In a world where AI seems to be cropping up in every app and on every device, it's nice to read a news post where it's absolutely trounced at something by a games console from the 1970s. It might be able to make a picture of you looking like a viking warrior or tell you how to write an essay on silica sand without you having to do any research (can you tell I don't like AI?), but when it comes to playing chess against a retro Grand Master, it doesn't stand a chance.
Robert Caruso, a Citrix software engineer has taken to LinkedIn to give his account of how ChatGPT got, in his words, 'absolutely wrecked' by an Atari 2600 console from 1977.
For regular readers of Retro Dodo, the Atari 2600 is a name you'll be very familiar with. If you're new to retro gaming, however, then this machine was one of the most pivotal home consoles from the second generation of console gaming. It sold over 30 million units worldwide, 27 million more than its next competitor the Intellivision, which coincidentally Atari now owns.
You would think that something as advanced as ChatGPT would be able to take an old timer like the 2600 for a walk in the park and beat it at chess while helping someone from Brussels get cheaper broadband at the same time, but no. The Atari 2600 managed to beat it on the easiest level, with the Atari only ever thinking 1 to 2 moves ahead.
Our review of the new Atari 2600+
'It started as a simple experiment: pit ChatGPT against the Atari 2600’s chess engine,' says Caruso, 'I figured it would be a lighthearted stroll down retro memory lane. ChatGPT got absolutely wrecked on the beginner level. This was after a conversation we had regarding the history of AI in Chess which led to it volunteering to play Atari Chess.'
Caruso goes on to say that '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'
In the end, the 8-bit behemoth just did what it's done for decades, quietly making moves without any flash or a language model. All it had was the level of difficulty that developers put into games back then. Come to think of it, I'd like to see ChatGPT have a go at Zelda II!
IT took 90 minutes for ChatGPT to admit defeat. Hopefully it will go back and tell its other Skynet pals that while they think they're the most powerful beings in the universe, the Earth's retro consoles are still in control.
