If you're new to AI

Read these two sections first. They will help you understand the current lay of the land, along with what all this new jargon means so that you're not completely lost.

Getting started with code

This is the first part of the "Choosing ML Models" section.

Evaluating LM Results

Why this guide?

Large language and image/video generation models are opening up new ways to create, plan and execute the things we do at work and home every day.

So far, onboarding into the state-of-the-art within AI has been challenging for engineers of any level who are new to the AI space. Current onboarding options require piecing-together various solutions in disjointed ways. Most of these solutions are also not default open source, and there are vast amounts of new terminology, entities and libraries to learn.

As of this moment, the historical pattern of engineers starting and ending projects with primarily open source components is at risk. Foundation models are very expensive to train from scratch, and open source versions have a hard time competing with well resourced for-profit companies. Mozilla is planting a flag in the ground in this space and creating a new hub for open-source AI projects, kickstarting a vibrant open-source community, with clear shared goals and improved open-source visibility for this important new sector.

Why Mozilla?

Mozilla has been developing and shipping AI through its products and initiatives for many years. From Common Voice to Firefox Translate and beyond, we see in AI the potential to empower humans, make technology more accessible, and improve lives.

But like many others, we also see the risk of real-world harm. We don’t have to cast our gazes to some far-flung science fiction future to envision it: people are already experiencing real-world impacts from this technology, right here and now.

Much will depend on how this technology is developed and deployed, but it will also depend on who guides it. We believe that the open source community has a critical role to play in ensuring that AI develops in a way that is responsible, trustworthy, equitable, and safe. We hope that this guide contributes to mobilizing the open source community to join us in this important work.

AI Basics