Andrej Karpathy’s trick for getting better answers from AI: Just ramble at it for 10 minutes
Anthropic researcher Andrej Karpathy suggests that users can get better results from AI by providing long, rambling voice inputs. He argues that this 'stream of consciousness' approach helps the model better understand the user's intent and context.
Why it matters
As AI voice agents become more prevalent, users are learning new ways to interact with LLMs to improve output quality.
Andrej Karpathy said LLMs are good at extracting important bits from a long, rambling voice note. Witthaya Prasongsin/Getty Images A leading AI engineer said he likes to ramble his thoughts at LLMs. Andrej Karpathy, a top Anthropic researcher, said he finds "nice, long ramble sessions" useful. It comes as AI labs are releasing updated voice agent models — and hardware to keep humans talking. Andrej Karpathy has some advice for AI users trying to write beautifully crafted prompts. Stop. Instead, you should just ramble. On, and on, and on. In a Tuesday post on X, Karpathy — a leading AI researcher at Anthropic who coined the term "vibe-coding" — said he sometimes turns on voice mode and prattles at a large language model for about 10 minutes. He called the sessions a "total mess, anything goes, full stream of consciousness."
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in