Why build a reading app in the age of AI?

Most information no longer needs to be read.

That much is true, and there’s no point pretending otherwise. If you want a fact, a definition, a summary of something well-trodden — ask a model. You’ll have it in seconds, and reading three hundred pages to get there would be a poor use of your life.

If this app existed to deliver information, it shouldn’t exist.

But there is a difference between knowing something and understanding it, and a further difference between understanding something and being changed by it. Information can be handed over. Insight has to be arrived at. As far as anyone can tell, the routes to it are still the old ones: your own experience, someone who has been where you’re going — and reading, slowly, the people who thought hardest about it before you.

No model has written a wise book.

Not a competent one. Not a useful one. A wise one — the kind you finish and find you’re standing somewhere new.

The reason may be structural. Philosophy tends to come out of finitude: we die, we regret, we lose something real in every choice we make, we cannot take the other road as well. A system that loses nothing can produce sentences about wisdom without ever having occupied the situation that produces them.

Which is not a complaint about AI. It’s a description of where the good material still lives — in what people wrote.

Two kinds of reading AI is worst at.

They sit at opposite ends of everything ever written, and they’re the two that still repay the effort of reading in full.

Everything in the well-trodden middle, a general chatbot handles fine. These two ends are where reading still matters — and, not coincidentally, where the door is hardest to get through.

The barrier was never interest. It was cost.

Almost nobody bounces off Zhuangzi or off a dense paper because they don’t care. They bounce off because page three assumes a decade of context they don’t have, and there is no one in the room to ask. Historically, that door opened for people who had a teacher, a peer group, and time.

This is the one thing AI genuinely breaks open. Not by reading on your behalf — by being the person in the room. Someone to ask what does this sentence actually mean at eleven at night, without embarrassment, as many times as it takes.

That is the entire idea behind Resage: not to spare you the reading, but to make the hard things enterable.

“If AI has no wisdom, why let it explain Zhuangzi?”

The right question, and the product is the answer to it.

Resage doesn’t supply the wisdom. It supplies passage. It is a guide, not an oracle, and it’s built to hold that line even when a fluent answer would be easier:

• If it says your file says something, it shows you where. The passage and the page, as a link — tap it and the reader jumps there so you can check.
• Common knowledge is labeled as common knowledge — not quietly passed off as something your text said.
• When it doesn’t know, it says so. Reaching for a tool first, admitting the gap second, and inventing never.

Getting an old text wrong is worse than getting a modern one wrong: there’s no obvious tell, and a confident fabrication can sit in your head for years. So the standard is deliberately unfashionable. Better a visible “I’m not sure” than a beautiful sentence nobody can check.

Verification stays with you. That’s not a limitation we apologize for — it’s the part where the understanding actually becomes yours.

Where the name comes from.

resage  = re(ad) + sage

A sage: someone worth learning from. But the sage was never meant to be the software.

The sages are the people who wrote what you’re reading — Zhuangzi, Marcus Aurelius, whoever it is you’ve been meaning to get through. They’ve been on the shelf the whole time. The reason you haven’t read them isn’t that they weren’t available.

Resage doesn’t aspire to be your sage. It walks you to them, and stays beside you while you read — translating, unpacking, remembering where you got stuck last time. The last stretch is yours. Nobody can do your reading for you, and you wouldn’t want them to: that stretch is the part that changes you.

What Resage will never do.

Principles are cheap. These are the ones with teeth — the things we’ve given up the option of doing.

• Never train on your files. They stay on your Mac and in your own iCloud. Only the passages needed to answer a question are ever sent, and only to answer it.
• Never ship with a library of its own. Resage arrives empty. Every file in it is one you chose — including the ones nobody would think to recommend.
• No feed, no recommendations, no streaks, no leaderboards. Nothing engineered to bring you back. Reading is not consumption, and an app that treats it as consumption will quietly ruin it.
• Never summarize you out of the reading. The text stays a real, full page you turn yourself. The help waits at the bottom of the screen until you want it.

Who this is for.

People who read — and in particular, people who want to read the difficult things and have never quite found the way in.

If AI can give this era one thing worth having, it isn’t another surface of answers. It’s that the doors which used to open only for people with a teacher now open for anyone patient enough to walk through.

Resage is a door. What’s on the other side was always yours.