ChatGPT will get ‘wrecked’ by a easy 1977 Atari chess program

Regardless of ever-growing curiosity in AI instruments and assistants, it’s value remembering that they’re nonetheless fairly restricted with quite a few shortcomings. They’re not as good as they may appear on the floor. Living proof, ChatGPT is fairly ineffective in the case of enjoying chess.

As reported by Futurism, ChatGPT misplaced a chess sport in opposition to the traditional Atari 2600 gaming console. Robert Caruso, an engineer at Citrix, organised the sport between the AI and a easy 1977 chess program launched for the Atari 2600.

Throughout the sport, ChatGPT made a collection of embarrassing errors, misinterpret strikes, and stored shedding monitor of its personal items. “ChatGPT obtained completely wrecked on the newbie stage,” Caruso wrote. In the long run, the AI chatbot merely gave up and misplaced.

The fiasco is a crucial reminder that LLMs—even the “reasoning” ones—are nonetheless simply language prediction fashions at their core. It’s clear proof that devoted instruments which can be educated or coded for a particular objective, just like the Atari 2600’s chess program, are nonetheless higher than AI assistants which can be purported to do “every part.” This must be an essential lesson for tech corporations like OpenAI and customers who depend on AI instruments.