It’s 2026 and I still read books — physical, paper, sometimes a Kindle. Not because I’m nostalgic. Because in an AI-summarised world, the people who actually finish long-form arguments have a competitive advantage that compounds quietly while everyone else skims.

What summaries miss

An AI summary of a great book gives you the conclusions. It can’t give you the texture of how the author got there — the side roads, the half-formed ideas, the sentences that change the way you think. The summary tells you what the author concluded; the book teaches you how to think the way the author thinks. That’s a meaningfully different transfer.

The competitive lens

Most knowledge workers in 2026 outsource their reading to GPT or Claude. They walk into meetings having “covered” 30 books and 200 essays. What they don’t have is depth on any of them. The person who actually read three books in the last quarter — slowly, with margins full of notes — runs circles around them in any conversation that goes past the surface. I’ve written about this kind of focus tax before.

What I read this year

Five books that earned their pages, in order of impact:

  • The Beginning of Infinity — David Deutsch. The most useful epistemology framework I’ve encountered.
  • The Hard Thing About Hard Things — Ben Horowitz. Re-read. Different lessons land at different stages.
  • Story — Robert McKee. Not just for screenwriters; the cleanest theory of dramatic structure I’ve found.
  • Antifragile — Taleb. Re-read. The AI era is the most antifragility-relevant moment of our lives.
  • Range — David Epstein. Vindication for multidisciplinary creators with data behind it.

The reading practice

One book at a time. Forty pages a day, five days a week. Notebook open beside it. Write down the sentence that hits, then the question it raises, then your guess at the answer. The note is what makes the reading durable; without notes, the book leaves no residue.

The honest part

Most books aren’t worth finishing. I quit one in three around the 50-page mark with no guilt. The point isn’t to consume more; it’s to find the small number that warrant the depth, and then go deep. Quality time, not quantity finished, is the metric.

The 2026 case for reading

Every advantage I have in conversations, in pitches, in product decisions, traces back to long-form thinking compounded over years. AI flattens shallow knowledge to zero. It can’t flatten depth. Depth is the only moat left.

Got a book that changed how you think? Send it my way.