In screen time Writes Tech Reporter Rutger Otto Weekly About the Internet. This time he follows a chess tournament for ai models.
It was really exciting in the final of the chess tournament between ai Models. It was between o3 from chatgpt maker openai and grok from elon musk’s ai company xai. Final Score: 4-0 for O3.
The Chess Matches Between the Ai Models Could Be Followed Live on YouTube. You could see the pawns on a virtual chessboard and in a Window on the side you could read how the models determined their next move. “It’s child’s play,” Chess Grandmaster Magnus Carlsen concluded after the final.
I Asked Nathalie van der Lende (from the Royal Dutch Chess Federation) in the run-up to that final if she wanted to watch a few matches. I am More of a Checkers Player Than A Chess Player; That says Enough About My Knowledge of the Game. Van der Lende concluded the Same As Carlsen: “The Level is not very impressive at the moment.”
Accordance to her, the moves of the ai models of look “random and suboptimal”. “It indicates that thesis models are not yet focused on actual playing good chess.”
Carlsen Does Play Chess Against Chatgpt Sometimes. “Sometimes I Get Borde While Traveling,” He’s writes on x with screenshots of a game. The 34-year-old Norwegian Did Not Lose a Single Pawn in 53 moves. Chatgpt Lost Them All.
“The top of the chess world has legs working with specialized chess computers for years,” Says van der Lende. “They are far superior in analyzing moves.” Chess Players Can Learn From That. Those programs are special designed to play chess and are therefore very good at it. Language Models Like Chatgpt Understand The Game And Can Discuss It, But They Are Not Made For Playing Chess At The Top Level.
I do find it interesting that computers can learning from the eldest board games. Years ago, I Enjoyed Following the Development of the Alphago Program, which possible managed to defeat the best the Human Go Player in the World.
With thesis games, The Rules Are Clear, There Is A Goal and Every Move Influeces How the Rest of the Game Proeds. Because there are so many possibilities, almost no game is the same. That makes it so suiteable for training algorithms, I read in an explanation from the university of twente.
For Lee Se-Dol, the Top Player who Lost to Alphago, The Fun was over. Hey possibly stopped as a professional go player, because he could never really be the best again. “Even if I am the best human, there is always an unbeatable entity.”
Even if machines are better at chess than we are, I don’t think there is any reason to stop. Just like I keep running without the pretense of ever winning a marathon. The Russian Chess Legend Garry Kasparov Thinks that we can learn a lot from ai. He said in 2017: “Machines can calculate, we can understand. Machines follow instructions, we have a goal. Machines are objective, we are passionate.”
And if you still want to win from ai, you still have the chance. The Current Chatbots Still Play at A Level Where You Can Beat Them.
49:10