Digital publishing DRM for LLM integration is the next wedge

Whoever solves digital publishing DRM for LLM integration is going to make a lot of money.

Now for those that know me well, this is probably a surprising statement. I have been somewhat notoriously skeptical of DRM because usually it means "bad customer experience with no real upside to IP protection". But hear me out.

I'm using several different ebook providers in the course of my MBA that are positioned as a "solution" for the fact that a lot of college students are both poor and technically savvy, so even a well-protected ebook in standard formats quickly gets traded around. They've been clunky, feature poor, anti-user, and generally unusable messes. But they barely staggered over the finish line of "good enough", so they decided "good enough".

That won't work for much longer. As LLM context windows have grown, and foundation model providers have integrated workspaces, document uploads, and custom knowledge bases, students are going to begin demanding access to their books through the filter of an LLM agent. Automated notecards. Custom lessons. Mastery-based learning. There's a hundred out-of-the-box use cases here and handing the students a personal tutor who knows the EXACT material the class is covering isn't going to be an innovation--it'll be table stakes a year from now.

I'm sure the academic ereader providers will eventually scrape up a half-baked solution here. It will be slow, it will be awkward, it will be wasted effort, it will be non-native to the student's preferred GenAI interface. Mostly, it will be too late. Given that they still struggle to break text across pages sized to my screen, which Kindle has been doing without problem since 2007, I'm not confident we'll see meaningful first-party improvement in the next decade.

If they continue to stand in the way of forward progress, the rest of the world won't wait. They'll find themselves with the same problem the record labels and television studios had before streaming. I'm reminded of the quote by Valve founder Gabe Newell: "We think there is a fundamental misconception about piracy. Piracy is almost always a service problem and not a pricing problem."

This is a huge problem, and a huge opportunity, which leaves the only real solution in the hands of whoever manages a universal application of the tech and the relationships and the distribution to let me query my books in ChatGPT, create on-the-go study podcasts in NotebookLM, and write those final Anki flashcards with Claude Cowork.