Anthropic's chatbot can read full novels

 


Anthropic, an AI startup founded by former OpenAI engineers, says its chatbot Claude can now read and analyze full novels in under a minute. 

The company bumped up the chatbot's context window from 9,000 to 100,000 tokens, which translates to approximately 75,000 words.

  • This means Claude could read and comprehend entire novels such as "A Farewell to Arms," "Frankenstein," and "The Adventures of Tom Sawyer," which range from 69,000 to 74,800 words in length.
  • In a test, Anthropic altered one sentence in "The Great Gatsby" and asked Claude to identify the change, which it did in just 22 seconds.
  • Anthropic says the average person can read 100,000 tokens of text in about five hours, while Claude can do this "in less than a minute."
  • The company says the upgrade will help businesses with tasks that require processing large amounts of information at a time.
  • An AI model's context window refers to the text it takes into account prior to generating additional text, whereas tokens are the fundamental elements of text.
  • For example, the word "fantastic" is divided into the three tokens of "fan," "tas," and "tic."
  • Claude's context window is now larger than any other chatbot. For example, OpenAI's GPT-4 model can process up to 8,000 tokens, while its extended version can handle 32,000 tokens. ChatGPT is currently limited to about 4,000 tokens.

Post a Comment

Previous Next

Contact Form