Given the current data sovereignty concerns, and risks of relying on loss making vendors, I want to teach you that you can use your existing hardware to try out cutting edge models. This is a stark contrast to the current status quo of sending over all your data to foreign LLM inference companies, where you've no control over their software stack and can't improve it.

I also want you to see that you can contribute to making these tools better for everyone. I hope that, by the end of the session, you will publish you Pi extensions under an open source license so that other people can get the benefit of your work!