{"id":375371,"date":"2025-02-26T06:05:25","date_gmt":"2025-02-26T00:35:25","guid":{"rendered":"https:\/\/dripp.zone\/news\/claude-ai-can-now-play-pokemon-and-its-winning-crypto-news\/"},"modified":"2025-02-26T06:19:53","modified_gmt":"2025-02-26T00:49:53","slug":"claude-ai-can-now-play-pokemon-and-its-winning-crypto-news","status":"publish","type":"post","link":"https:\/\/dripp.zone\/news\/claude-ai-can-now-play-pokemon-and-its-winning-crypto-news\/","title":{"rendered":"Claude AI Can Now Play Pok\u00e9mon\u2014And It\u2019s Winning – Crypto News"},"content":{"rendered":"
\n

It turns out robot lawnmowers<\/a> and ChatGPT are not the only ones that can play video games.<\/p>\n

Anthropic said on Tuesday that Claude\u2019s latest version, 3.7 Sonnet, can play the classic video game Pok\u00e9mon.<\/p>\n

In a thread posted<\/a> to X, Anthropic said an early version of Claude 3.7 Sonnet could defeat opponents within hours of playing Pok\u00e9mon.<\/p>\n

\u201cThe results were striking. Within hours, Claude defeated Brock. Days later, it trounced Misty. Progress that older models had little hope of achieving,\u201d Anthropic wrote. \u201cTurns out extended thinking is super effective.\u201d<\/p>\n

According to Anthropic, Claude 3.7 Sonnet keeps notes in its knowledge base, observes the screen, and employs function calls to click buttons and navigate the game.<\/p>\n

\n
\n

A few researchers at Anthropic have, over the past year, had a part-time obsession with a peculiar problem.<\/p>\n

Can Claude play Pok\u00e9mon?<\/p>\n

A thread: pic.twitter.com\/K8SkNXCxYJ<\/a><\/p>\n

\u2014 Anthropic (@AnthropicAI) February 25, 2025<\/a><\/p>\n<\/blockquote>\n<\/div>\n

In addition to screenshots, Anthropic linked<\/a> to a Twitch channel called \u201cClaudePlaysPokemon\u201d showing Claude playing the game.<\/p>\n

What made defeating the Pok\u00e9mon opponents possible, Anthropic said, was Claude 3.7 Sonnet\u2019s ability to plan its next moves and adapt its strategies, where previous models like Claude 3.5 Sonnet<\/a> would wander or get stuck in a loop.<\/p>\n

\u201cWith a few tools to help it see the screen a bit better, Claude acts as an agent, applying its abilities to a novel task,\u201d Anthropic wrote. \u201cIn this, we start to see glimmers of AI systems that tackle challenges with increasing competence, not just through training but with generalized reasoning.\u201d<\/p>\n

Claude 3.7 Sonnet is the latest AI model to play video games successfully. Last March, researchers used ChatGPT<\/a> to play classic first-person shooter Doom, managing to get to the last room in the game once.<\/p>\n

That same month, Google DeepMind launched its Scalable Instructable Multiworld Agent (SIMA).<\/a> This generalist AI, capable of performing various tasks such as text generation, image analysis, and translation, was trained to play video games such as No Man\u2019s Sky<\/i>, Teardown<\/i>, and Valheim.<\/i><\/p>\n

\u201cOur AI agent doesn\u2019t need access to a game’s source code, nor bespoke APIs,\u201d Google DeepMind wrote. \u201cIt requires just two inputs: the images on screen and simple, natural-language instructions provided by the user.\u201d<\/p>\n

Edited by Sebastian Sinclair<\/a><\/i><\/p>\n

\n
<\/p>\n

GG Newsletter<\/h3>\n

Get the latest web3 gaming news, hear directly from gaming studios and influencers covering the space, and receive power-ups from our partners.<\/p>\n<\/div>\n<\/div>\n<\/div>\n