Back when this profession of ours was trying to find its place in the world, some smart people who thought the state of computing was shit got together and started to make some noise. The L0pht fellas testified in front of the US Congress, the cDc folks released tools that not only mocked Microsoft but pretty much made the company do something to at least save face, and hacker conferences were all the rage. Those were the days of smashing stacks for fun and profit, and the days some people realized that powerful technology companies were putting us all at risk for profit and no fun.
Some of those folks went on to start their own cybersecurity companies, some went on to work for governments, or work for the very companies they were ridiculing not too long ago. Some have tried to keep the flame alive and bring back that vibe. And, somewhere along the way, shit got weird.

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.