# Setting up your own Articial Intelligence on your own computer.

**What you need:**

1. **A recent computer.** An Apple M1 would be best for speed, but most machines with more than 16Gb of RAM will function OK. More RAM is better. A computer with a recent AMD or NVIDIA GPU is better still. If you don't know what any of that means just ask me: I don't bite.

2. You need **plenty of disk space** for the Large Language Model Files. These can be BIG! 

3. **An account with [Ollama.com](https://ollama.com/)** This is **free**. Just go to their website and create one. Remember your login details.

4. **Patience and confidence** in using commands at a Terminal.

## The Recipe

- Get an Ollama account. Go to [https://ollama.com/](https://ollama.com/) and sign up.

- Open a terminal window. In Windows you:

 *Press Win + R to open the Run dialog.*

 *Type _wt_ and press Enter.*

- Now you need to **install the Ollama software**. Make sure your connected to the Internet, then type (or copy/paset) the following (precisely!), and hit Enter:

 *irm https://ollama.com/install.ps1 | iex*

(For other operating systems, see **[https://ollama.com/download/linux](https://ollama.com/download/linux)** -- which gives the correct command)

- _Wait_. Ollama will install itself.

- Now you need a Large Language Model to use. Let's pick a quick and easy one to install --- a small Phi3 model from Microsoft. The command to install this is:

 *ollama pull phi3:latest*

- Hit Enter, and wait for the download to finish.

- OK, you now have **your very own LLM**! But how to talk to it? That's easy. Type the following command:

 *ollama run phi3:latest*

- Hit Enter. Wait a little while until the LLM cranks up. Then start chatting!

**NOTE:** *This LLM can't access the Internet, so its answers are limited to the data it has at the time it was built. Nevertheless, for most situations it can do the same kind of jobs more powerful AIs can do. It may be slower than you expect. Be patient!*

Bea Groves-McDaniel, 24th May 2026