HP Forums

Full Version: evolution of computer power/cost
You're currently viewing a stripped down version of our content. View the full version with proper formatting.

i just thought someone here might find this interesting.
http://www.dai.ed.ac.uk/homes/cam/images/MoravecsLaw.jpg
i wonder where the 41 and 48 fit in here.

Stopping at 200MHz (PII) (or PPC601 180MHz). Extrapolating 15 years of exponential growth (or is it even faster?) is quite hard - and then there's the "target" of 100 Million MIPS, is it equivalent to a human brain? Or do we need 10 times/100 times/a million times as many MIPS? I'd say nobody knows :)

Brain is all about parallel processing and automatic pattern matching. I wouldn't try to measure that in MIPS.

Very true. That is why chess engines tend to rely on the brute force approach with ever faster computers until they overwhelm even the best grand masters. Still, AI remains an elusive goal and Lieutenant Commander Data's positronic brain is just as much fanciful fiction as a mass transporter.

hmm, some people beg to differ. I recently heard Ray Kurzweil at a conference in New York and then went on to read his books. He has very convincing arguments that we will achieve human brain capacity calculations in our life-time (and far beyond).

He estimates the processing power of the human brain to be between 10^14 and 10^16 MIPS (100 bn neurons x 1000 synapes per neuron x 200 calculations per second)

Quote:
The human brain has about 100 billion neurons. With an estimated average of one thousand connections between each neuron and its neighbors, we have about 100 trillion connections, each capable of a simultaneous calculation. That's rather massive parallel processing, and one key to the strength of human thinking. A profound weakness, however, is the excruciatingly slow speed of neural circuitry, only 200 calculations per second. For problems that benefit from massive parallelism, such as neural-net-based pattern recognition, the human brain does a great job. For problems that require extensive sequential thinking, the human brain is only mediocre.

With 100 trillion connections, each computing at 200 calculations per second, we get 20 million billion calculations per second. This is a conservatively high estimate; other estimates are lower by one to three orders of magnitude. So when will we see the computing speed of the human brain in your personal computer?


(the whole article can be found here)
Another related article about this can be found on his very informative web-page The Law of Accelerating Returns
His book, 'The Singularity is Near' is definitely worth a read and I would guess the crowed here would generally enjoy it. A Wiki-summary can be found here

Last but not least, I found the slides he showed during his presentation. They are worth a quick peak. I find they quite convincingly show the smoothness and predictability of exponential growth. Highly recommend to take a peek...Graphs & Slides

Cheers

Peter

Looking at slide #81 in the Powerpoint presentation, we're only a year away from the disappearance of computers, when we'll all be wired directly to the internet. Pretty scary, really, but the ultimate dream for any totalitarian state.

Edited: 2 Apr 2009, 1:44 p.m.

I have met Ray and read his "The Singularity is Near' book. All good stuff and well worth a perusal.

Frank Tipler, a professor of mathematical physics at Tulane University wrote "The Physics of Immortality" in the early nineties. Also a good read. In it, he estimates the storage capacity of the brain at 10^15 bits and the processing speed at 10 teraflops. He predicted that we will have machines capable of human-level processing capability (power and storage capacity) by 2030.

Time will tell...

Jeff Kearns

http://i.gizmodo.com/5196845/robot-makes-autonomous-scientific-discovery-for-first-time

A timely article. Granted it is not a full human replacement, but computer-aided-discovery is a start.

Yes, I read that one, too. Very impressive. There is actually a weekly snippet service I subscribed at kurzweilai.net which sends em once per week the most 'futuristic' actual research projects published and this one was one of it. There was another I read about Honda showcasting their Aimo robot being steered with just with thought. Granted it does only a few things, but still...

Cheers

Peter