Listen

Description

This podcast explains what happens when 66,000 research papers are used as training data