A few years ago, I was struck by the tension between how Wikipedia was used, and how it was talked about. Everyone was on it all the time, and yet its reputation had stalled at “don’t trust it”, because it was the “free encyclopedia that anyone can edit”.

Clearly people were trusting it, so much so it had become part of the fabric of everyday life since it was founded by Jimmy Wales and Larry Sanger in 2001. But despite the fact it now has been consulted trillions of times by billions of people (English Wikipedia alone is 5.2bn words long), it inspires an odd lack of curiosity. Even people donating to Wikipedia might say the money was “for the factcheckers”, though there aren’t any – not paid ones, anyway. Meanwhile, the internet in the social media era had begun to feel like a much uglier place, and some digital nostalgia was setting in: here was this holdover that had somehow survived; a website that functions – and even looks – much the same as it did 20 years ago.

You could say that I fell down a Wiki rabbit hole, especially when it came to the website’s early years. Almost everything important about the project came together quite quickly. Its amateurish, homegrown ethos was off-putting to a lot of writers and journalists, and that meant many questions about why it had worked had no answers. Most of the people seriously interested in Wikipedia were a handful of academics – so I decided to write a book about it.

I’m not a tech journalist. Instead, I have a background in the humanities. Wikipedia actually does, too: philosophy, and even off-piste disciplines like town planning and poetry, played a much bigger role in its inception than pure technology. Ward Cunningham, who invented the wiki in the 1990s – that’s a webpage anyone can edit, and the foundational tool of Wikipedia – thought very carefully about how the power of computers might be democratised. He drew on all sorts of oblique concepts for it, such as constructivism and the Zen concept of wabi-sabi – beauty in imperfection – and it’s remarkable how well the culture he created has persisted. Especially wabi-sabi – Wikipedia is, after all, always in a state of incompleteness.

Philosophy drew Wikipedia’s co-founders together as well. Though Wales worked in tech and finance, he corresponded with philosophers for years. And though Sanger was an academic philosopher, he spent lots of time arguing online – only “amateurs” seemed to share his sense that philosophy had life-and-death stakes. Lots of people in the dotcom boom started online encyclopedias, but Wales was the only one to hire a philosopher. In fact the only person he hired was a philosopher. It was such an unpredictable choice, and an important clue to what made Wikipedia work: it sprang out of this digital “republic of letters”, where all sorts of people were thinking about facts and why they were important.

Because Wikipedia is so massive, writing an “exhaustive” book on it is impossible, like trying to represent a mountain with a grain of sand. Instead, I concentrated on the human history – the small, sometimes very small, number of people who really influenced what Wikipedia became. Wikipedians can be difficult, and the most ornery get a bad wrap – even the Wikimedia Foundation’s leadership has made comments about their “Dorito-stained fingers”, or tendency to stay home on a Saturday night. But by the end of the process, I was in awe of them, and I still am.

Wales and Sanger are prominent, but many other important early Wikipedians are unknown. The first woman to edit Wikipedia, Gaytha Langlois, hasn’t had much attention. She wrote about two things dear to her: Rhode Island and standard poodles – small states and big dogs. Wikipedia has always had a gender imbalance, which has proved stubborn – Sandi Toksvig is only the latest in a long line of critics trying to address it. She is part of an effort to create a more balanced alternative to Wikipedia called Mappa Mundi, which has the potential to feed back into Wikipedia as well.

I became fascinated too with a little-known contributor who had the first biographical page for a woman on Wikipedia (today, it would be called a user page). Her name was Ruth Ifcher, and she was a very intelligent grammar pedant, with a PhD, who had been forced out of the workforce with severe agoraphobia. Wikipedia came along later in her life and opened a whole new avenue for creative and meaningful work.

Proust said that neurotics alone “composed our masterpieces” and that’s true here.

Neurodiversity is so central to Wikipedia that one former head of the Wikimedia Foundation told me the encyclopedia is a “city built by Asperger’s”. You can sense that energy in familiar topics – planes, trains and automobiles, comic books, weather, and all the rest. Often, Wikipedians are so prolific that they write too much, a phenomenon called “fancruft”. “Road geeks” (like trainspotters for roads) kept having their work about minor roads deleted, so a whole lot of them left in a huff, and started a separate online wiki encyclopedia for roads. That’s almost a microcosm of the inclusionists v deletionists debate, which is the biggest philosophical rift on Wikipedia – a rolling argument about what should be included, and what shouldn’t.

The co-founders had once believed that 100,000 articles would be “a reasonable size” for an encyclopedia, because that’s how big Britannica was. Today, Wikipedia has 68m, and Pew Research estimates that it served 1.9tn page views between 2015 and 2025. Because it is so monumental, I never expected Wikipedia to feel fragile or vulnerable – but as I was finishing the book, AI had begun to encroach on Wikipedia in an existential way by diminishing the web traffic it relies on, for instance, or by editors contributing huge volumes of poorly written AI-generated text. People ask me “is Wikipedia under threat?”, and it is, but so is everything from Hollywood to pure mathematics. I do think it will survive, but the quaking foundations made it feel urgent to say “we saw this made, and we won’t see anything like it again”.

• The Last Best Place on the Internet: A Human History of Wikipedia by Richard Cooke is out now through Black Inc in Australia, and WW Norton in the US. It will be released in the UK on 5 November, via Granta