An underappreciated fact about large language models (LLMs) is that they produce “live” answers to prompts. You prompt them and they start talking in response, and they talk until they’re done. The result is like asking a person a question and getting a monologue back in which they improv their answer sentence by sentence. This […]