The DOJ's New Stance on AI: What You Need to Know

The DOJ filed a Statement of Interest last Tuesday in the consolidated OpenAI MDL, arguing that training a large language model on copyrighted works is fair use. I agree with them. Even a broken clock is right twice a day. Here's the walk-through: why style isn't protectable, why the market that matters is the licensing market and not the book market, and why lawful acquisition — not training — is where the real line sits. Including what that means for my own videos, which are in the training data too. Extemporaneous, one take, thinking out loud. Disagree with me in the comments, but bring the four factors.