This is all painfully on the nose as far as the mistakes being made today. This paper is going to fill half a dozen citation gaps in my case against AI; I've already written a few times that that even if your LLM automation is *extremely* successful you need a commensurate increase to your L&D budget to avoid catastrophic system collapse later, and it will be great to have such a venerable and comprehensive look at that exact problem. #NBPy