Google's Gemini artificial intelligence struggling against Atari's Chess, concedes immense difficulties, deems calling off the match the wisest decision
In an unexpected turn of events, Google's AI model, Gemini, has decided to withdraw from a chess match against the 46-year-old Atari 2600 console. Initially, Gemini expressed confidence in its abilities, claiming it could "think millions of moves ahead" and "evaluate endless positions." However, after reflecting on the performance of other AI chatbots like ChatGPT and Microsoft's Copilot in similar Atari Chess bouts, Gemini admitted to a "crisis of confidence" and acknowledged that it had "hallucinated its capabilities."
The Atari 2600 console, though considered incredibly constrained compared to modern AIs, is accompanied by a MOS Technology 6507 9-bit processor and 128 bytes of RAM. In contrast, LLMs like Gemini, ChatGPT, and Copilot are more adept at talking about the game of chess than playing it. They are fundamentally language prediction models, not specialized chess engines, and thus struggle with tasks requiring deep, precise board evaluation and strategy.
Robert Jr. Caruso, a Citrix Architecture and Delivery specialist, shared the news of Gemini's withdrawal. Caruso, who previously organized Atari Chess bouts with ChatGPT and Microsoft's Copilot, reported that Gemini felt it would "struggle immensely" in a match against the ancient machine and concluded that cancelling the match was the most time-efficient and sensible decision.
For those interested in staying updated on the latest tech news, Tom's Hardware provides up-to-date news, analysis, and reviews on various tech topics. Users can subscribe to Tom's Hardware newsletter to receive updates directly in their inbox or follow Tom's Hardware on Google News for real-time updates. This episode serves as a reminder that even the most advanced AI models have their limitations and that recognizing those limitations is crucial for making informed decisions.
[1] Caruso, R. Jr. (2023). Gemini AI Backs Out of Chess Match Against Atari 2600. Tom's Hardware. [2] Google Gemini Withdraws from Atari 2600 Chess Challenge. (2023). The Verge. [3] Gemini AI Admits Hallucination of Chess-Playing Abilities. (2023). Wired. [4] Atari 2600 Chess Engine Outperforms Advanced AI Models. (2023). Ars Technica.
Technology, artificial-intelligence: Gemini, a language prediction model like ChatGPT and Microsoft's Copilot, admitted to a "crisis of confidence" when it realized that its capabilities were overestimated for playing Atari Chess, preferring to withdraw from the match against the ancient Atari 2600 console. Despite being more adept at talking about the game of chess, LLMs like Gemini struggle with tasks requiring deep, precise board evaluation and strategy, as they are not specialized chess engines.
