Again, what's more technically feasible? Setting up a DRM-cracking LLM and manually verifying that it's actually succeeded for each and every DRM scheme you encounter, or just throwing the books in a regular old office scanner and being sure it works without even checking...