Well just because a computer can beat us in a chess game doesn’t mean that they are better than us in every other way. Computers can calculate faster than a person, obviously proven many times but we can still out think them in every other way.
No computer can ever be as intuitive as a person.
Also they are not capable as us in being creative or coming up with creative ideas.
So no need to worry about them taking over the world and controlling us.
Like in that old sci fi movie “The Colossus Forbin Project”.
“Ever” is a very long time. It would be more accurate to say that currently AIs are neither intuitive nor creative. There are AIs that have created to do creative tasks but at this time they are very primitive. As for intuition, currently the goal of most* AI research is to produce rational thinking machines, which by definition is not intuitive. Could some future machine have instinct? Maybe, but it certainly isn’t a research focus. For one thing, in terms of computational theory, what does it mean to derive something by intuition? Someday we may get there though with greater understanding of the underlying processes.
I’d say it is extremely close to all, but there’s probably some people working on non-rational thinking because it seems like there’s a niche for just about anything imaginable.
Of course, outside of the narrow walls of academia, most real work on AIs is on making them better at things that humans aren’t good at. After all, we already have humans to do the things that humans are good at. But no human can do what, say, Google does.
You’re correct that speaking very broadly you could look at at the research as a split between making machines that think like a human versus a computational aide to humans. Overall, though there’s probably more research on AI as a computational tool then on machines that think like a human.
Writing it out makes it sound complicated, but it would only take 2-3 seconds with practice. For instance, “modulo 1000” isn’t actually a separate operation; it just means that you never have to do the math to more than 3 digits. And making a mistake occasionally doesn’t really harm anything. The computation could be done in parallel with the physical part, so basically you get the entire round time to do it. If necessary, you can almost double the throughput by taking the middle number as two ternary digits. That is, 0=>RR, 1=>RP, … 8=>SS.
In a real life tournament, even easier, just count something completely random. Number of seconds on the clock. Letters in your opponent’s name modulo 3. Number of strokes in each letter of your current opponent’s name. Number of flies in a certain window pane.
Plenty of random seeds in nature, no need to come up with mental pseudo random generators when random is all around you.
Tournament players exploit some of the same weaknesses a computer does, like predictable play by the opponent, as well as things alien to the computer like physical tells. Real-life tournament results are as a result far from random.
In a tournament, you’re playing 10-20 games in a row against the same opponent, so any method that’s based on your opponent’s name would fail horribly. The number of seconds on the clock won’t work if you have a good rhythm going, because it’ll cycle through according to some pattern of a type that a computer would pick up on. And if you’re playing somewhere where any number of flies on a windowpane modulo 3 is equally likely, and where that number changes stochastically over the time that it takes to play one game of RPS, well, let’s just say that that’s an awful lot of flies.
Chronos is quite correct. But I’m sure Silver Lining will be along soon to tell us that Magnus Carlsen is wrong and that he really could beat them. LOL
[QUOTE=Magnus Carlsen]
Computers are very important for sure. All the analysis I’ve done has been with the help of the computers. You constantly need them.
[/QUOTE]
[QUOTE=Magnus Carlsen]
Yes, to some extent it takes a bit of the mysticism away. But as for chess, we’ve known for a long time that computers are better, so the computer never has been an opponent. It’s a tool to help me analyze and to help me improve at chess.
[/quote]
I just stumbled upon this thread. By coincidence, I decided about two months ago that I wanted to learn chess. I play lots of other games, but I just never happened to learn chess beyond the “I know how the pieces move” level.
To the OP’s original question, it’s a blast. There’s so much to learn, and I enjoy a real “level up” every time I advance in a significant way. I’m using computer analysis to review basically every game I play, so it’s great that computers are so fast and good. The fact that I can’t beat them on their strongest settings is beside the point. I find human opponents more enjoyable mostly because the play feels much more varied and the sorts of puzzles I find myself facing seem different than against an engine.
As glee put it so succinctly upthread:
This certainly holds true for me, even as a beginning player. The rush of besting an opponent or a solving a puzzle is present even if there are harder opponents and harder puzzles out there.
I can certainly understand why chess.com has them listed as a novelty; however, as mentioned previously, you can’t really compare the Elo rating of chess programs and humans anymore since the ratings for the programs come from machine-only pools.
My understanding was that the Komodo rating at chess.com is from the games played there, i.e. the same player pool. If I click on that account’s history, I see a list of matches played. Is that unlikely to be the case?
It must be: Carlsen’s FIDE rating is 2822, so chess.com must be drawing from some other source of data. Vachier-Lagrave’s, meanwhile, is 2789. Kasparov’s is the same at both, 2812, but that might be because he’s retired. And FIDE has someone else in between Kasparov and Caruana, who isn’t listed in chess.com’s top 5.
