On Fri, 2007-05-18 at 20:55 -0400, Chris Fant wrote:
> > my thinking is that in a situation like this, hardware advances aren't
> > yet the main limiting factor.  and since the strength of 19x19 programs
> > of any variety are still many hundreds of ELO away from strong play, big
> > algorithmic changes (and not how those algorithms scale in the
> > distant hardware future) are still what really matters.

I apologize for this, but it really irks me for some reason when people
keep comparing humans to computers like you are doing here.  No
disrespect intended on a personal level to you, but here is my rant:

What does this comparison have to do with algorithmic changes?   You
seem to be using the logic that because computers are not at some
arbitrarily chosen level that you specify, that it's some kind of proof
that algorithmic breakthroughs are possible.

I hate this reasoning.   Algorithmic breakthroughs are either possible,
or they are not and it has nothing to do with how strong a few player
happen to be.

But you are not even being precise.  What is "strong play?"   How can
you be many hundreds of ELO away from strong play?   I think 5 kyu is
strong.   But strong compared to what?   And whatever you compare it to,
why is that specially significant?

I personally don't care how far much weaker or stronger computers are
than humans.   That means nothing to me and I don't know why so many
consider this important.   Computer chess program used to be a lot
weaker, now they are lot stronger.   Who cares?   I never assigned
special significance to the transition.    

And by the way,  you can never be sure major algorithmic changes are
even possible.  It's always nice to dream and imagine that they are (and
I think they probably are.)    However, we definitely know that hardware
will continue to improve for a while.  

So what we should be doing is exactly what we are doing now.   Continue
to look for the breakthrough's and improvements, but also continue to
build increasingly elegant and efficient hardware.   I don't know where
you've been recently,  but it's been demonstrated that hardware matters
a LOT in computer GO.     We should face that fact that you will have a
big advantage if you have a more efficient machine.    Even Lance
Armstrong would lose badly if he had to race using the balloon tired
heavy steel bicycle I had as a kid.

- Don



> Is skipping super-Ko in the tree search one of those big algorithmic changes?
> _______________________________________________
> computer-go mailing list
> [email protected]
> http://www.computer-go.org/mailman/listinfo/computer-go/

_______________________________________________
computer-go mailing list
[email protected]
http://www.computer-go.org/mailman/listinfo/computer-go/

Reply via email to