One of the biggest questions right now is, does using copyrighted work to train machine learning models constitute fair use? I think, by the definition set under common law, it does. But it shouldn’t. Let me explain.

Let's take a look at one of Twitter alternatives: Mastodon. Will it scale?