Chess, a Drosophila of reasoning | Hacker News Reader

Chess, a Drosophila of reasoning

science.sciencemag.org

72 pointsprostoalex20 comments

I followed the recent championship matches on chess24.com and was blown away at how compelling and engaging computers have made the game. There was a chessboard on the screen with the current board state and a computer analysis on the side estimating which player had the advantage and by how much. While waiting for players to move, I was able to test out my own ideas on the board of which moves I thought were best and see what the computer thought of them (every move I made pretty much threw the game to the opponent).

After losing to Deep Blue, Kasparov came up with the concept of "Advanced Chess," where players team up with a chess program of their choice to augment their play. At the time, a chess program with a computer coach was still superior to a chess program alone. That didn't last long, but I think Kasparov was on to something and I would love to see students paired up with open-source chess programs for tournaments that blend chess-playing proficiency with computer-programming.

Completely agree that Kasparov was on to something with Advanced Chess. The first iteration of the idea failed to gain traction but that was more due to cumbersome tech (players staring at desktops playing slow games, not much of a spectator sport) than to the idea itself. I’m prototyping some ideas in this space with a team of hackers/chess enthusiasts. Ping me if interested in contributing or beta testing, contact info in my profile.

I think it has been shown since then that for modern computer chess engine the human member of the centaur team is always detrimental. It's strictly superior to systematically ignore the human suggestions.

Interesting, I'm not aware of this but conclusively shown, but please provide a link to back that up as I'm certainly curious.

That said, if you do centaur blitz or bullet chess using over the board pieces, not computer move input, I'm fairly certain that the skill of the human player will be a factor :-)

Here's a much better article with a similar title: https://pdfs.semanticscholar.org/c9e7/3fc7ec81458057e6f96de1...

It's always nice to see Kasparov, who could have easily been salty for losing to a machine and given up on chess forever, instead keep in touch with technology and science to make the game better.

He's at the age where you could have expected him to be willfully ignorant on these sorts of things, but he chose not to be. I will respect him for that.

Why would he give up on his job because he lost a match? He made tons of money from that match and he's still making money out of it today.

Also, he is retired from professional chess for a long time now.

He is still active in the chess community. What i meant was he could have become cynical and stopped acknowledging the game altogether now that he is rich, retired, and involved in politics.

Kasparov was the first player to significantly profit from computer chess, selling many many chess computers I fantasised about owning as a child. He has always profited by playing and talking abut computers, often in ways that baffled those in the field:

http://content.time.com/time/subscriber/article/0,33009,9843...

(the 'insightful' pawn sac was suggested by most run of the mill computers, and seemed an obvious moves to strong humans at the time).

Sure, that was like 20+ years ago. I'm saying today, he has no reason to keep writing about chess and yet here we are.

> he has no reason to keep writing about chess

What does that mean?

edit: I mis-read the tone of your message, ignore.

He's "retired". It was just last month he was competing in a Fischer Random event with many of the top players today. [1] And he still regularly provides commentary, above and beyond that of an informal observer, on various games and events.

[1] - https://www.uschesschamps.com/tags/2018-champions-showdown

He is a well known figure in chess. More people recognize his name in connection to chess than Magnus Carlsen and Fabiano Caruana combined! Of course he writes about chess - when he does people listen and assume he has something insightful to say (weather or not he does), thus when he wants money he needs to talk about chess. He can be (and reportedly is) a great inspirational speaker, but he is expected to talk about chess at some time in each speech.

do you really think Kasparov writes about chess when he runs out of money?

It's not really that surprising. The chess world has embraced computers as much as they can be expected to. And Kasparov was world champion, which means people value his opinion about chess.

He is a genious in a broader sense. I am also following him on philosophical and political issues that he expresses his opinion

I'm a beginner violinist who for a long time looked warily at the fact that robot violinists who'll play better than us aren't more than a few decades away.

But I realize nowadays that technology isn't just about automating away what we do as humans - its can also be about removing anything that might be standing between you and your fullest potential.

e.g.

- even if I couldn't find a professional orchestra willing to play with me as soloist, someday a robot orchestra could do that for me.

- a teacher that can spend unlimited amounts of time observing me and guiding me in my learning in the most personalised way

I wonder if artificial learning could recreate the intellect of Socrates.

Are we peak ML yet? I understand the need for hype, but I'm yet to see a killer app.

Not even close. Chess is a relatively easy target for AI: the rules are well-defined, all of the pieces are visible to both players, and while the state space is massive, it's a tractable problem.

Now look at something like DOTA: while each avatar has specific moves, the combination of when you play them, how you chain them, and where you aim them is very unclear; half the time, you've got no idea where your opponents are due to fog of war; and the state space is infinitely larger than chess' (which is already huge to begin with).

In short, we're just getting started. Chess was always just a stepping stone for the AI folks -- a chance to show how a machine could reason. Buckle up -- it's gonna be a wild ride :-)

Reply on news.ycombinator.com