How to deal with the economically redundant? Tax the rich

[Munger]

Re: How to deal with the economically redundant? Tax the rich

Not even close.

Evidence? Assertions that a computer will never beat a human at chess, Go, bridge, poker, etc etc have repeatedly been shown to be wildly overestimating the uniqueness of human thought. As uniquely human endeavors have fallen one by one to the wayside for centuries, I think the burden falls on you to demonstrate why it cannot continue.

[ThePythonicCow]

Re: How to deal with the economically redundant? Tax the rich

Evidence? Assertions that a computer will never beat a human at chess, Go, bridge, poker, etc etc have repeatedly been shown to be wildly overestimating the uniqueness of human thought. As uniquely human endeavors have fallen one by one to the wayside for centuries, I think the burden falls on you to demonstrate why it cannot continue.

For many specific problems within sufficiently narrow constraints, of course computers beat humans, sooner or later.

Similarly, for many specific physical tasks, machines beat humans.

Machines are a bazillion times stronger than I am. But after years of effort developing robots, this old man can still walk across broken terrain better than a robot.

The subtlety and complexity of the human mind is vastly beyond what man made digital computers will ever achieve.

[tacito]

Re: How to deal with the economically redundant? Tax the rich

Good stuff . But, I would have to disagree with the sentiment of this statement. Mathematical formulations are ALWAYS effectively metaphorical (That is what is wrong with Wall Street today) when mapped to the physical world. A real world understanding of Godel's theorem is that a model of the world cannot be realized that contains a representation of the model itself. The idea that a human being cannot model and predict its own behavior perfectly is the likely source of irrational behavior. Sometimes this paradox is realized as a limits to degrees of freedom (capabilities) and sometimes it is realized temporally (too much happening too fast). Why bring this up an economics forum? Because it puts you on the road to the paradox we face when considering the importance of a central bank or a global currency. IF irrational behavior in a system is proportional to the proximity of the current state to the maximum complex state then since we are hitting the limits of global economy and trading speed irrational behavior will (and is) increasing dramatically because of this fact and no other. In the past Free Trade was a great way to relieve this pressure as was computer trading and bring things back to a rationally functioning economy. What markets are there left to open? How to we make things move faster? What do we do now?

1. If the system is consistent, it cannot be complete.
2. The consistency of the axioms cannot be proved within the system.

The quote above is taken from the wikipedia. What I was thinking that was the Godel theorem was number (1). That means that (under the mathematical theories that the theorem considers) there are assertions that are true, but cannot be proven. It seems that number (2) is what sustains your interpretation of the theorem. I'm afraid that this is just what P. Cow calls metaphorical.

About your theory about complexities in economics, my feeling is that excessive complexity cannot speed up anything, and that the best way of dealing with it is getting an "Occam's knive", cutting flesh and starting anew.

[Munger]

Re: How to deal with the economically redundant? Tax the rich

For many specific problems within sufficiently narrow constraints, of course computers beat humans, sooner or later.

Similarly, for many specific physical tasks, machines beat humans.

Machines are a bazillion times stronger than I am. But after years of effort developing robots, this old man can still walk across broken terrain better than a robot.

Ok. So far. 10 years ago scientists couldn't get a bipedal robot to balance itself while standing. Now they can run, pick themselves up when they fall down, and climb stairs. You think that's far as it will progress? And your reasoning?

The subtlety and complexity of the human mind is vastly beyond what man made digital computers will ever achieve.

Again, this assertions lacks evidence or reasoning. The same thing has been said too many times to too many other fields.

[ThePythonicCow]

Re: How to deal with the economically redundant? Tax the rich

Again, this assertions lacks evidence or reasoning. The same thing has been said too many times to too many other fields.

Of course my claim lacks convincing evidence. One can't easily show that such is unlikely.

[flintlock]

Re: How to deal with the economically redundant? Tax the rich

Getting back to the theme of this thread, when you tax the rich, they take their money somewhere else.

This is exactly right.

I just caught the tail end of it this morning, but Congressman John Linder from GA was discussing on the Neal Boortz syndicated radio show how Nancy Pelosi and her gang were bandying about the idea of taxing retirement accounts, something I mentioned a while back as a distinct possibility as they become more desperate. Imagine the capital flight that would cause.:eek:

[ASH]

Re: How to deal with the economically redundant? Tax the rich

Imagine the capital flight that would cause.:eek:

And the human flight.

[sunskyfan]

Re: How to deal with the economically redundant? Tax the rich

I tend to agree with Python on this one. We can't even come to good definition of intelligence let alone emulate it at smart animal level (including dolphins and such) let lone human beings. There is a huge difference between optimization given an objective and problem solving. Software can be written that searches for the best optimization (as in chess), or to implement control theory to walk down some steps. You would be hard pressed to find software capable of identifying a new problem in its environment and solving it.

[tacito]

Re: How to deal with the economically redundant? Tax the rich

And no one knows if Moore's law will hold up.

Moore's law will not hold up. Exponential functions do not hold up in a finite universe. Even Moore agrees that miniaturization will stop at least when we arrive to the atom level. And it's not holding in RAM and HD access speed now.

At which point a home PC will have the computing power of a human brain.

I believe that that is possible.

Can we program such a computer to work like a human brain? Not sure. There are several reasons I think we will be able to though: (1) we can artificially evolve the solutions

There is a constant pattern in artificial intelligence:
1) You identify a problem that only a human can handle. Playing chess, for example.
2) Someone makes a program that beats humans at that problem. (Deep Blue).
3) You analize the solution and arrive to the conclusion that this problem is easier than you thought. The people that are good at it are not that intelligent, only have good memory and concentration. But that program is still not really intelligent. Then you go back to point (1).

(2) we can reverse engineer the brain.

A lot of people think that if a teenager can be a hacker, an experienced programmer should have no problem hacking whatever computer program or breaking whatever password. That is not true. Teen hackers just google for a program that works and that they don't understand. Making those programs is not easy at all, even thought that computer programs are built having three principles, and having them in order:

(1) It should be understandable.
(2) It should be correct.
(3) It should be efficient.

If you want (3) and do not have (2), you have to get (2) first. If you want (2) but don't have (1), your best option is to start over again.

The human brain is way more complicated that any existing computer program, AND lacks (1).

So, I basically agree with you: it is possible to replicate (by reverse engineering or artificial models) the human brain. But this is a really, really hard task.

[Munger]

Re: How to deal with the economically redundant? Tax the rich

Moore's law will not hold up. Exponential functions do not hold up in a finite universe. Even Moore agrees that miniaturization will stop at least when we arrive to the atom level. And it's not holding in RAM and HD access speed now.

Agreed. I should have stated that the question is whether Moore's law will break down much sooner than at the atomic level. If it holds up that long there is no question we will pass the computational level of the human brain. This does not include factoring in quantum computing.

So, I basically agree with you: it is possible to replicate (by reverse engineering or artificial models) the human brain. But this is a really, really hard task.

Right. I didn't mean to imply it would be easy. I can envision it happening though. Combine

(1) exponentially increasing computing power with
(2) exponentially increasing MRI resolution with
(3) super-exponentially increasing ability to sequence DNA ...

The result?

[ASH]

Re: How to deal with the economically redundant? Tax the rich

Moore's law will not hold up.

Agree. The word "law" is a little grand, given what Moore's law really is.

If we're able to make quantum computers practical, then all bets are off. However, like reverse-engineering the brain, it is a damned hard problem.

[ASH]

Re: How to deal with the economically redundant? Tax the rich

Right. I didn't mean to imply it would be easy. I can envision it happening though.

The pessimist in me says that our rate of technical progress may slow dramatically for economic reasons, including peak cheap oil, the aftermath of the credit orgy, and associated societal strains. I think we're about to be a lot poorer as a society, and a lot more preoccupied with meeting the basic material needs of our citizens, and that spending on advanced research is likely to contract as a result.

[Diarmuid]

Re: How to deal with the economically redundant? Tax the rich

I suspect they said similar things when Watts steam engine threatened to replace the laborer.

Once again, humans will come to better understand what separates them from machines and simpler life forms.

PC I agree strongly.

Having worked in the technology industry for near 15 years , I have come to see even the most technologically advanced server or server farm as nothing more then a jumped up calculator, while software and hardware has improved exponentially, the underlying architecture has not, that is even the most sophisticated machines are still just dumb terminals flashing 1s and 0s albeit at an ever faster rate. Software can be created to harness these improvements in speed to do ever more complicated manipulations, but the software itself is still bound by the parameters and constraints of the software, which in turn is a function of the imagination, creativity and skill of the programmer.

For example eluded to in this post were improvements in chess -

"Computer chess programs consider chess moves as a game tree. In theory, they examine all moves, then all counter-moves to those moves, then all moves countering them, and so on, where each individual move by one player is called a "ply". This evaluation continues until a certain maximum search depth or the program determines that a final "leaf" position has been reached (e.g. checkmate)."

The computer does not play chess as such in that it is using imagination or creativity - it is just calculating probabilities in a Matrix according to input formula and giving a result - that result is portrayed on the screen and interrupted by the human player as a "move". The computer no more "understands" the idea of the move or chess as does a calculator understand the idea of the no. 4 when 2+2=Enter is hit on a calculator.

Chess computing software is a match for human players because the calculation horizon is now within a acceptable time frame.

However take a game like GO

"The game emphasizes the importance of balance on multiple levels and has internal tensions.


It has been claimed that Go is the most complex game in the world due to its vast number of variations in individual games.[70] Its large board and lack of restrictions allow great scope in strategy and expression of players' individuality. Decisions in one part of the board may be influenced by an apparently unrelated situation in a distant part of the board. Plays made early in the game can shape the nature of conflict a hundred moves later.

The game complexity of Go is such that describing even elementary strategy fills many introductory books. In fact, numerical estimates show that the number of possible games of Go far exceeds the number of atoms in the known universe.[nb 13]

Go poses a daunting challenge to computer programmers.^[71] While the strongest computer chess programs can defeat the best human players (for example, the Deep Fritz program, running on a laptop, beat reigning world champion Vladimir Kramnik without losing a single game in 2006), the best Go programs only manage to reach an intermediate amateur level. "

http://en.wikipedia.org/wiki/Go_(game)#Computers_and_Go


Even if Moore's law holds it will be many many years before an acceptable calculation horizon for Go is met by hardware limits. Go is also only a game with a limited fixed set of rules, what chance does a computer with current architectual and software limitations have for computing the complexities of situations where not only are the choices infinite but the rules change and vary and also express infinite possibilities?.


Computers are bound by the finite, a computer possesses intelligence only so far it is able to perform rudimentary calculations over and over again albeit very quickly, the computer is bound by the limits of this type of intelligence or lack of, as I would hold.


Human intelligence is bound by no such limits, creativity, imagination, intuition etc.. allow it to overcome what once may have been considered finite limits of particular problems or situations. Until such time as a computer is made to even rudimentarily emulate these types of qualities - imagination, creativity, intuition as opossed to just geting quicker and quicker at passsing 1s and 0s, I will maintain a computer as it is today has about as much chance of surpassing human intelligence as does a shovel.

[flintlock]

Re: How to deal with the economically redundant? Tax the rich

The pessimist in me says that our rate of technical progress may slow dramatically for economic reasons, including peak cheap oil, the aftermath of the credit orgy, and associated societal strains. I think we're about to be a lot poorer as a society, and a lot more preoccupied with meeting the basic material needs of our citizens, and that spending on advanced research is likely to contract as a result.

I agree and have said as much in other posts. Not all a bad thing imo.

[Munger]

Re: How to deal with the economically redundant? Tax the rich

For example eluded to in this post were improvements in chess -

"Computer chess programs consider chess moves as a game tree. In theory, they examine all moves, then all counter-moves to those moves, then all moves countering them, and so on, where each individual move by one player is called a "ply". This evaluation continues until a certain maximum search depth or the program determines that a final "leaf" position has been reached (e.g. checkmate)."

The computer does not play chess as such in that it is using imagination or creativity - it is just calculating probabilities in a Matrix according to input formula and giving a result - that result is portrayed on the screen and interrupted by the human player as a "move". The computer no more "understands" the idea of the move or chess as does a calculator understand the idea of the no. 4 when 2+2=Enter is hit on a calculator.

I take it you believe this is not how a human plays chess. I disagree. Humans do depth calculations just as computers. After enough practice humans can recognize patterns that serve as heuristics. There is no magic. No real difference.

However take a game like GO

"The game emphasizes the importance of balance on multiple levels and has internal tensions.

Go programs have progressed fantastically in the last year or so because of a focus on Monte Carlo methods and specifically the UTC algorithm. Give it 5 years and humans won't have any games that they can win.

Computers are bound by the finite, a computer possesses intelligence only so far it is able to perform rudimentary calculations over and over again albeit very quickly, the computer is bound by the limits of this type of intelligence or lack of, as I would hold.

We as humans are bound by no such limits...

Incorrect. Brains are bound by the finite, parallel computing power of about 100 billion interconnected neurons. There is no mysticism; you are not special.