iStockphoto
Artificial intelligence platforms like ChatGPT have been positioned as The Next Big Thing thanks to their ability to tackle complex tasks faster and better than many humans. However, there’s still some work to be done when it comes to mastering the art of chess based on how it fared while playing a match against a computer opponent on an outdated video game console.
Atari didn’t actually release the first video game console designed for residential use (the Magnavox Odyssey laid claim to that title upon its launch in 1973), but the Japanese company is still widely credited for leading the charge thanks to the success of the home version of Pong that came out in 1975.
Two years later, the developer cemented itself as the brand to beat with the release of the Atari 2600, which became a smash hit with the help of in-house games like Space Invaders and Asteroids as well as licensed titles including Pac-Man and Frogger.
The chess game that was released for the console may not have been a blockbuster in comparison, but it was still a more than serviceable option for players who were looking to get their fix against a computer opponent that could be set to one of eight difficulty levels.
A computer engineer named Robert Caruso recently decided to see how ChatGPT would fare in a chess match where its virtual foe was operating on the beginner level where it is only able to plan a couple of moves ahead, and according to what he had to say in a LinkedIn post chronicling the experiment, the answer is “Not great!”
A computer first beat a chess grandmaster in a game in 1988 (IBM’s famed “Deep Blue” toppled reigning world champion Garry Kasparov in a match close to a decade later), and based on the amount of processing power at its disposal, you might think ChatGPT wouldn’t have much trouble handling an opponent that was designed to be beat fairly easily on a video game console that was released nearly 50 years ago.
However, Caruso says the showdown (which he ran on an emulator) was firmly one-sided in the opposite direction, noting ChatGPT “made enough blunders to get laughed out of a 3rd grade chess club” and failed to improve after he attempted to accommodate the needs of an A.I. chatbot that claimed it was struggling to identify icons that were “too abstract to recognize.”
Caruso noted he spent 90 minutes attempting to stop ChatGPT from opting for “awful moves” while attempting to steer it in the right direction without much success before it conceded defeat.
The best part? ChatGPT was responsible for floating the idea of the match in the first place, as Caruso added he was using it to discuss the history of A.I. in chess before it suggested the idea of going up against the Atari opponent.
You hate to see it.
Content shared from brobible.com.