[ExI] AIG Bail out

[Damien Sullivan]

On Wed, Sep 24, 2008 at 09:44:52AM -0700, Lee Corbin wrote:

> see a difference between our efforts to create an AI and our
> efforts to re-design societies and economies. Have you read
> Hayek at all?

Yep!  I remember thinking he was a lot more plausible and nuanced than
other libertarian writings I'd read.  Market eliciting information, not
assuming perfect information; market as a "fair game" if people
agreed to the rules and executed them, not because it produced fair
outcomes.

...later I wondered why people should agree to those particular rules.

> We have no choice but to try to engineer solutions to our own
> creations such as AI, something never before successfully tried.

Hmm.  The idea isn't just AI engineering; the idea is that the hybrid
model might capture something about how we actually work.  Strange Loops
and all.  Of course pure connectionism and anti-symbolism seem more
popular in cognitive science, at least locally, but they still have
conceptual problems with some phenomena.  Research and debate continue.

> Hayek warned strongly (as did Von Mieses) about the arrogance
> and vanity humans have when monkeying with societies. So far,
> all planned societies or planned economies have failed abysmally.

But those were pure top-down economies, not hybrids.  (I'm also not sure
if it's true; descriptions of Incan or Sumerian temple economies sound
pretty planned.  But yes, the Communists haven't done well.)

> But we have ample evidence (a good deal of it from the 19th
> century) that free market economics work. Perhaps the only

I think any real economy, apart from states of anarchy or Communist
ones, has been hybrid to some extent, with laws providing what I think
of as top-downness.  The differences have been in degree and goals.  The
19th century had economies which were pretty free in some respects;
after some decades of experience with them, people opted for more
regulation and safety nets, and continued to prosper.  Not always out of
charity; Bismarck just wanted to buy off the foreseen revolution of the
poor.  Whether this constitutes the economics "working"...

> issue is whether some people are so concerned about what
> may happen to the very poor that they believe governments

They're concerned about the very poor, and the equality of opportunity
of the next generation, from the very poor up through the middle class,
and about the security of democratic practices in the face of
concentrations of economic power, and about economic practices which
affect third parties -- supposedly something libertarians are also
against.

> if you must, and give it to the poor. Fine. But we must stop
> monkeying with the markets and stop the ever expanding layers
> of regulation, which just never seem to work.

The regulations expand, at least partly, to match ever expanding layers
of complexity and methods of interaction.  New technologies allow for
monkeying with societies and the environment in new ways.  The
laws needed by a bunch of scattered dirt farmers are not the laws needed
by 17th century cities in densely packed wooden buildings, or by 21st
century cities in an era of high explosives and nukes, or by people in a
world where one moderately wealthy country could change the global
climate with a determined effort.  (Should Denmark have the right to get
rid of the Greenland ice cap, which could raise sea levels by over 6
meters?)

Lots of those regulation are pre-emptive "don't kill us", "don't poison
us", or in the areas of asymmetric information, pre-emptive anti-fraud.
Sometimes tort works, sometimes a simple "no, don't do that" works
better.

-xx- Damien X-)