The Artificial Intelligence guru Ray Kurzweil believes that technology, accelerating exponentially, is converging on a point some 20 years ahead when in a flash of universal illumination he calls "The Singularity". It will solve all the problems of mankind. cloudillustration Adoption of cloud services is up about 60 to 61 per …
TLA is an initialism not an acronym (unless you pronounce it as one word) and is therefore not self-describing, which I thought was rather the point.
"technology, accelerating exponentially"
Oh really? It's true that Moore's law has held, but outside of counting the number of transistors on a chip, do we really see much game-changing progress in technology? 40 years ago we could land men on the Moon and cross the Atlantic by scheduled airliner in 3 hours. Today? - not so much. Key technologies such as transport, power generation and storage have seen a few percent increase in efficiency but no real novelty.
Even software hasn't improved much in the last decade or two - a few more buttons in Word and lots of pointless shiny, shiny bits to go on web pages about covers it.
I'd love Ray to be right, I really would, but I just can't see it.
Yes - accelerating exponentially
Ah, Mr Miller. You've committed the classic schoolboy error of looking at output of things which are not derived from information technology. Yes, 40 years ago we could do wonderful things. 100 years ago we could do just as many marvelous things. Tut tut, we haven't really changed that much, right?
Go back to anytime in recorded history, and you'll probably find things that haven't changed that much. Consider the height of buildings. Although we have billions times more processing power at our fingertips, our buildings aren't billions of times taller. Nor are our trains or planes billions of times faster. Nor are our lives billions of times longer.
But we're not talking about output. We're talking about information technology and its exponential growth. Anything which is inherently an information technology does grow exponentially, whether we like it or not. DNA, communication, computing power, photovoltaic cells. Once we start to treat something as an information technology, it will improve in price-performance exponentially.
Moore's law rightly described the exponential grown of the price-performance of integrated circuits. However it was the fifth exponential growth in computing power. We are currently on the cusp of the 6th phase of exponential growth, all of which has followed seamlessly as expected. (For information, the first was electromechanical calculators, cicra 1890; the second was Turing's relay-based Robinson machine; the third was vacuum tubes; the fourth was transistors; the fifth was integrated circuits; the sixth will be 3 dimensional circuits).
In terms of your examples - storage (of information technology) has in fact seen exponential growth. Consider how much it costs you to save 1 MB of data in an external hard drive now, versus 20 years ago.
In terms of transport - well, that is not an information technology. Its purely mechanical - at the moment. Although you could look at certain aspects of transport which are I.T. For example, consider how much it costs to transport 1MB of information electronically - now, versus 20 years ago.
And for power generation, well, we're still in a mechanical phase for most of our power generation. Even something like solar (concentrated solar) is just a mechanical process (heating water to create steam to turn a turbine). However, photovoltaic cells are an information technology, as we're dealing with electrons and light. Solar P.V. is following a predictable exponential growth in price-performance. But because we're still in the very early stages, it's hardly noticeable. But it's growing just the same (see http://bigthink.com/ideas/31635). The growth rate is a doubling every 2 years (as opposed to every 18 months for processing power).
You're right in a way, Ben
If you consider only *information* technology (as I rather thought my original post pointed out).
So, compared to 15 years ago, I can store 1,000 times as much on my PC and process it 1,000 times more quickly*. But (I would argue) it isn't doing anything significantly *better* than it was 15 years ago. It certainly hasn't made anyone 1,000 times (or even twice) more productive.
*Actually, I can process it about 20 times more quickly (3GHz vs 166MHz**), but in 16 multiple streams, which really doesn't help me much unless I want to do vector arithmetic. In order to stop our fancy hardware sitting twiddling its thumbs, we've reinvented Virtual Machines - good news for power consumption and floor space in data centres, but otherwise pointless.
** Or if you like you can throw in another power of 2 for the move from 32-bit to 64-bit, but it doesn't affect my argument.
Clouds and the illusive singularity
The clouds have their problems for SMEs. Problems that MNCs may struggle to work with.
Security in the clouds is but one. Disaster recovery another, then there is regulations, data protection, etc, etc... Just migrating into the IaaS model introduces risk with the migration itself. Not too mention that the migration is going to be costly in HR hours, documentation, procedure changes, version control...
They probably just cannot be bothered.
The singularity... hmmm that illusive concept. I hope Ray and university are considering that the actual Singularity event itself may FAIL spectacularly. After all version 1.0 has been known to do that in the past in many cases... indeed "blindly".
Kurzweil is a utopian fantasist
About 60 to 61% ?
Is that close to 60.5% ?
"Solve all humanity's problems"
sounds like a euphamism for "wipe us out".
Firstly, this cloud stuff is going to make it much, much easier for customers to 'tweak' their licensing costs. If you're operating MS Office, say 50 workstations and then find you can get away with Google Docs for half of them, you just tell your cloud supplier you don't want them any more. Yes, I know you'll all be quoting back to me that cloud suppliers will try to stop that with tricks like annual licensing, 'standard' packages and higher costs for lower volumes, but you're ignoring the marketplace effect if you do that. This market will become much easier to enter in the future.
Secondly, and this is really good, many organisations will not want to move their data off-shore (I even think this is a legal requirement for accounting, but IANAL). This means a huge amount of general 'off-shoreing' may revert to home territory again, with a consequent rise in employment - for call centres and cloud infrastructure.
- Boffins attempt to prove the UNIVERSE IS JUST A HOLOGRAM
- China building SUPERSONIC SUBMARINE that travels in a BUBBLE
- Review Raspberry Pi B+: PHWOAR, get a load of those pins
- Experimental hypersonic SUPERMISSILE destroyed 4 SECONDS after US launched it
- That 8TB Seagate MONSTER? It's HERE... (You'll have to squint, 'cos there are no specs)