The first is what I think of as awareness reading. You're scrolling X, skimming headlines, flipping through books at the bookstore. You're not trying to remember everything—you're trying to maintain a sense of what's happening in your field. It's like monitoring radar. You need to know what's on the horizon without studying every blip in detail.

I... See more

The second type is information retrieval. This is what we do with Google, though we don't usually call it reading. You need a specific fact, a particular technique, an answer to a concrete question. When I look up how token architecture works or search for a Python library, I'm doing retrieval reading.

I think LLMs are essentially retrieval reading... See more

The third type is what I'm calling connectome reading, though I'm not entirely satisfied with the term. This is reading that literally rewires your brain—forming new neural pathways, pruning old ones, building new patterns of thought. It's metabolically expensive. Research suggests deep reading burns significant glucose because your brain is... See more

The Technical Practitioner

This is for engineers who need to build real systems. I've watched too many developers try to read Deep Learning cover to cover and burn out. That's not how technical learning works.

These books become your reference library. You're building something, you hit a problem, you look up the solution. AI becomes your pair... See more

The Informed Professional

This one's close to my heart because it's for people like product managers and designers who work with AI but won't write training loops. You need to know what's possible, what's dangerous, what's hype.

The balance here is tricky. You need enough technical understanding to push back on engineering ("this seems like it... See more

Before I start a book, I tell AI what I'm trying to accomplish. Not just "summarize this book" but "I'm a founder trying to understand if LLMs are right for my product. What should I focus on?" (And yes, I do sometimes read without AI entirely, and I think that’s a good practice too!)

For deep reading, I use AI like a reading companion. After each chapter: "Help me think about this. What questions should this raise? How does this connect to what I read last week?" I'm not asking for summaries—I'm asking for thought partnership.

I read hungrily, in whatever scraps of time I can steal between writing, making videos, and taking care of the kids. I'm like Danny Ocean in Ocean's 11—I'm looking to raid and read with stealth and style.