Search whole site: site:twitter.com
Search title: Ethan Caballero on Twitter: "Whisper is how OpenAI is getting the many Trillions of English text tokens that are needed to train compute optimal (chinchilla scaling law) GPT-4." / Twitter
See how to search.