These supercomputers are many smaller computers linked together, the supercomputer as a whole is massive. What we are lacking is the programming knowledge to create Artificial Intelligence on a human level (or even a mammal level).

[ecoli]

the difference is, that if those computers were un-linked, they would still work fine, individually. If you removed parts from the giant computers of old, they wouldn't work anymore... (I suppose that might be dependent on what part you removed, but you get the idea)

[bluesmudge]

no one can predict anything films are a great one for this - back to the future predicted flying cars in 2015 - ok theres 9 years to go but thats *probably* wrong . . . . Bicentennial Man predicted artificial Intelligence in 2005!

[the tree]

Oh, but you’re so wrong, computers are powerful.

Fast and high capacity yes, but not sentient. The difference is huge.

[Pleiades]

You’re right, it is a huge difference, but my point was that the hardware will soon be advanced enough, but the software necessary for intelligence is far, far behind what it needs to be. I did address that in the last paragraph.

What we are lacking is the programming knowledge to create Artificial Intelligence on a human level (or even a mammal level).