Round one had been a basic lesson in language models, using this year’s congressional legislation as our collection of open-source language. Our audience members, each one skilled in coding already, heard some quick speeches describing the mechanisms of LLMs, and an exposé of the public policy research landscape.

Our main aim had been to share sample code in Python (runnable in the cloud), making use of LangChain to access OpenAI’s models. Our example builds up a Question & Answer program based on legislation produced by the US Congress in 2023.

This round, our organizers plan to bring in more speakers, to discuss some of the ethical quandries that can be caused by relying too much on ML models. This is going to also be a chance to look beyond federal legislation, and to bring in more sources from regional or non-legislative legal sources.