Author: Albert Silver
Date: 08:23:50 05/20/00
Go up one level in this thread
On May 19, 2000 at 11:52:25, Dann Corbit wrote:
>On May 19, 2000 at 11:24:03, Albert Silver wrote:
>[snip]
>>Really? You think that if I were to try to qualify for the Olympic 100m dash by
>>inscribing a slow remote controlled 10 cc motorbike that could go at 5 mph tops,
>>the competitors would have no problem with this? Everyone would give me the
>>thumbs up? You can organize races with such participations, and have the bike
>>win or lose, but are you going to declare the bike a national champion if it
>>wins? Not at all, you'll just register the result, note the progress of machines
>>in the area and that's it. Maybe get a nice newspaper piece if you're lucky, but
>>so what?
>
>When I think of it, I am sure that you are right about that one. In fact, there
>are lots of mechanical devices you can put on runners that really would make
>them faster or better but we don't. I'm really not sure why that is, but I
>suppose it has something to do with nostalgia.
>
>>Computers and thus the programs, are getting there, and there is no question
>>about it. Now we are waiting for entry-level PC programs to to reach GM level,
>>or at least have it generally recognized. Soon, it will be 2600 I guess, and
>>then even more. Eventually, programs won't even have to be as specialized
>>state-of-the-art programming in order to beat the best of us. Fine. I have NO
>>doubts about it, and any concern I have isn't whether they are stronger, but on
>>the future of my passion as openings really are overanalyzed, killing much of
>>the real adventure and discovery.
>
>Here, I am fairly sure that you are exactly wrong. That's because you don't
>know how large the canvas is. The advances from human learning will continue to
>outstrip those from computer learning for years to come. Look at the chess
>played today verses back in Paul Morphy's day. The rules were very nearly the
>same, but many openings that were thought to be sound have been busted and many
>new openings have been discovered. Are these discoveries bad? Should we still
>be playing the old, busted openings in blissful ignorance?
>
>Chess is a search for a level playing field. If one opening has a decisive
>advantage for either player, people will stop using it. If it is an advantage
>for your opponent, you will obviously aim for something else because you will
>have become aware of that. If we discovered a million new openings a year, it
>would be thousands of centuries before we even uncovered the tip of the iceberg.
>
>Real adventure and discovery can come from humans or computers. What we will
>see is more and more adventure and discovery, not less and less.
>
>>Fritz could very well be the absolute best.
>>Unbeatable, unstoppable, a steam-roller. I still wouldn't make it national
>>champion. I just don't understand the point. You can measure it, and it can
>>outperform us, but that doesn't mean it is competing. It is a machine, that's
>>it.
>
>It enters a contest. It wins or loses. That's competing.
I disagree. It doesn't enter a contest. It is entered into a contest. It may
seem like nitpicking, but I see that as a fundamental difference. A competition
is a moment where each tries to outdo themselves and others. A national
championship is where those qualities/elements are at their highest, and the
title of national champion is the summit of this (on a national level). I do not
see a program as accomplishing any of these things. It may be stronger and it
may be faster, but it will never be a competitor. Sentience is required for
that. You can inscribe a motorcycle in a race (yes, more of that tiresome
analogy), and it may finish the race in 1/10 the time, but it certainly never
'competed'. To declare a machine champion is baloney.
It is certainly a matter of time before we can no longer compete against
machines, but it is always we who are competing, not the machines. It is OUR
vanity and pride that bring us to want to measure ourselves against our
creations, but in the end it will be just another John Henry story, and we will
undoubtedly glorify the people who still can until they can't. WE compete,
programs do not. They perform.
>
>>I'm into computers, chess, and chess programs as much as the next, and never
>>cease to be amazed at how things have evolved, but this makes no sense to me.
>>Really not.
>
>I don't think it has to make sense to you. For me, I want to see the best chess
>possible from whatever source. Whether it is Kasparov or Deep Blue or Crafty on
>a 10,000 node Beowulf cluster, I don't care much.
I agree. The best chess is great, but let's not confuse the issues. I said that
I am in favour of seeing how Junior will perform (not compete) in Dortmund, but
a national championship is another issue. It isn't just a tourney, and what it
is about has no place for measuring machines. That's all.
>
>The reasons I think machines should be allowed are manyfold.
>1. Have the best play the best, whatever that is.
>2. Chess is not as interesting to the general public as it once was
>3. The "John Henry" thing
>4. It gives people something interesting to talk about
>5. If faced with a challenge that is undecided, I would like to see it finished
>
>I really, honestly don't think your objection about national champion is why we
>truly don't want them in the contests. If, for instance, we let them compete,
>but the results did not count and they could not get any prize money, I still
>think people would object. It's fear of the machine. There are poems and
>stories about our fear of mechanization. James Thurber's works are wrought
>through and through with it, and that stuff was written decades ago.
>
>If a machine cannot be national or world champion, the single real reason for
>that is fear. And if it is better than the human, with or without a tag
>proclaiming it champion, it really is the champion. We just refuse to recognize
>it. In fact, by our refusing to play, it may be champion by default.
It will never be champion, though it may be the strongest.
Albert Silver
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.