The Library of Babel: Infinite Information and the Scarcity of Judgement
6 min read
In 1941 the Argentine writer Jorge Luis Borges published a short story about a library.
It's no ordinary library. It's made of an endless series of hexagonal galleries, and its shelves hold every possible book: every combination of the alphabet's letters, the comma, the full stop and the space, arranged into volumes of exactly 410 pages. Somewhere in the Library sits every book ever written and every book that ever will be. There's the true history of your life, and also a million subtly false ones. There's the cure for every disease, a faithful catalogue of the Library itself, and thousands upon thousands of false catalogues.
Borges describes the librarians' first reaction when they understood what the Library contained. In James E. Irby's translation:
"When it was proclaimed that the Library contained all books, the first impression was one of extravagant happiness. All men felt themselves to be the masters of an intact and secret treasure."
The happiness doesn't last. Because the Library holds everything, it holds the truth in exactly the same measure as it holds nonsense, and nothing distinguishes one from the other. Librarians wander the galleries for a lifetime searching for a single meaningful page. Some go mad. Some destroy books they deem worthless, not understanding that every volume they burn has countless near-identical copies elsewhere. The treasure turns out to be a prison.
I come back to this story more than any other when I think about cyber threat intelligence, because I think we've built Borges' Library, and we're still in the extravagant-happiness phase.
The Doors Are Open
Consider what an analyst can reach today. Roughly 40,000 new CVEs were published in 2024 alone, more than a hundred a day. Add to that vendor blogs and threat reports, government advisories, open and commercial indicator feeds, Telegram channels, dark-web forums, paste sites, GitHub repositories, academic papers and the endless commentary of security social media. A determined analyst could read for a lifetime and never reach the end of a single week's output.
This was supposed to be the good news. For much of the history of intelligence, the hard problem was collection: getting hold of the information at all. That problem hasn't gone away for the truly secret things, but for most of what a commercial CTI team needs, the information already exists. Somewhere in the Library is the report describing the exact intrusion set targeting your sector, the indicator that would have caught last week's phishing campaign, and the vulnerability that will matter to you next quarter.
Before moving into threat intelligence I worked in security operations and incident response, and I learnt this the hard way. The intelligence we needed usually existed. What didn't exist was the time to find it, read it and work out whether it applied to us. Every shift was a small walk through the Library, and the shelves only grew.
The False Catalogues
The cruelest detail in Borges' story is that the Library contains not only the true catalogue but countless false ones, each as convincing as the real thing.
Our version of the false catalogue is familiar to anyone who has worked in this field for long. There are indicator feeds full of stale infrastructure that was burnt years ago, now generating alerts on innocent IP addresses. There are threat-actor profiles stitched together from other profiles. Worst of all is circular reporting: one original claim is repeated by a second vendor, cited by a news site, summarised by a third blog and quoted back in a fourth report, until a single unverified assertion looks like five independent sources in agreement. Intelligence services have been burnt by circular reporting for as long as there have been intelligence services, and it remains one of the easiest traps in open-source work.
Generative AI has added a new wing to the Library. The cost of producing plausible, authoritative-sounding security content has fallen to almost nothing. Some of that content is useful. Much of it is a confident rearrangement of what was already on the shelves, and some is simply wrong, written in exactly the same polished register as the truth. For the first time, the Library is writing its own books, faster than any of us can read them.
The Search for the Vindications
In the story, many librarians devote their lives to searching for their Vindication: a book that justifies and explains their own existence. They're certain it's out there. Statistically, it is. They will almost certainly never find it.
I see a version of this in CTI teams that have mistaken having information for producing intelligence. The belief is that somewhere in the feeds is the answer, and that if we ingest enough sources, buy one more subscription and add one more dashboard, the answer will present itself. It won't. Intelligence isn't found on a shelf. It's made, by someone who brings a question to the information, weighs it, discards most of it, and states what the remainder means for a specific decision.
That distinction is the whole of our profession. Information is the book. Intelligence is the reader's judgement about the book.
Start With the Question
The intelligence cycle doesn't begin with collection. It begins with direction: someone with a decision to make states what they need to know. In practice that means intelligence requirements, the specific questions a security programme exists to answer. Which adversaries are likely to target us, and how? Which vulnerabilities in our estate are being exploited in the wild right now? What should the board understand about our exposure this quarter?
Without requirements, the Library is unnavigable. Every book is equally relevant, which means none is. With requirements, it becomes a resource. The question tells you which galleries to walk, which shelves to skip, and when you've read enough to answer.
It's remarkable how many CTI programmes still start from the other end, with a list of feeds and tools, and only later wonder what question all of it was meant to answer. If you take one thing from this article, let it be this: write your questions down before you buy another source.
Better Librarians, Not More Books
Borges mentions a figure the librarians call the Man of the Book, a librarian said to have read the catalogue of catalogues and so become, in effect, a god. No one has met him. People make pilgrimages anyway.
We'll never have a Man of the Book. What we can have, and what technology is finally good enough to give us, is a far better class of librarian. Machines are now genuinely excellent at the parts of the work that used to swallow an analyst's day: reading at volume, translating, deduplicating, tracing a claim back to its original source, spotting that five reports are really one report, and mapping behaviour to the frameworks we use to reason about it.
This is, in the end, why I co-founded Deltabridge. The problem I kept running into was never a lack of intelligence. It was a lack of hours to read it. Handing the reading to machines isn't the same as handing over the judgement, though. A good librarian can bring you every book relevant to your question and tell you which ones are copies of each other. Only you can decide what the answer means for the organisation you're defending. The art is in drawing that line deliberately, and in never letting a fluent summary convince you the thinking has already been done.
An Elegant Hope
Borges ends his story on an unexpectedly hopeful note. His narrator suggests that the Library may be unlimited but periodic: that an eternal traveller, crossing it in any direction, would eventually find the same volumes repeating in the same disorder.
"…which, thus repeated, would be an order: the Order. My solitude is gladdened by this elegant hope."
I've always read that as a description of what analysts actually do. Adversaries repeat themselves. They reuse infrastructure, tooling, habits and mistakes. Campaigns rhyme with the campaigns before them. What looks like endless disorder, viewed from a single gallery, resolves into patterns once you've walked enough of the Library with a question in hand.
The Library will keep growing. It will never stop. Our job was never to read all of it. Our job is to know what we're looking for, and to recognise the order in the disorder when we find it.



