Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Kurzweil's rebuttal to Paul Allen

technologyreview.com

Post reply on HN
Print view View original

51–60 of 123 posts

Re: Kurzweil's rebuttal to Paul Allen

#51

Post by hugh3 »

I grow tired of Kurzweil's vague arguments against people who disagree with his vague predictions.

What I think Kurzweil doesn't understand is that in any argument about what's going to happen in the future, the onus of proof inevitably lies with the guy saying "This is what's going to happen", not the guy saying "Ehh, maybe not".

I don't know what's going to happen in the future, and I don't pretend to know what's going to happen in the future, but whatever happens either (a) I'll find out eventually or (b) it'll happen after I'm dead anyway. But john_b's point about Kurzweil's lack of a null hypothesis is a good one.

So my question for Kurzweil is this: what will the world look like if you're wrong? What possibilities are your predictions excluding? If I'm still alive in 2060, and I look around at the world around me, under precisely what conditions am I entitled to say "Well whaddya know, looks like Kurzweil was wrong about that Singularity thing after all"?

Re: Kurzweil's rebuttal to Paul Allen

#52

Post by api »

lupatus wrote: post #50
api wrote: post #44

Earlier quoted context omitted.

But how much do they want all this, and does this require major improvements in computing power? It looks to me that almost everything you listed could be done on the computers of five years ago. It's all nothing but software improvements. Existing hardware is good enough.

Two years ago, I was developing some HVAC equipment modeling software as part of a sales automation package whose worst case scenario needed to calculate the max cooling capacity, and a few other thermodynamic stats, for roughly 18 million different configurations. No matter how much I optimized it, the CPU didn't have enough juice to "instantaneously" plow through all of that math. The best I could get it down to wa…

But how much did that 20 seconds delay cost you? It would have been nice to have the answer faster, but how much would you have been willing to shell out for it?

I'm talking about economics. I'm talking about what people are willing to pay for.

Re: Kurzweil's rebuttal to Paul Allen

#53

Post by cpeterso »

Here is a rough transcript of a Long Now Foundation talk by SF author Vernor Vinge (and coiner of the term "singularity") entitled "What If the Singularity Does NOT Happen?". He sees:

  * Scenario 1: A Return to MADness (nuclear war)
  * Scenario 2: The Golden Age (peace and prosperity)
  * Scenario 3: The Wheel of Time (catastrophic natural disaster)
http://www-rohan.sdsu.edu/faculty/vinge/longnow/

Re: Kurzweil's rebuttal to Paul Allen

#54

Post by api »

hugh3 wrote: post #51

I grow tired of Kurzweil's vague arguments against people who disagree with his vague predictions. What I think Kurzweil doesn't understand is that in any argument about what's going to happen in the future, the onus of proof inevitably lies with the guy saying "This is what's going to happen", not the guy saying "Ehh, maybe not". I don't know what's going to happen in the future, and I don't pretend to know what's g…

IMHO, the future cannot be predicted. Period. I'm not talking impossible-as-in-hard. I am talking impossible-as-in-perpetual-motion.

Nice well-behaved linear Newtonian systems can be modeled and predicted. There are systems that are chaotic but that can be modeled in the aggregate very well too, like thermodynamic systems and certain kinds of fluid flow.

Life isn't like any of that. Life is complex, chaotic, computationally irreducible, and full of feedback loops on top of feedback loops. Even worse: predictions often create economic incentives to prove them wrong. Take a position on the stock market and you have created an incentive for your prediction to not come true.

People have always wanted to deny the fundamental unpredictability of history, and have always clung to woo-woo prophecy superstitions toward this end. The ancients had Tarot cards and pig entrails. We have graphs and computer models.

Re: Kurzweil's rebuttal to Paul Allen

#55

Post by hexagonc »

api wrote: post #30

Another response: Kurzweil also ignores economics. The advance of technology is driven in part by economic forces. Computing power may stagnate not because we have reached physical limits but because present-day computers are good enough for what 98% of the market wants. I see this trend developing. If anything, the trend in consumer computing is toward less powerful but lower-power and more portable computing device…

I don't think Kurzweil ignores economics at all. In fact, many of his arguments are largely based on economics, such as the cost to obtain a given amount of computation. It is undeniable that this has gone down steadily with time. Computation is so cheap that cloud computing (e.g., E2C, Azure) allows individual developers access to more computing than most know what to do with.

Your example of the MacBook Air is invalid because the MacBook Air of today isn't priced principally on performance -- it is priced mostly by build quality and performance/watt/kilogram. Never before have computers had a better performance/watt/kilogram profile. You have to consider the whole package.

The argument that there is less need for performance is also flawed. It is not that humans have less need for computation, it's that the distribution of computation is changing. As mentioned before, a lot of computation is moving to the cloud. In many cases, this is simply the most efficient place for it to be; instead of having an abundance of computation sitting mostly unused on a laptop or desktop, computation is becoming a service where it is used on demand. Think of the computation required for cloud services like speech to text processing or Google Goggles. Most of the computation is farmed out the servers in cloud and only minimal.

I'm not even sure I agree that the decrease in local computational needs is a long term trend. For one thing, migration to the cloud can only continue as long as network bandwidth keeps up with cloud data demands. If network bandwidth is not able to keep up then I predict we'll see local computational needs spike again.

Even with some computation siphoned off to the cloud, AI will create a whole new class of applications for even local computation because it is so data intensive. When the next crop of AI applications emerge and become more common, the early algorithms will probably be inefficient, which will itself cause a spike in computational needs. The applications are here yet but there are many (including myself) that are working hard to change that.

Re: Kurzweil's rebuttal to Paul Allen

#56

Post by lupatus »

api wrote: post #48
lupatus wrote: post #46

Earlier quoted context omitted.

I totally get EROEI. In fact, I used to be the canned beans buying type. But, after years of thought, I realized that the EROEI has a fatal flaw. It completely discounts humanity's ability to discover creative efficiency improvements. The EROEI calculations are done with today's technology, and do not include future improvements. For example, I grew up in Western Pennsylvania, where the oil industry started back in t…

I am in neither the doomer nor the pollyanna camp on this, as I said. Keep in mind that EROEI includes not only the energy cost of operating technology but also the energy cost of developing it. Every engineer who worked on that fracking technology? Every car trip they made to/from the office, etc., all has to be included in the energy investment required to get that shale gas out of the ground.

Yup, I totally get EROEI. But, it ignores possible efficiency improvements when it makes it's future projections. What do they say in the finance world? Something like "past performance is not an indicator of future returns."

What if in 5-10 years those engineers ride to work in self-driving robo-cars that are all electric, using cheap and powerful graphene-based super capacitors as batteries. They will have time during the commute to daydream about "wild" ways to make the fraking technology itself more efficient. And the all electric robo-cars are charged up from a cheap, efficient liquid fluoride-thorium reactor generator station. Or maybe the they are just daydreaming about an upcoming vacation to outer space while riding in luxury two-stage zepplins[1]. Oh, and the engineer is healthy and performing at peak mental and physical performance nearly continuously because he/she eats a ketogenic diet[2] centered around high-latitude reindeer herding products[3].

There are a lot of ways the world can be better.

[1]http://www.jpaerospace.com/

[2]http://www.ketogenic-diet-resource.com/

[3]http://en.wikipedia.org/wiki/Lomen_Company

Re: Kurzweil's rebuttal to Paul Allen

#57

Post by bh42222 »

endtime wrote: post #43
bh42222 wrote: post #23

Allen writes that "the Law of Accelerating Returns (LOAR). . . is not a physical law." I would point out that most scientific laws are not physical laws, but result from the emergent properties of a large number of events at a finer level. A classical example is the laws of thermodynamics (LOT). If you look at the mathematics underlying the LOT, they model each particle as following a random walk. Oh that's a terribl…

>Oh that's a terrible point! Thermodynamics laws are nothing like predictions about the future. I would have thought linguistic slight of hand like this is beneath Kurzweil. Could you elaborate on this? I'm not a huge Kurzweil fan, but as far as I can tell he's saying something reasonable here - that when he talks about LOAR, he's describing a phenomenon rather than a physical process, and that this is an accepted us…

Our understanding of thermodynamics is very thorough. Our understanding allows us to make a plethora of predictions, all of which are falsifiable, and have been throughly tested over the years.

This is what makes our theories about thermodynamics real scientific theories.

Predictions about the future, no matter how simple or based on long running past trends, are only falsifiable in exactly one way: wait until the predicted date passes.

I think a very, very informal use of the term "law" could cover both. But what irks me as a science minded person, is that Kurzweil is attempting to equate the informal meaning of "law", generic description of a phenomenon, with the scientific "law", an actual testable, falsifiable theory with predictive powers.

Re: Kurzweil's rebuttal to Paul Allen

#59

Post by lupatus »

api wrote: post #52
lupatus wrote: post #50

Earlier quoted context omitted.

Two years ago, I was developing some HVAC equipment modeling software as part of a sales automation package whose worst case scenario needed to calculate the max cooling capacity, and a few other thermodynamic stats, for roughly 18 million different configurations. No matter how much I optimized it, the CPU didn't have enough juice to "instantaneously" plow through all of that math. The best I could get it down to wa…

But how much did that 20 seconds delay cost you? It would have been nice to have the answer faster, but how much would you have been willing to shell out for it? I'm talking about economics. I'm talking about what people are willing to pay for.

I am also talking economics.

The customer would have loved for it to go faster, because then it would have looked like magic.

In 1998, it would have been nice if I had had photo-realistic computer games, but I patiently made do with what I had. Today, people eagerly shell-out money for hardware that can run this year's even-more-photorealistic version of Madden Football.

I think the demand for more power is definitely there.

Re: Kurzweil's rebuttal to Paul Allen

#60

Post by cadalac »

api wrote: post #32
kenjackson wrote: post #31

What are the big advances in linear programming that happened since 1988? As api mentions in the comments here on HN there are areas of work where progress stopped. For example passenger jet speed was once thought to continue at a rapid rate such that LA to Europe flights would be a few hours. Skyscraper height was thought to continue, with advances in various technologies and engineering methods making it desirable.…

What passenger jet speed and skyscrapers really hit is economics: demand limits to growth. Most super-tall skyscrapers are economic disasters. There seems to be a maximum economically rational height to a skyscraper, and it's already been reached. You can build higher, but if you do you're wasting your money. A human-level or beyond AI would probably be like the Burj Khalifa: an economic disaster. Why build it when s…

The super tall skyscrapers are indeed an economical disaster in them selves, but they skyrocket the value of the surrounding land. The Saudi Prince who's building the Kingdom Tower owns a considerable amount of land around the building site so he'll make a lot of money off the land he sells. It kind of reminds me of how Google makes money with adverting on their free products.
Post reply on HN