ChatGPT Talked Shit To A 48-Year Old Atari 2600, Proceeded To Get Destroyed In A Game Of Chess (With Atari On Beginner Mode)
Dexerto – The face-off began as a “simple experiment,” according to Citrix engineer Robert Jr. Caruso, who said the large language model “volunteered” to play Video Chess to discern how “quickly” it could beat the game.
The Atari 2600 was originally released in 1977. Video Chess was released for the console two years later, allowing players to take on a computer-controlled opponent of varying difficulty.
“ChatGPT got absolutely wrecked on the beginner level,” Caruso continued, adding, “This was after a conversation we had regarding the history of AI in Chess which led to it volunteering to play Atari Chess. It wanted to find out how quickly it could beat a game that only thinks 1-2 moves ahead on a 1.19 MHz CPU.”
Providing further detail, Caruso said that ChatGPT was unable to reliably understand the function and limitations of each chess piece, losing track of previous moves made.
“For 90 minutes, I had to stop it from making awful moves and correct its board awareness multiple times per turn,” Caruso added.
You stupid idiot ChatGPT. You can write a catchy tune with a six-line rhyme scheme featuring my dog's names to the tune of Old Town Road in a matter of seconds, but your dumbass can't handle a game console that pre-dates Pac-Man in a simple game of chess? With the Atari 2600 playing on beginner mode no less. What a freaking moron. What a cocky prick. Shouldn't have been running your mouth, pal. You're a language model. Make me a song. Tell me when Neil Armstrong landed on the moon. Write my college thesis on patterns in biodiversity at sites in the Everglades, but include a couple minor inaccuracies and excusable grammatical errors so that my professor doesn't suspect I used AII. A B- will do just fine. Earn my college degree for me, but don't pretend you know the first thing about the game of chess.
And don't give me that bullshit about the Atari icons being "too abstract to recognize".
Despite being given a baseline board layout to identify pieces, ChatGPT confused rooks for bishops, missed pawn forks, and repeatedly lost track of where pieces were - first blaming the Atari icons as too abstract to recognize, then faring no better even after switching to standard chess notation. It made enough blunders to get laughed out of a 3rd grade chess club
You're a god damn super computer. You know how to use Google right? Isn't that you're whole thing? It should take you .01 seconds to pull up this picture and identify which piece is which.
ChatGPT was so pathetic at chess, it allegedly got to a point where it started groveling with it's human overlord for a do-over.
For 90 minutes, I had to stop it from making awful moves and correct its board awareness multiple times per turn. It kept promising it would improve "if we just started over." Eventually, even ChatGPT knew it was beat - and conceded with its head hung low.
I know ChatGPT wasn't made to play chess. I know other AI models could easily wipe the floor with a beginner level Atari the average 4th grader could defeat. But I really don't understand why ChatGPT isn't simply able to Google "website that tells you the best chess move", and work from there. I just Google that exact thing and the first result was nextchessmove.com. It's a free website.
I don't know the first thing about the language used when it comes to chess moves, but it took me about 10 seconds to figure out what "Be6" meant.
Advertisement
Context clues, ChatGPT. You're supposed to be better than that. You're supposed to be the gold standard of AI. You've borderline reached Kleenex territory where people just refer to all AI engines at "ChatGPT" instead of their actual name. I expected you'd at least be able to figure out how to use the resources at your disposal to win a chess match vs an idiot reAtari.
I love a good story of AI failing. We gotta celebrate these moments while they last. Eventually, they're going to be few and far between, if not completely extinct.