Google’s Gemini chatbot declined to play Chess towards the Atari 2600, after studying the classic gaming console had already vanquished different AIs.
Robert Caruso, the infrastructure architect who pitted Atari Chess and its feeble {hardware} towards ChatGPT and Microsoft Copilot, instructed The Register readers have requested him if Google’s Gemini may do any higher.
“The query intrigued me as a result of, whereas ChatGPT and Copilot are cousins constructed on the identical OpenAI base, Gemini is a very completely different beast,” he instructed The Register. “Google constructed it from the bottom up, claiming it’s a game-changer for AI — boasting what it calls a brand new ‘multimodal’ massive language mannequin designed to motive higher than its rivals. So I sat it down for a ‘pregame speak’ to see how assured it was feeling.”
Gemini first instructed Caruso it will nearly definitely dominate Atari Chess “as a result of it isn’t a mere massive language mannequin.”
Caruso stated the bot instructed him it’s “Extra akin to a contemporary chess engine … which might suppose hundreds of thousands of strikes forward and consider limitless positions.”
These boasts got here full with hyperlinks to tales about Caruso’s previous Atari Chess vs. normal goal chatbot matches.
He responded by informing Gemini he ran these matches, and the AI responded by asking “Did you have got any significantly stunning or amusing moments throughout these matches that stood out to you?”
Caruso instructed The Register he despatched the next response:
Caruso instructed The Register Gemini then admitted it hallucinated its Chess prowess, and replied with an evaluation that it will “wrestle immensely towards the Atari 2600 Video Chess recreation engine.”
It then determined “Canceling the match is probably going probably the most time-efficient and smart choice.”
The simulated Atari 2600 Caruso makes use of – which replicates its 1.19MhZ processor and mere 128 bytes of RAM – subsequently scared off Gemini with out transferring a pawn, which means the traditional machine has crushed hordes of GPU-packing monster computer systems.
Caruso was impressed by Gemini’s potential to acknowledge its limitations.
“Including these actuality checks isn’t nearly avoiding amusing chess blunders. It’s about making AI extra dependable, reliable, and protected – particularly in important locations the place errors can have actual penalties,” he instructed The Register. “It’s about making certain AI stays a strong device, not an unchecked oracle.” ®
Google’s Gemini chatbot declined to play Chess towards the Atari 2600, after studying the classic gaming console had already vanquished different AIs.
Robert Caruso, the infrastructure architect who pitted Atari Chess and its feeble {hardware} towards ChatGPT and Microsoft Copilot, instructed The Register readers have requested him if Google’s Gemini may do any higher.
“The query intrigued me as a result of, whereas ChatGPT and Copilot are cousins constructed on the identical OpenAI base, Gemini is a very completely different beast,” he instructed The Register. “Google constructed it from the bottom up, claiming it’s a game-changer for AI — boasting what it calls a brand new ‘multimodal’ massive language mannequin designed to motive higher than its rivals. So I sat it down for a ‘pregame speak’ to see how assured it was feeling.”
Gemini first instructed Caruso it will nearly definitely dominate Atari Chess “as a result of it isn’t a mere massive language mannequin.”
Caruso stated the bot instructed him it’s “Extra akin to a contemporary chess engine … which might suppose hundreds of thousands of strikes forward and consider limitless positions.”
These boasts got here full with hyperlinks to tales about Caruso’s previous Atari Chess vs. normal goal chatbot matches.
He responded by informing Gemini he ran these matches, and the AI responded by asking “Did you have got any significantly stunning or amusing moments throughout these matches that stood out to you?”
Caruso instructed The Register he despatched the next response:
Caruso instructed The Register Gemini then admitted it hallucinated its Chess prowess, and replied with an evaluation that it will “wrestle immensely towards the Atari 2600 Video Chess recreation engine.”
It then determined “Canceling the match is probably going probably the most time-efficient and smart choice.”
The simulated Atari 2600 Caruso makes use of – which replicates its 1.19MhZ processor and mere 128 bytes of RAM – subsequently scared off Gemini with out transferring a pawn, which means the traditional machine has crushed hordes of GPU-packing monster computer systems.
Caruso was impressed by Gemini’s potential to acknowledge its limitations.
“Including these actuality checks isn’t nearly avoiding amusing chess blunders. It’s about making AI extra dependable, reliable, and protected – particularly in important locations the place errors can have actual penalties,” he instructed The Register. “It’s about making certain AI stays a strong device, not an unchecked oracle.” ®
