With my mom and younger sister I’m visiting the Yucatan Peninsula for the first time. My wife Z is joining us partway through the trip.

I’m impressed by the Tren Maya. Not by how well it works but by the mere fact it exists. It’s not the kind of thing I would expect to see in Mexico. Admittedly, it’s no bullet train. From Cancun to Tulum it takes an hour forty, the same travel time by car. And it only runs a few times per day. But it only opened last year and it connects the Mexican southeast. We’ve mentioned it to servers, cab drivers, desk clerks, but no one we’ve spoken to has taken it. On this one today there are a few dozen passengers but what strikes me is the number of workers.

To enter the platform, we show our printed tickets and identification before progressing to automated gates where a lady instructs us to scan our QR code in the typical place at the front of the gate. It beeps and opens to let us through. A step past it on the other side another lady motions us forward toward the metal detector a few feet away and tells us to line up on the right side. Four or five uniformed officers stand at the one metal detector, feeding in bags and baskets into it on the conveyor belt on one side and taking them off on the other side. Once seated on the train a man comes down the aisle asking for tickets but when I mutter I don’t know where I put it he just confirms our destination before continuing down the aisle. At each station we pass uniformed officers of the Guardia Nacional stand and traipse around while several cleaning ladies sweep and mop already tidylooking floors.

This morning when we came out onto the train platform a worker greeted us and queried about our assigned train car. We told him and he led us to a particular spot on the platform marked with yellow and black tape to which our car would align when the train arrived. Already I thought this was unnecessary information but then to my surprise he explained a few more things. Please stay behind the yellow line, when the train arrives let passengers get off before boarding, and please bear in mind while doing so that there will be a gap between the station and the train, and if you happen to drop a belonging into it do no try to retrieve it – be it a cellphone, a bag, a pair of sunglasses – do no try to retrieve it, that’s what we are here for, we will retrieve it for you.

There are many kinds of books.

In 2025 I did a few things.

I wrote an essay called The Virtual Book but I never defined the term. By virtual book I mean a book unbound by the traditional and physical constraints of printed books. Even though I think the greatest possibilities await in the virtual world of computers, I don’t think virtual books need to be digital. The possibilities that excite me challenge not only the physicality of books but also their intangible attributes.

A virtual book can be multimedia. It can consist of words, images, video, audio. There, we got the obvious one out of the way.

A virtual book can be readerdriven. Instead of forcing readers to follow the author’s thought process, a virtual book can let each reader steer the way. Wikipedia does this already. It lets you search the page for keywords, skip to the section you’re interested in, and even escape into a tangential topic, never to return. This is a natural way to consume Wikipedia because its form affords it.

Books generally have one start and one ending, but a virtual book can be nonlinear. Wikipedia is again the obvious example. But letting the reader drive is only one way to create a nonlinear book. It’s also possible to create multiple entrypoints, or even multiple endings, like Black Mirror: Bandersnatch.

A virtual book can be dynamic. It can change after its initial creation. Printed books, on the other hand, are static snapshots laboriously rendered by a particular author at a particular time. But what if a theory is debunked? Or a hypothesis confirmed? Or a record shattered? Or, in the case of storytelling, what if a loose end can be tied up neatly?

A virtual book can be nonmonolithic. It does not need to be discrete or selfcontained. It can consist of many interconnected parts that make up the whole but can exist without it. It can reference other virtual books, borrow bits from them, and lend bits of its own. For example, if Herbie Hancock’s memoir was a virtual audiobook, it could allow its snippets to be reconstrued into a documentary about jazz. (If Ken Burns’ Jazz documentary series was also ‘virtualized’, it could have been updated 15 years after its release to include bits of Herbie’s narration.) In fact, it could provide material for documentaries about many different topics: jazz, funk, hip hop, Miles Davis, Black Nationalism, Nichiren Buddhism, meditation, and crack addictions, to name some of the obvious ones.

A virtual book can be responsive. What if a reader could expect a book to field spontaneous questions? ChatGPT is an obvious candidate here, but the possibility is broader. What if Herbie Hancock returned to his memoir every now and then to answer questions that readers had left behind while reading it? What if readers could raise flags on issues that factcheckers would then verify or return to the author for amendment?

The possibilities are plenty, and they are thrilling. The difficulty in realizating them is not technological, but legal and political. Powerful companies – and therefore governments – are hugely incentivized to prevent the free exchange of “intellectual property”. To make virtual books possible, we need not only the technological power of software, but also its progressive politics.

This page is inspired by Viktor Lofgren’s The Small Website Discoverability Crisis and Tom Critchlow’s list of digital bookshelves. It’s a reference list of websites that I have visited in the past and may want to visit again.

There are a handful of sites that I frequent these days. Normally I start by going to sive.rs/now to see what Derek Sivers has been up to recently. Reading it inspires and mobilizes me. I also tend to check macwright.com/reading to see what Tom MacWright has been reading. I recently picked up

because he gave it five stars and it wouldn’t be hyperbole to say it has changed my life. Not majorly, but in a smaller, meaningful way. As I wrote at the end of my most recent

update, I begin 2025 with strong intentions regarding what I do with my attention.

In his January 16, 2025 /now update, Derek Sivers linked to this essay about the joy of riding an electric bike. I spent the next couple hours learning about this guy, Craig Mod. I’m thrilled to find another person who is and has been thinking deeply about books as technology and creating

(he calls “book-shaped”) experiences online. Craig is not the first technologist-writer whose site I’ve encountered recently. I also found robinsloan.com and jsomers.net, both of whom are accomplished writers and experienced programmers.

Although personal sites are perhaps the most exciting, there are also some great ones run by organizations. The most special of these is probably Capitol Hill Seattle Blog, a neighborhood blog turned crowdsourced news site turned indepedent and ultralocal news site. It’s special not only because it is independent but because it offers volumes of dense and specific information about a focused locale. I’m yet to find anything else like it.

I also used to be a big fan of pudding.cool, but lately have lost some of my enthusiasm for them because they seem to pick topics for novelty and appeal. They are perhaps more an entertainment company than a journalism firm.

I often come across new sites and I like to save them here to remind myself to check back on them and consider adding them to the small set of sites I frequent. Here are sites that haven’t made it into my special list but I want to revisit soon:

I’m particularly interested in websites that organize their content as an interconnected network:

On the surface, this is an ordinary blog. For its techno-literary lean, it could be categorized with the likes of robinsloan.com, craigmod.com, sive.rs, jsomers.net, and macwright.com. My goal, however, is not just to publish my thoughts on subjects that interest me but to promote

of #reading and #writing. I attempt to do this not only through the ideas I discuss on site but also through the custom features I build into its interface, which allow readers to explore the content on this site according to their curiosity:

I use GitHub Issues to track problems and potential improvements.

I started developing this site in February 2023. The fundamental unit of this site is a post. What you’re reading right now is a post. I often call them “pieces,” too. I write each post as a Markdown text file using a text editor, usually VS Code and occasionally vim. From those text files, Jekyll generates the site’s HTML and CSS using the Minima layout theme and the Solarized color palette. I have tweaked fonts, spacing, and other UI details by modifying the _sass/minima/_layout.scss file. The features described in the table above are implemented using Ruby plugins and Liquid.

The site is static and consists mostly of HTML and CSS and a little bit of JavaScript. It’s hosted for free on GitHub Pages. The only money I’ve spent so far was on the domain okjuan.me, which I bought in June 2023 from bookmyname.com for ten years for €156 and configured okjuan.github.io to point to it.

To publish a post, I commit it to the main branch of vbook GitHub repo. I configured that repo to publish on GitHub Pages as okjuan.github.io/vbook, which gets resolved by DNS to okjuan.me/vbook. GitHub provides a default process for building Jekyll sites and publishing them to the web, but I use the community-built Jekyll Deploy Action because I have several custom plugins and GitHub only supports a few select plugins.