• Google
    This Blog Web

October 2011

Sun Mon Tue Wed Thu Fri Sat
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          

RSS Feed

Bookmark and Share

Email Feed

  • Powered by FeedBlitz

« Nanotech and Cyborgs | Main | Nanotechnology Law »

March 05, 2006


Feed You can follow this conversation by subscribing to the comment feed for this post.


Ian also thinks computers will outperform humans at just about any task by 2015, if I remember correctly.

And yet he insists there will be no hard-take-off Singularity.

I have posted a discussion between me and Ian on my blog about this.

It's at:


Tinyurl for your convenience:


Brian Wang

I have made an article with a set of predictions which will will be in the March, 2006 monthly nanotech now report.

The background references for those predictions appears in articles on my site. http://advancednano.blogspot.com


I find the idea of exponentially self-improving AIs kind of silly. Why do computers have to be equal to the human brain to allow this? If you could really program an AI into an endless positive feedback loop like that, it should be possible to do it already. It could start by designing very simple systems but, since we're talking about exponential progression, it wouldn't take long to reach "infinite" intelligence.

Come to think of it, why would a human-level AI have any advantage over a human anyway? If 6 billion humans can't design a superintelligence, who's to say a human-level computer could do the same? Well I know the answer: because computers can interconnect more effectively. But in that case the globally-connected web of humans today, allowing for faster and faster research, is like a singularity occurring in slow motion. The problem is that, while computer power is improving exponentially (for now), software isn't. My computer today is 10 times as fast as ten years ago, but does it do 10 times as much? Not really, it just has slightly better graphics.

Even if infinite increases in computer power did occur, I don't see how it would make all the other predictions go out the window. New technology cannot be adopted instantly because people still think and operate at a fixed speed; and as long as people live in "real reality" they cannot do things at infinite speed. And I'm still very sceptical about simulated consciousness, since consciousness is a subjective experience that can never be proven. People will be very reluctant to live in a simulation unless they're sure they retain (a) consciousness and (b) free will. It doesn't matter how wonderful the computer says it is. So a singularity would not put all the other predictions "out the window", at least not for those who remained behind.

To me the "singularity" is more like a slow-motion event as we develop into an interconnected global mind of fyborgs (functional cyborgs), and develop the technology to spread beyond the earth. I'm sure we will soon develop computers that can simulate a convincing human, but I don't see this as a particularly meaningful event. And we probably will develop smarter-than-human intelligences to perform an array of functions, but I don't think "merging with our computers" is quite as easy as Kurzweil makes it sound. I'm sure we will continue evolving, but we won't magically become transcendent gods.

I'm very excited by the progress of technology and if a benign singularity does occur, it'll be a neat bonus, but I'm sceptical of all this neat invisible-hand, exponential-increase, rapture-in-time, purpose-of-the-universe stuff precisely because it's so neat and explains the universe so cleverly. It all sounds too religious. We should be wary of any theory that (a) claims to reduce the universe to simple explanations or (b) announces we are at the end of the world or the culmination of history. I like Mike's "uplift" scenario more than his "nirvana" scenario. But anyway.


Holy moly that was long. Sorry. :-)


One other thing - the average home PC is about 100th the power of the human brain. Supercomputers are undoubtedly as powerful as human brains, and we've got hundreds of them. We've got more than enough computer power to create a singularity right now, so why isn't it happening? Software, software, software. First we need to write a program that can think as effectively as a person (very difficult). Then we have to teach it to improve itself without human intervention, faster than tens of thousands of IT researchers are already doing. This is all really hard. For this reason I see computers of the future as our "transcendent servants" (to quote Kurzweil), maybe our virtual companions, our global administrators, our space explorers ... but instant God? Give me a break.


I find it funny that they concentrate on hyper computertechnology, internet and spacetravel. Where's the impact of nano (on biology) on individual organisms?


The human brain is estimated to clock in at about 10 petaflops/sec. 1/100th of that would be 100 teraflops/sec. Home PC's are nowhere near that and won't be for 10 years. I believe they come in somewhere around the 20 - 40 gigaflop range today. But, even that may be a high estimate.

Brian Wang

My list of predictions that I was referring to is at

they include more molecular nanotechnology predictions than the BT predictions.


Ian Pearson reckons that even a Playstation is 1% as powerful as a brain. Is he wrong? I guess it's hard to estimate how fast a brain actually is. But why does a self-improving intelligence have to be as smart as a human anyway? And how do we know that a superintelligence will be any better at improving itself than thousands of computer scientists? These are the questions that need to get answered. Meanwhile it's very exciting how computers and the internet are bringing peace and breaking down barriers - all I'm saying is let's not expect them to magically turn into a God that can solve any problem, since the theory is a bit tenuous. :-)

Lady luck

Actually the home pc comes in at ~1.5 gigaflops. Now the Cell Broadband Engine can come in at ~25 gigaflops and it's 9 SPE processors can come out at a wopping 256 Gigaflops. So we have a long way to go yet.


Pearson has a long history of colorful predictions. You can see some of his guesses from 1997 at the Internet Archive:



That didn't quite work; try this one and click on the "June 06, 1997" link at lower left.


michael vassar

Got any reports from Nostradaemus you want to share Mike?
Pearson is a characature of a futurist. Please don't associate with him, it hurts CRN's credibility.

Daniel Hazelton Waters

The exponential curve starts to be noticed alot more in 2007- 2013. Just google up quantum computer timeline and see how close to the curve we are! Quantum computers are not just a way to crack current encryption schemes they will render nanotechnology machine learning and desktop manufactoring plus brian machine interfacing. By 2013

mrs smith

shut this websight down please :)

Michael Deering

Mrs Smith, is that a reference to the movie? If you want to stop people from getting ideas about nanotechnology you are going to have to shut down a lot more than just this one website. You're going to have to shut down the whole internet.

Michael Deering

Cool! I like your new human verification system. I wonder how long it will work? Only until the machines get smarter than us.

Tom Craver


Successful terrorist attack on Saudi Oil Infrastructure, oil goes to over $10/gallon 2006-2010

I think that's a minor slip - you must have meant "gas goves over..."

Otherwise, I'd agree that your list has far fewer odd predictions.


Now lets look at what people in 1950s thought:

2000-Flying cars
2001-Robot maids
2003-Nuclear tech making oil obsolete
2005-Cities on the moon

What really happened:

2000-cars using 80yr old combustion engine tech and still 4 wheels just like the 1950s.
2001-the only robot in sight for purchase is a lame robot dog that does 5 extremely simple pre programmed things
2003- Oil prices high and climbing with more demand than ever.
2005-we couldn't even make new orleans safe from flooding, the only time we were up on the moon was 40 years ago.

Brian Wang

the robot things are happening.

Millions of broombas for vacuuming.

Robots for industrial assembly.

also, there is intelligence and time saving for food preparation. Timed cooking, microwaves, food processors etc... The automation is there in appliances just not all in one place.

What are the desired functions that people will pay for ?

flying cars : the people in the 1950s were wrong not about the technology for a cheap flying vehicle but about how much people would really want it to overcome issues of regulation and safety and affordability.
cheap jets as an air taxi service between regional airports is closer to an easy and affordable availability and widespread air travel service. 43,000 die from driving cars, I do not think we really want the driving population flying without a supersafe automated system.
the lesson for predictor is not just is it technically feasible but what form is the easiest way to deliver it and are there training and safety and business issues that are involved.

Nuclear tech versus oil: The lesson for predictors of the 50's here is do not underestimate the current technology or switching costs.

Cities on the moon: this was also not a deeply thought out issue. Plus I would think that most people did not think we would be on the moon. Most people were surprised and amazed that the space program delivered the results that it did. Some serials had predictions on men in space (Analog etc...) But this was not a collective expectation.

Comparing what we had in the last major colonization phase. From Europe to North America.
There were large fleets (hundreds) of sailing ships from many nations able to move hundreds of people at a time.

a big commitment to an Orion rocket. could have and could still land thousands of people and tens of thousands of tons of supplies and equipment. Making a viable self sustaining colony. It would take a president with 100 times the vision of Kennedy to have done this.

costs have to come down for space.
what is the economy for a moon colony ? Platinum maybe. helium three in the future. tourism.

We have a longer wait for more countries and companies to get into the space game.

Brian Wang

References on historical colonization.
About 350,000 people migrated to the Americas in the 1600's. 1.5 million in the 1700's. By 1670, There 500 crossings going up to 1500 by 1730. Each of the ships could carry up to 200 or so colonists. Many were moving goods back and forth.


Point being: to colonize space. You have to make a lot of trips (or move a heck of lot of stuff in fewer trips) and move a lot of stuff and live off the resources that are there. Plus there should be economic reasons or strong societal reasons for it.

The comments to this entry are closed.