Quoted at Brain Pickings, which expands to something said by "the great Polish-born British mathematician, biologist, writer, and historian of science Jacob Bronowski":
Evolution is built up by the perpetuation of errors. It runs counter to the second law of thermodynamics by promoting the error to the new norm so that the second law now works on the error, and then a new error is built up. That is also central to all inductive acts and all acts of imagination. We ask ourselves, “Why does one chess player play better than another?” The answer is not that the one who plays better makes fewer mistakes, because in a fundamental way the one who plays better makes more mistakes, by which I mean more imaginative mistakes. He sees more ridiculous alternatives… The mark of the great player is exactly that he thinks of something which by all known norms of the game is an error. His choice does not conform to the way in which, if you want to put it most brutally, a machine would play the game....
Posts
27 comments:
An explanation for Trump. of course.
Fear of doing things "wrong" supposes there's a definite and infallible set of rules governing choices. But that doesn't apply to things like cooking, marriage and especially these days, politics. How often has Trump done or said something "wrong" only later to be proven right, and in doing so rewrites the rules as he goes. Remains to be seen, but I'm reluctant to say the outcome of the recent health care bill was result of "errors", the rules of governance are being reshaped as we speak. Unless of course as Spicer said, Trump is caught putting Russian dressing on his salad tonight, then we'll have something "wrong" to talk about.
The only people not making any mistakes are the people who aren't doing anything.
The fear of doing anything wrong is terribly disabling.
This article goes way beyond inviting people to cast off that fear (which I experienced far to much in my life and know how much it held me back).
It says mistakes are a positive good, crucial to creativity and progress.
Far TOO much.
Man, I'm glad I made that typo. The good that may arise...
Guy in the Blue Corolla says....
You have to make mistakes to figure out what the mistakes are. I used to just drive around the city, trying to talk girls into my car, but it rarely worked. I tried pretending I was lost and asking for directions, then I pretended I needed help looking for my lost kitty kat, and things got a bit better, but I was still unsuccessful far more often than not...
Then -- when I DID get a girl in the car -- I soon realized you can't get the duct tape on their mouth while you are also trying to drive: it just doesn't work. Also: if they are not properly tied up they can jump out of the car at red lights -- people notice those kinds of things, and some may even be able to remember your license plate number...
I realized you get a head-start by picking up girls that are visibly drunk outside of clubs at the end of the night. You tell them "Hey? Remember me? We talked at the bar a week ago -- you told me about that girlfriend of yours who was being such a bitch lately. Need a ride?" Important thing I have learned: drunk girls ALL have a girlfriend who has been being such a bitch lately...
The best part about the drunk girls is that most of them will have passed out by the time you drive a few blocks, which then makes tying them up a LOT easier. Of course, when you duct-tape their mouths it increases the chances that they will choke on their own vomit, so you have to pay attention: this is why it is also important to keep a plastic bucket in the back seat -- I learned THAT one the hard way...
I could tell you more, but I don't want to give away ALL of my hard-earned secrets. Sometimes you just have to decide you're going to do what you're going to do, and then figure things out on the way. One last piece of advice, though: if you do not know how to tie a good, sturdy knot you are NOT ready yet....
I am Laslo.
"mistakes are a positive good, crucial to creativity and progress." True, but not very informative. Gotta control for the kinds of mistakes and the ways they are handled. Plus not all mistakes are created equal; some are just dumb. "_____ are a positive good, crucial to creativity and progress": quite a few right answers.
Not sure if the whole Bronowski point about chess is actually true at all.
The mark of the great player is exactly that he thinks of something which by all known norms of the game is an error. His choice does not conform to the way in which, if you want to put it most brutally, a machine would play the game....
I don't claim to be a good chess player but I know the computer always beats me. ;-) My theory about the losing chess player is the he/she is often so caught up in his/her strategy that he/she loses track of his/her opponent's tactics.
OMG! You mean I've been winning all my life by fucking up!?!?
Awesome!
Sebastian said... [hush][hide comment]
Not sure if the whole Bronowski point about chess is actually true at all.
It is true, occasionally, at lower levels
Computers win without any imagination, and does not see more ridiculous alternatives.
It is true, occasionally, at lower levels
Computers win without any imagination, and does not see more ridiculous alternatives.
Aha! So that's the secret of beating the computer? Making ridiculous moves? I'll try it!
I used to play bridge with some guys during lunch at work and one of them had [seemingly] no clue of how to bid but would usually win his contract, anyway. Of course, it didn't hurt that he could also remember every card in every trick and who played it.
I learned to weld from a man called Edsel Ford. He was an aviation professional, and did not tolerate anyone goofing off, or doing something that was exactly what he told us not to do.
He came to my welding station one day, and watched me. He then tapped me on the back, grabbed my work with his pliers and threw it into the sand pit.
I told him I was experimenting with trying to save the piece from my error. To study my error. He told me that the error deemed the piece worthless, and just think if someone depended on it for their life. The most important thing he wanted in our minds was:
Practice does not make one perfect. Perfect practice makes one perfect.
If what you are doing is not perfect, then you need to throw it in the sand pit, and start over. All we want to deliver to people who are going to be flying at 100 MPH is perfection. It really pisses pilots off when a weld breaks.
Let me tell you, the history of aviation is filled with good intentions. But none of that is going to bring back Lt Selfridge, or fix his crushed head. Orville Wright did not deliver perfection when his plane came apart in the air. It ruined his life.
Thankfully, Etienne, not all of us build aircraft.
Mockturtle, when I was playing chess years ago, we learned that you needed in the early stages of the game to make a move that would take the computer out of its "book."
The book was its memory of the great games and the most effective countermove to any move whether you're white or black.
If you see grandmasters play, the first 20 or so moves go very very quickly. That's because they're not brilliant (at that point), they're going through moves that were deemed most optimal for that position. Once someone does something that's not in the book, the game slows down and the real match begins.
Bronowski was born in 1908 and died in 1974.
His thoughts about chess and machines were written well before computer chess programs gained their present strength.
'...in 1976 Senior Master and professor of psychology Eliot Hearst of Indiana University wrote that "the only way a current computer program could ever win a single game against a master player would be for the master, perhaps in a drunken stupor while playing 50 games simultaneously, to commit some once-in-a-year blunder".'
https://en.wikipedia.org/wiki/Computer_chess
Things have since turned out to be 180 degrees opposite from Bronowski's:
"...The mark of the great player is exactly that he thinks of something which by all known norms of the game is an error. His choice does not conform to the way in which, if you want to put it most brutally, a machine would play the game...."
Just as there are sites that report changes in the ratings of the top chess players:
#1 Magnus Carlsen Elo rating 2838 (World Champion, Norway)
#2 Wesley So Elo rating 2822 (USA)
http://www.2700chess.com/
...there is also a similar one for top rated chess programs:
#1 Stockfish 8 64-bit 4CPU Elo rating: 3390
#2 Houdini 5.01 64-bit 4CPU Elo rating: 3387
http://www.computerchess.org.uk/ccrl/4040/
Top chess programs are now overwhelming stronger than the best human players.
A strange (but strong) "computer move" is now considered one that does not conform to the choice a human would make!
That is because the top chess programs can calculate moves accurately by sheer brute force so much further ahead than a human can.
Bill Peschel explains: That's because they're not brilliant (at that point), they're going through moves that were deemed most optimal for that position. Once someone does something that's not in the book, the game slows down and the real match begins.
So how does one decide when to make the unorthodox move?
Evolution is built up by the perpetuation of errors.
If he's speaking of biological evolution (link didn't work), that's teleology. Mutations and such are not errors.
I have 'Houdini'. Next time I play I'll throw a curve ball--or maybe a change-up--and see what happens. I'm sure I'll lose, anyway. It's most demoralizing, which is why I seldom play.
Evolution is a chaotic process by virtue that it is incompletely or insufficiently characterized and unwieldy. For example, a human life evolves from conception until its natural, accidental, or elective abortion.
My mistakes lapse into your virgin sister's asshole.
But only if she's younger.
Follow your heart
keep your ears
suicide is not painless, it's evil and hurts people
that song is so stupid
don't be afraid to paint
and feel the passion!
there's a great Van Gogh art show circling around, a 3-D immersive experience, with actors playing Vincent and Theo, a bunch of paintings and writings on the wall. It came to Charlotte, it was amazing. If you get a chance, go!
At one time I thought I was a player. Started when I was 6 years old. Dominated the high school chess club. Beat everyone in college. One night I ran into a guy in a bar. We played two boards simultaneously. He played with his back to the boards. I dropped thirty bucks in about 15 minutes. I haven't played much since.
I came up with a related speculation while reading Thinking, Fast and Slow by Daniel Kahneman: Most entrepreneurship and innovation is, strictly speaking, based in fallacious optimism about how things could turn out. Most entrepreneurs crash and burn, most innovations aren't necessary, many new businesses fail, etc. But the ones that do prevail, against the odds, drive economic prosperity.
So we may be rich, not in spite of our flawed mental processes, but because of our flawed mental processes.
"One night I ran into a guy in a bar. We played two boards simultaneously. He played with his back to the boards. I dropped thirty bucks in about 15 minutes. I haven't played much since."
Hey, I remember you!
Apparently the computer that is now beating top players at Go is doing so by using approaches that no human player would use for fear of being mocked. Since part of the fear of mistakes is a fear of mockery, artificial intelligence might actually prove to be more creative (i.e., more willing to make mistakes and learn from them) than human intelligence.
Post a Comment