I haven't read Good to Great. Here's how I'm going to understand it.

I haven't read Good to Great by Jim Collins.

The book came out in 2001, studying 11 companies that transformed from "good" to "great" between 1965 and 1995. Pre-internet. Pre-cloud. Pre-AI entirely. It's been cited millions of times and appears on nearly every "must-read business books" list.

And I don't intend to read it the traditional way.

Instead, I'm trying something else: dialogue with AI section by section, push back, distill into blog posts, then listen back as a podcast.

This isn't lazy reading. This is an experiment — and a question I think many people are curious about but no one has done seriously: can you actually understand a classic book deeply by having a dialogue with AI instead of reading linearly?

The problem with traditional reading

I'm not against reading. I've read a lot. But the more I read, the more I see three problems that are increasingly serious in the AI era.

One is the context lag problem. Good to Great was written 25 years ago. The context has changed fundamentally. Linear reading forces me to absorb each chapter through Collins's logic — his arc, his examples, his conclusions. What I want isn't "what did Collins think in 2001" but "which parts of that thinking still hold in 2026." Reading straight from page one doesn't answer that question.

Two is anchoring bias. The human brain tends to agree with the author when reading alone. The longer you read, the more you absorb — by the time you put the book down, you think you're thinking, but you're really just paraphrasing the author in your own words. Having a debate partner (AI) helps me hold a critical stance. AI doesn't care that Collins is "a classic." It just answers the questions I pose.

Three is input without output. Reading and then forgetting is the biggest trap of reading. You highlight, underline, nod along — then six months later can't recall three main ideas. The reason: you never had to produce anything from the knowledge. This series forces me to distill, write, and publish — the only way to truly internalize.

The 5-step method

For each chapter (or module composed of several short chapters):

Step 1 — Load the content. I ask AI to summarize the chapter: main ideas, Collins's examples, core arguments. Not to skip reading — to have a baseline before debating.

Step 2 — Brain dump. Before diving into deep dialogue, I write down three things: what's persuasive, what's doubtful, what I want to dig into. This step matters because it forces me to take a position before the AI "primes" my thinking. Skipping this step skips the whole experiment.

Step 3 — Debate with AI. I ask AI to steelman both sides — not just defend Collins. I ask it to point out survivorship bias, selection bias, academic critiques. I ask it to connect to the 2026 AI-era context.

Step 4 — Distill into a blog post. I write the draft, AI reviews. Each post has a fixed structure: what Collins said (short), critique through the AI-era lens (main body), reframe for AI-era companies, methodology notes.

Step 5 — Audio. Convert the blog post into audio to listen back while commuting or exercising. Repetition is the key to internalization.

This isn't lazy reading

I know the first reaction from some people will be: "You haven't read it but you dare to write about the book?"

I'm not claiming to understand Collins as fully as someone who's read him ten times. I'm claiming something different: in an era with too many books and too little time, we need a new way to engage with knowledge — not to replace reading, but to supplement it.

Linearly reading a 300-page book takes 15–20 hours. After that you know the content but haven't necessarily thought about it. This series takes a few days of focused work with AI — faster than reading straight through — and the output is 9 blog posts, a podcast series, and a thinking framework that has been put through real critique. The trade-off is clear.

This is also not "just ask AI, no need for books." That position is bad and incorrect. AI doesn't know what isn't in its training data. For very recent books, or books that need to be read in the original to feel the writing, this method doesn't work. I chose Good to Great because it's old (AI has enough data), famous (training data is thick), and it's a framework book (clear structure — no need to read the original to feel it).

Transparency commitment

Because I'm not reading the original, I depend on AI summarizing Collins correctly. There's a risk AI distorts or emphasizes the wrong things. I'll be transparent:

• Every Collins summary will cite the chapter
• At the end of each post, there's a "Methodology Notes" section — I admit where I trusted AI without verifying
• If you've read the original and see me (or AI) misunderstanding, please correct me

This transparency isn't an apology in advance. This is what honest reading looks like in the AI era — not pretending to have read it carefully.