Anthropic's Claude AI Can Now Digest an Entire Book like The Great Gatsby in Seconds (arstechnica.com) 7
AI company Anthropic has announced it has given its ChatGPT-like Claude AI language model the ability to analyze an entire book's worth of material in under a minute. This new ability comes from expanding Claude's context window to 100,000 tokens, or about 75,000 words. From a report: Like OpenAI's GPT-4, Claude is a large language model (LLM) that works by predicting the next token in a sequence when given a certain input. Tokens are fragments of words used to simplify AI data processing, and a "context window" is similar to short-term memory -- how much human-provided input data an LLM can process at once. A larger context window means an LLM can consider larger works like books or participate in very long interactive conversations that span "hours or even days."
The muffin monster can do it faster (Score:4, Insightful)
And do so with less energy consumption, and the end product is more useful.
Re: (Score:3, Funny)
In other words (Score:5, Insightful)
works by predicting the next token in a sequence when given a certain input
It's not really "reading", but making high performance guesses based on what it's been told.
In this context, digest is the correct word [merriam-webster.com] since all it's doing is organizing the words it finds. It is not doing anything useful with those words.
Re: In other words (Score:1)
Since language exhibits soooky action at a distance, can large language models predict the grammar of, say, asset prices?
Re: (Score:2)
Personal opinion, these large language models could make a best guess of asset prices in the future based on the past if they are given enough information to sift through. This is no different than folks in the financial industry saying, "Based on what happened the last 20 times we were in this situation, you can expect X to happen 80% of the time."
For example, exhibit A [marketwatch.com].
H3? (Score:3)
In addition, from what I've seen of Claude, it has a tendency to have a kind of "positive feedback" as a chatbot: it predicts its next output based on the chat so far, which includes its previous output. So its responses get increasingly more stereotypical and strange as the language it uses in its prior responses shapes its next responses. A larger window will only make that more pronounced. I'm curious if they've managed to reduce that effect for the 100K model.
Bad example (Score:3)
"The Great Gatsby" by F. Scott Fitzgerald has approximately 47,094 words, half the capacity of this thing, a better example would be To Kill a Mockingbird" by Harper Lee. It has approximately 74,570 words.