Large language models (LLMs) are incredibly powerful general reasoning tools that are useful in a wide range of situations. But working with LLMs presents challenges that are different from building traditional software: Calls tend to be slow, and streams generate output as it becomes available. Read this story
