edit: Without looking up who this guy is and what he himself says he plans to do, this seems to me an obvious case of popular media oversimplifying an otherwise reasonable (albeit ambitious) project. The same thing happens with artificial intelligence all the time. Invariably, someone would be working on, say, simulating a promising, but tiny piece of how the brain works and the media would say "artificial brains, intelligent robots zOMG!1"

That said, leaving out any normative statements about whether such a predictive technology should exist, and whether it would do more harm than good, the fact is that it is close to impossible to implement at this point.

One obvious reason is that many of the subsets of the problem this article claims will be solved have been tackled by brilliant people for many years without solution.

Even more discouraging is that for some of these problems, it has been shown that there are inherent theoretical limitations on what we can know and predict.

For example, many of the systems this article claims will be simulated could be chaotic. You might have heard of a chaotic systems before called "the butterfly effect," also known as "if a butterfly flaps its wings in japan, it can cause a hurricane in the us." Mathematically, this means that in a dynamical system, small changes in the initial configuration of the system can cause dramatic changes in how the system ends up behaving. Suppose the stock market is a chaotic system. This means that if you make a perfect model to predict how stock prices evolve given an initial set of stock prices, but you make a measurement error of 0.00000001 in the initial stock prices, the results of your simulation will be wildly different from reality.

There are other theoretical problems as well. The chaos theory example assumed you could learn perfect models. But you can't even do that. Learning from data has theoretical limitations as well. In the limit of infinite data, everything is fine from a statistical-learning-theoretic point of view, but the fact is that we have finite data, if the model is very complex, there might not be enough data to make any accurate predictions.

That said, I am incredibly interested in complex systems theory, especially the dynamics of social and economic networks, two things whose complexity has so far prevented any deep understanding. The former (social networks) is a relatively new field of study, and the latter (economics) has so far relied on simple models. Simple models are not necessarily bad -- in fact, a simple model that explains X,Y and Z is preferable to a complex on that explains X,Y and Z. But consider the newtonian model of gravity, or initial models of electromagnetism -- they were simple, and explained a lot, but digging into how they really worked produced a lot of insight into the nature of the universe. In the same way, digging into complex systems and trying to uncover general principles could result in a profoundly different understanding of many things we take for granted.

Last edited by tachobg on May-07-2010 at 19:38

quote:

Originally posted by woscar Still haven't grasped the point of that talk, eh?

still haven't grasped the implications of his points, eh?

quote:

Originally posted by tachobg That said, I am incredibly interested in complex systems theory, especially the dynamics of social and economic networks, two things whose complexity has so far prevented any deep understanding. The former (social networks) is a relatively new field of study, and the latter (economics) has so far relied on simple models. Simple models are not necessarily bad -- in fact, a simple model that explains X,Y and Z is preferable to a complex on that explains X,Y and Z. But consider the newtonian model of gravity, or initial models of electromagnetism -- they were simple, and explained a lot, but digging into how they really worked produced a lot of insight into the nature of the universe. In the same way, digging into complex systems and trying to uncover general principles could result in a profoundly different understanding of many things we take for granted.

http://www.wired.com/techbiz/it/magazine/17-03/wp_quant

thank god that guy was an economist and not a living earth simulation scientist

quote:

Originally posted by tachobg This means that if you make a perfect model to predict how stock prices evolve given an initial set of stock prices, but you make a measurement error of 0.00000001 in the initial stock prices, the results of your simulation will be wildly different from reality.

exactly. and the more variables the even more wild the difference will be, and maybe even the set of variables deemed to be of enough importance are later discovered to only be partially or tangentially important according to the next theory du jour.

It's just a misguided notion, bottom line. Mechanics works on a reduced, abstracted level of complexity, not with a level of realistic complexity.

Last edited by nefardec on May-07-2010 at 19:44

quote:

Originally posted by nefardec http://www.wired.com/techbiz/it/magazine/17-03/wp_quant thank god that guy was an economist and not a living earth simulation scientist

quote:

it became so deeply entrenched—and was making people so much money—that warnings about its limitations were largely ignored.

This is not a problem with science, but a problem with the application of science. If the motivation is a humble desire to understand, this would not have done any harm, as the theory was known by Li to have limitations.

quote:

Originally posted by nefardec exactly. and the more variables the even more wild the difference will be, and maybe even the set of variables deemed to be of enough importance are later discovered to only be partially or tangentially important according to the next theory du jour.

Absolutely -- scientists unavoidably have to build simplified models, and the choices made in building those models are undoubtedly biased by everything from our perceptual capabilities as humans, to the culture of our workplace/nation/scientific field and the scientific modes of thought at the time.

quote:

Originally posted by nefardec It's just a misguided notion, bottom line. Mechanics works on a reduced, abstracted level of complexity, not with a level of realistic complexity.

I think most scientists don't adhere to reductionist thinking. They acknowledge the fact that they probably can't ever come up with perfect set of rules. But just because a set of perfect rules is elusive or nonexistent doesn't mean you can't uncover valuable scientific knowledge.

Even if as a scientist, you believe that there is a perfect theory of physics, you must acknowledge that when these rules come together, they can create emergent phenomena that are complex and difficult to understand. For example, if you understand all the fundamental laws of physics, this doesn't automatically make you understand turbulent fluid dynamics. You might not be able to predict everything about such a complex system. And that's okay! (as long as you know its limitations and don't do anything stupid)

Last edited by tachobg on May-07-2010 at 20:01

quote:

Originally posted by nefardec still haven't grasped the implications of his points, eh?

LOL. I said what I said above because it's mind-numbingly obvious that the main point of that talk has flown over your head when you make statements like:

quote:

Originally posted by nefardec some science fanboy wanted to create a global set of rules for people based on scientific observations and deductions.

How on Earth did you get the notion that he wants "to create a global set of rules"? Like I said countless times on that thread, he is merely advocating for a somewhat unified moral criteria that is reached by taking a scientific approach.

___________________

My Set Archive - MY BLOG

Just another attempt for humans to explain the pretentious mystery of our existence and future.

quote:

Originally posted by woscar How on Earth did you get the notion that he wants "to create a global set of rules"? Like I said countless times on that thread, he is merely advocating for a somewhat unified moral criteria that is reached by taking a scientific approach.

um... 6 of one half, dozen of another?

quote:

Originally posted by tachobg This is not a problem with science, but a problem with the application of science. If the motivation is a humble desire to understand, this would not have done any harm, as the theory was known by Li to have limitations.

Sure, as I said, it was a problem with reductionist, industrial modes of thinking.

quote:

Originally posted by tachobg I think most scientists don't adhere to reductionist thinking. They acknowledge the fact that they probably can't ever come up with perfect set of rules. But just because a set of perfect rules is elusive or nonexistent doesn't mean you can't uncover valuable scientific knowledge.

The guy who came up with this idea obviously does adhere to such thinking, and if he doesn't he's never going to be able to come up with a result, a budget, or funding.

I agree that you can uncover 'valuable knowledge' in the process, but you have to ask to whom this knowledge is valuable, and ultimately 'Why?'. Why should we do this? Who is going to benefit the most? What are its implications for freedom, creative expression, human experience?

I think this enterprise ultimately presents more danger to humanity than good.

quote:

Originally posted by nefardec um... 6 of one half, dozen of another?

You can be gigantically dense some times. There is a huge difference in trying to unify morality so that women having the misfortune of being born in certain Islamic or Animist cultures don't have their clitoris chopped off excused with idiotic arguments of "cultural differences", and wanting to create write a new moral code into stone tablets.

___________________

My Set Archive - MY BLOG

quote:

Originally posted by woscar You can be gigantically dense some times. There is a huge difference in trying to unify morality so that women having the misfortune of being born in certain Islamic or Animist cultures don't have their clitoris chopped off excused with idiotic arguments of "cultural differences", and wanting to create write a new moral code into stone tablets.

And you can be repeatedly, needlessly, personally offensive - for what reason?

So if you're saying that the only implication of what this TED talk guy was saying is that people should fund another international committee that tells people to play nice, then what he's saying is even sillier, and we already have something called The Universal Declaration of Human Rights, so what he's saying is not even original either.

quote:

Originally posted by nefardec And you can be repeatedly, needlessly, personally offensive - for what reason?

Why is being called "dense" so offensive to you? I mean, that's what I perceive from the discussions that we have recently engaged in. I perceive a smart man simply wanting to miss the point on purpose.

quote:

Originally posted by nefardec So if you're saying that the only implication of what this TED talk guy was saying is that people should fund another international committee that tells people to play nice, then what he's saying is even sillier, and we already have something called The Universal Declaration of Human Rights, so what he's saying is not even original either.

Seriously, why do you keep insisting that what Sam Harris is proposing is a written document of any kind telling people what they have to do, and then make sure it is enforced by authorities or an international committee?

He's simply arguing that a scientific approach based on the well-being of conscious creatures is a much better approach to morality than scripture, superstition, ancient myths and folklore. This is an individual process that each person is to take, and not imposed by him or any other moral philosopher or scientist for that matter. That's the difference between his thesis and that of the Rabbis, Popes, Ayatollahs and dictators that you so desperately want to have it associated with.

I don't think I can say it clearer than that, so if you don't grasp it you are either dense, stupid, or a monumental troll and Jay would be proud of you. And I know for a fact that you're not stupid.

___________________

My Set Archive - MY BLOG