Model

A model is the trained “brain” of an AI system. Under the hood it’s a huge set of numbers (called parameters) that were tuned, during training, to be very good at predicting what comes next in a sequence of tokens.

A few things worth knowing:

• Models come in sizes. Small ones (a few billion parameters) run on a laptop; huge ones (hundreds of billions) need big data centers.
• A model is a file, not a website. You can download one and run it on your own machine — that’s what “local AI” means.
• A model isn’t an agent. The model just reads tokens and writes tokens. Everything around it — the tools, the loop, the app — is the harness.