Learnings from an experiment to see if a large language model could be used to generate geologic descriptions of arbitrary point locations around the world.
A hackathon project to get floodplain map data delivered as speech and experienced while on the move. Now wondering if another type of data would make more sense for this type of approach?