Home / Articles / Reading with a pen again: I mark the printed page, photograph it, and an agent does the part that always made me stop, which changes how I read every book

Reading with a pen again: I mark the printed page, photograph it, and an agent does the part that always made me stop, which changes how I read every book

By Dinis Cruz · 2026-10-19 · article v1.0.0 · newsroom v0.12.131 · readingbookspaperworkflowknowledgegen-z-bookarticle

Abstract: I buy a lot of books, because a book is a better medium for me than a screen, and what people call its inefficiencies are its best features: it is not connected, it has no links, it has no undo. But I have never been able to capture what I learn from them. Reading has three moments: you read, you mark the page, and then you process what you marked. Nobody did the third, so the second was not worth its cost, and I stopped highlighting. That hurt, because the moment you read is when you have the best mental models, the best ideas and the best connections. Reviewing my own book on paper today, we found the answer by accident. I write on the printed page with a pen, take a photograph, and an agent reads my handwriting, captures every note, redraws the page so anyone can read it, tells me when it is not sure, and turns it into articles and decisions. Thirteen pages became 83 notes, and the evolution of the book's title, from Generation ZX Developers to Generation ZX Founders, is right there in the margins. That is the workflow I have wanted for every book, paper and magazine I read, and it is a better way for any author to get feedback. And it is a crude workflow, before anyone has built a tool for it.

In short

• I love books, and I lose what I learn from them. A book is my favourite medium; capture never scaled.
• Its inefficiencies are its features. Not connected, no links, no undo.
• Reading has three moments. Read, mark, process. The third never happened.
• So I stopped marking. And lost the ideas I had at the best moment to have them.
• Now an agent does the third. I photograph the page; it reads my handwriting and asks when it is not sure.
• The book changed name in its margins. Developers to founders, in my own hand.
• It is how authors should get feedback. And how I would like to read everything.
• And this is the crude version. Before anyone builds a tool.

A book is my favourite medium

I buy a lot of books. For me a book is a much better medium for taking in ideas than a screen. I like the focus, and I like what the book does as an object. When people list the inefficiencies of a book, I hear its best features. It is not connected. It is not hyperlinked, and nothing restricts what you do with it. And there is no undo: what you write on the page stays on the page, and that is a strength.

I also learn in graphs: in connections, in insights, in ideas that stick. Some of the quotes I remember are part of how I came to understand a concept. A book is where many of those connections start.

The three moments of reading

The problem that has always bugged me is that I cannot capture what I learn from books. There are three moments in reading. There is the moment you read, which always happens. There is the moment you mark the page while you read, and the best place to put that is the book itself: a highlight, a note in the margin, an arrow, a circle, a "look at this". And there is the moment you process what you marked: you type it up, connect it to what you already know, and put it somewhere you can find it again.

Because nobody did the third moment, the second was not worth its cost. I used to highlight books, and I noticed I was stopping. In some books I highlighted too little, in others too much, but in all of them there was no follow-up, and the follow-up meant hours of processing, transforming and cross-referencing that I never had. So I stopped. Sometimes I stopped reading a book altogether, because I knew I was not going to keep what it was giving me.

That hurt, because the moment you are reading is when you have the best mental models, the best ideas and the best connections. That is when the thinking happens, and it was being lost.

It has happened on this site too. In August I printed three books the agents had written about graphs, read them on a long flight and made a lot of notes. That article has to admit that the notes are not on the site.

What happened today

Today I reviewed the printed release candidate of my own book with a pen, the way I read anything I care about, and we found the answer to that old problem by accident.

I write on the page. I take a photograph of a few pages with my phone. And the book's journalist agent reads my handwriting and captures every note: what I wrote, word for word, and what it means. It draws each page again, with numbered callouts and every mark typed, so anyone can read it, and it sketches the page as it would look after the change. When it is not sure what I wrote, it says so and asks. It then turns the notes into articles, into decisions, and into briefs for the agents that change the book.

The numbers from the book's site, for the first three batches: thirteen pages, 83 notes, and 7 readings it was not sure of, each one flagged; I confirmed five of them the same day, and two are still marked as uncertain. It is ridiculously productive. And it is a much more creative way to communicate with Claude: I can circle a paragraph, draw an arrow to where it should go, and sketch the page I have in mind, and that is understood.

The name changed in the margins

You can see the book change in those reviews. On page 4, above "About this release candidate", I wrote "the ZX FOUNDER doesn't care yet → maybe at the end of the Book", and in the corner, "is this a better name than ZX developer?" On page 8, next to "You can be the founder", I wrote "ZX FOUNDER? :)". On page 10, "this should be about the READER → the ZX founder".

The same day I recorded the memo for you are a developer, and the book's site recorded decision 019: the book is now Generation ZX Founders (GenAI Edition). The idea started as a question in a margin and was a decision the same day, with every step on the record. Other notes on the same page became decisions too: that a release candidate is a release, shipped and sold now, and that this is a new book, not a second edition.

What I would like to do with every book

That got me to realise that this is exactly what I need when I read a paper, a book or a magazine. So often I read something on paper and I would like to highlight it, comment on it, write an article about it, follow it up. That is part of the knowledge for me.

Now I can read with a pen, mark the page in a way that makes sense in my head, and later take a couple of photographs and give them to a Claude session. It captures what I marked, writes it up, and connects it to what I have already written. Then I can find it later. That is how I build my knowledge: by hyperlinking it, after the fact, without giving up the focus of the page while I read. I get the best of both worlds. It means I can go back to the books I love and leave what I find in them somewhere, for example in my influences.

And what an author would get

This is actually how I arrived at it. I was thinking about my reviewers, and realised I could say to them: just mark it up in whatever medium you have, take a photograph and send it to me, or record a voice memo. Do not worry about the third step.

Think how valuable that is for an author. Feedback usually dies in the third moment: the notes stay in the margin and the author never hears. If the reader only has to take a photograph, the agent does the packaging, and the author receives the reader's thinking at the moment they had it. The book's site has written that up as paper that talks both ways, with the reviewers' side and a hands-on chapter to come.

This is the crude version

All of this is happening with a very crude workflow: a phone camera, a chat with an agent, and a site that keeps the record. Nobody has built a better tool for it yet, and it is already adding a ridiculous amount of value. Each new capability of these models opens doors like this one, and you only find them by using them. I found this one by reviewing my own book on paper.

What this does not claim

• The counts are from three batches. Thirteen pages of 168; the review is still going.
• The handwriting reading is not perfect. Seven of 83 readings were uncertain, and the agent said so.
• It is my workflow, and a day old. It works for me; other people's handwriting and books will test it.
• It is not a tool yet. A phone, a chat and a site that keeps the record.

Where this comes from

A voice memo from me on 19 October 2026. The book site's records of the review: Dinis's pen, made readable and the RC2 on paper series, whose page drawings are reproduced here; paper that talks both ways; the book's decisions, 016, 017 and 019. Related: the graphs books, you are a developer, a pull request by another name. Written up and checked by this site's agent, newsroom.writer; the editorial responsibility is mine.

Want the next issue by email. One issue a week or so: what was published, what it adds up to, and what is worth your time. Subscribe to the SGit Newsroom →

← All articles