> We are also open sourcing experiments, models, and research data to explore the strengths and weaknesses (e.g. potential unintended biases) of using machine l

> We are also open sourcing experiments, models, and research data to explore the strengths and weaknesses (e.g. potential unintended biases) of using machine learning as a tool for online discussion. Does that mean that practically everything to be able to run it on a server without any google API is possible (once those things release)?