The four pieces of the local AI landscape every founder should know

There are four core pieces to the local AI landscape. First, the model — the brain file. Examples include Gemma, Llama, Mistral, Qwen, and Phi. These are model families, and they differ meaningfully: some are better at reasoning, some at coding, some are smaller and faster, some handle images, and some are easier to run on your own machine. Second, the warehouse — where you find the model. Hugging Face is the biggest and the first place to go. Third, the software — what actually runs the model on your hardware, such as LM Studio or Ollama. Fourth, the workflow — the product you're building around all of it. Those four pieces — model, warehouse, software, workflow — are the full map of the space.