[TML] Future wars will be run and won by computer control

[Jerry W Barrington]

On 8/9/08 1:42 PM, "Garry Ward" <garry.e.ward at worldnet.att.net> wrote:

> For humans, yes. To use as a justification of what computers can do based on
> a rigged game makes the example meaningless.
> 
> Take away the special accomodations for the computer and if it still wins,
> then I'll be impressed.

It's just a handicapping system, like in golf.  Thing is, there are 46
levels of rating in go, and if I understood right, the "accommodations"
amounted to handicapping the master about 9 levels.  So if that places the
program about 9 levels below the master, then it's still pretty good.  Since
he's in the pro ranks, that would put the program in the advanced amateur
ranks.

Of course, *1* game doesn't yield a good rating.  It'd need to play many
games against many players to be able to tell.

> Both chess and go operate in a very small area with very specific rules
> which can be specifically coded for. The real world has a whole lot more
> variables to deal with and very few rules (outside of certain physical
> aspects) that are particularly precise.
> 
> Computers will always make very good tools, even very expert tools in
> specific areas, but they will always, eventually, come up short.

Well, that's an opinion, and of course you're entitled to it.  But we really
have no evidence to prove it.  As Douglas quoted earlier this week:
    When a distinguished but elderly scientist states that something is
possible, he is almost certainly right. When he states that something
is impossible, he is very probably wrong.
    Arthur C. Clarke, Clarke's first law