Hi, I’m Nhut

LLMs Don’t Actually “Call” Tools Here is a fact that surprises most people: LLMs cannot run code. They cannot call APIs. They cannot read files. All they do is predict the next token — they output raw text, one piece at a time. So how do they “use tools”? The trick is simple. During training, models learn to output text in a special format when they want to use a tool. For example, instead of guessing an answer, a model might output something like: ...

This is my first post. More to come soon. I started this blog to share what I learn about software engineering, AI, and whatever else I find interesting. Writing helps me think more clearly, and if it helps someone else along the way, even better. Stay tuned!