I guess I'll read this in bits as it gets swept up in training data and I'm getting LLM output based on its contents sometime soon.