Anthropic was launched in 2021 by former members of OpenAI who were seeking new ways to utilize AI for the public’s benefit. Some of the company members were former researchers on ChatGBT and brought their knowledge to Antropic’s own project, an AI assistant named Claude.
Claude was finally introduced in March after months of a closed alpha test. Since then, Anthropic has been continuously building on the capabilities of the assistant, boasting some pretty impressive functions.
Anthropic now says it has upped Claude’s “context window” so that the AI can now read a full book in just one minute. Claude is capable of reading 100K tokens, units that AI uses to process language, corresponding to about 75,000 words.
“The average person can read 100,000 tokens of text in ~5+ hours, and then they might need substantially longer to digest, remember, and analyze that information,” Anthropic shared. “Claude can now do this in less than a minute.”
The company described a test it ran in which it fed Claude the entire The Great Gatsby manuscript but changed a single line to read that Mr. Carraway was “a software engineer that works on machine learning tooling at Anthropic.” When asked to spot the difference, Claude identified the error in 22 seconds.
While generative AI is still limited in some capacities, the expanded context window of Claude paints a promising picture of AI assistant tools and how they can help decipher and communicate documents.
In other tech news, Teenage Engineering has revealed a TP-7 field recorder.
