The Best Books I Read in 2025

Four books. The shortest list I have put together for this series. Also the most deliberate.

This was the year I read less and reread more. Audiobooks on the commute, which I had resisted for years and should not have. A few old, small books doing more work than a shelf of new, thick ones ever managed.

So here’s what I only noticed when I wrote the four titles down together. Frankl published in 1946. Lao Tzu, if he existed at all, was writing something like two and a half thousand years ago. Saint-Exupéry published in 1943. And Harari published last year.

Three of these have been through the filter. One has not been tested yet.

And the fourth book, the new one, is about the filter itself. Nexus is Harari on information networks and what they select for. And that turns out to be a strange thing to read at the end of a year spent on books which already got through.

So why does that distinction matter? Because a book that lasted two thousand years has already been argued with by everybody, and the ones still standing are standing for a reason.

Four books, one filter, and one of them describing the machine.

Man’s Search for Meaning by Viktor E. Frankl

Frankl wrote this in nine days. It went out in 1946, originally without his name on it, and it comes to about a hundred and fifty pages.

And the book splits cleanly in two. First the camps, reported rather than dramatised. Then logotherapy, the theory he built out of what he watched.

And his argument is that the primary human drive is meaning, ahead of pleasure and ahead of power. He quotes Nietzsche for it. He who has a why to live can bear almost any how. And what Frankl watched in the camps was that the men who held a why kept going through conditions that took the men who had lost theirs.

And there is a detail from the arrival that I keep coming back to. Frankl had the manuscript of his first book sewn into the lining of his coat when he reached Auschwitz. It was taken and destroyed with everything else. He reconstructed it later, on scraps of paper, which is a fairly complete answer to the question of whether he believed his own thesis.

So what is the sentence everybody takes away? That when everything can be taken from a person, one thing holds. The freedom to choose your response to what is happening to you.

So does the theory survive the reporting? It does, and the order is the reason. He earned the second half of the book in the first half, and there is no way to fake that sequence.

Now, I have read a shelf of resilience books that traffic in exactly that sentiment and earn none of it. This one earns every word, because the man writing it lost his wife, his mother and his father to the camps, and he is describing something he stood inside rather than something he read about.

That’s the first one through the filter. Eighty years, nine days of writing, and not a spare paragraph in it.

Tao Te Ching by Lao Tzu, translated by Stephen Mitchell

Eighty-one short chapters. Mitchell’s version is from 1988. The whole thing fits in about a hundred pages.

And it keeps circling the same handful of ideas from new angles. Empty is useful. A wheel works because of the hole at the centre, and a room is worth having because of the space nobody filled. The soft overcomes the hard. Water wears down rock. The leader nobody notices is often the one doing the best work.

And the concept underneath it is wu wei, which gets translated as non-action and is better understood as action that goes with the grain rather than across it. Doing less, but doing it in accordance with how things already are.

So why is a two and a half thousand year old book still in print? Because Mitchell’s version reads like poetry, which makes it easy to finish in one sitting and impossible to absorb in one sitting.

And that gap is the point. I read the whole thing in an evening and then spent months getting maybe four chapters of it. Read a chapter a day. It resists being rushed, and rushing it is exactly what a person who reads nineteen books in a year would do.

So what does wu wei look like on a Tuesday? Mostly it looks like restraint. The engineer who does nothing while a system settles, rather than the one who reaches for the console. And I have been the second engineer far more often than the first, which is probably why the chapter kept finding me.

That’s the second one through the filter. Twenty-five centuries, a hundred pages, and it is still telling people to slow down.

The Little Prince by Antoine de Saint-Exupéry, translated by Richard Howard

A pilot crashes in the desert and meets a small prince who has left his own tiny planet to find out what is wrong with grown-ups.

And the crash is not invented. Saint-Exupéry went down in the Libyan desert in 1935, with his mechanic André Prévot. The two of them survived several days on almost nothing. Then a Bedouin on a camel found them. He published the book in 1943. He disappeared over the Mediterranean in July 1944. About a year.

So it is a children’s book only in the sense that children are still willing to take its questions seriously.

And the line everyone quotes is that what is essential is invisible to the eye. And it earns the fame here, because the whole book is built to make you feel it rather than read it. That is much harder engineering than a good sentence.

Now, here’s the thing nobody warns you about. It reads differently as an adult, and differently again as a parent. The businessman counting stars he will never visit stops being a joke somewhere around the second time through.

The rose is the other half of it. There are thousands of roses in a garden and his is unremarkable among them, and it is his only because of the hours he spent on it. Time spent is what makes a thing yours.

And the fox is the part I had forgotten entirely. He asks the prince to tame him, which in the book means to build a tie by turning up at the same hour, patiently, until the two of them matter to each other. Then the fox delivers the line about being responsible forever for what you have tamed. So is that a children’s lesson? It is the most adult sentence in the book.

That’s the third one through the filter. Ninety-six pages, eighty years, and the author only got one year to watch it travel.

Nexus by Yuval Noah Harari

And here is the new one, the untested one, and the one describing the machine that tested the other three.

And Harari’s argument is that more information has never automatically produced more truth. He traces it from the canonisation of religious texts, through witch hunts, through totalitarian regimes, and out into the algorithms deciding what you saw this morning.

So what are information networks actually built to do? Create order. Truth is one route to order. It is the slow, expensive route. And a network optimising for order will reach for the cheap methods long before it reaches for that one.

So I closed the year on this deliberately. After three books about meaning, stillness and simplicity, this is the one about the world I actually work in. Data. Systems. Networks moving information at a volume that no single person, anywhere, can sit down and audit.

Now, he is a popular historian and he moves fast over ground specialists would take slowly, and the book has been argued with on exactly those terms. Take the frame. Check the examples.

And reading it against the other three was the whole experience. Frankl, Lao Tzu and Saint-Exupéry made it through several centuries of a filter that was slow, narrow and brutally selective. Harari’s point is that we have replaced that filter with one optimising for something else entirely.

So what does he say a healthy network needs? Self-correcting mechanisms. Something inside the system whose job is finding its errors, which is expensive and slow and always the first thing cut when a network is optimising for reach.

That’s the fourth one. Still in the machine. We get to find out.

And here’s where I should be honest about the frame. Four books is a tiny sample, three of them are famous precisely because they survived, and building a theory about survival out of survivors is the oldest mistake in the book. Every writer who did not make it through the filter is unavailable for comment.

That’s my 2025 list. Short, quiet, and closed out by a book about the very networks I spend my working life inside.

A new book will tell you what somebody is thinking now. An old one will tell you what turned out to be worth keeping.