This week we celebrated my daughter’s 20th birthday. It is one of those occasions that forces a kind of reflection most ordinary days don’t. Not because the day itself is dramatic, but because the number is. Twenty years is long enough that you cannot hold the whole of it in your head at once. You have to go looking for it in pieces.
So I went looking. Partly out of sentimentality, and partly, I’ll admit, out of professional curiosity. What was actually happening in the world of data, storage, and technology twenty years ago? And how different does it look from where we are standing now?
“Almost unrecognisably different” is the answer. And yet, at the time, none of it felt like a revolution. It felt like a series of small, slightly interesting updates. That is the thing about genuine change: you rarely recognise it as it’s happening. You only see it clearly once you turn around and look back.
Where we were
Twenty years ago, we had just built and moved into our new storage warehouse. At the time, that felt like the future. A purpose-built facility designed around the serious, unglamorous business of keeping physical records safe, organised, and compliant.
The conversations happening across the industry at the time were almost entirely about two things: physical storage, and the early promise of electronic document management. EDM, as it was known, carried with it the great ambition of the era — the paperless office. It was the phrase everyone used and almost nobody achieved, at least not in the way it was originally imagined. The idea was straightforward enough: digitise what you have, store it electronically, reduce the physical footprint. It was about efficiency and space. It was not, in any meaningful sense, about insight.
And when it came to the archive itself (the actual records being kept) the purpose was almost entirely legal. Organisations stored data because regulation, compliance, and good governance required them to. Retention periods were observed because they had to be observed, not because anyone believed there was a second, richer use for that information sitting quietly in the archive. There was no second question being asked of the data. What could this tell us? What patterns are hiding in here? Nobody was asking, because at the time, there was no realistic way to answer.
There was no AI. Not as part of the conversation, not as a distant ambition, not even as a buzzword yet. It simply did not exist in the vocabulary of records management. The idea that an archive could be “AI-ready” would have meant nothing to anyone in that warehouse twenty years ago, because the concept itself hadn’t been invented.
What else was happening that year
It is worth pausing on what else was unfolding in the wider technology world at that same moment, because looked at individually, none of it seemed especially momentous. Looked at together, in hindsight, you can see the first scaffolding of the world we now live in being quietly assembled.
Google bought YouTube that year. A video-sharing site that, at the time, looked like a slightly chaotic home for amateur clips, and that few people predicted would become one of the most powerful media and advertising platforms on the planet.
Twitter launched the same year. A strange, stripped-back idea; short bursts of text, limited to a tight character count, with no clear sense yet of what it would become. Nobody quite knew what to make of it. It would take years before its influence on public discourse, politics, and culture became fully apparent.
Amazon Web Services launched S3 and EC2. Of everything on this list, this is probably the one that received the least attention at the time and has had the most lasting structural impact. Cloud storage and elastic compute, as concepts, were barely understood outside of a fairly narrow technical audience. What S3 and EC2 actually did was lay the groundwork for an entirely new model of infrastructure — one where businesses no longer needed to own and maintain their own servers, where storage and computing power could simply be rented, scaled, and accessed on demand. It also accelerated the wider build-out of fibre optic infrastructure, the physical backbone that all of this new digital activity would eventually depend on. Without realising it, this was the moment the foundations were being poured for everything that would later make cloud computing, big data, and ultimately AI possible.
And on a lighter note, the Nintendo Wii launched that year, with its motion controllers reframing what a games console could be and who it could be for. The PlayStation 3 launched too, pushing further into high-definition gaming and entertainment. Both, in their own way, were a sign of how quickly consumer expectations of technology were shifting. Gesture, immersion, and entertainment becoming as central to the conversation as raw processing power.
None of these things, on their own, looked like the start of a revolution. They looked like product launches. Incremental steps. Interesting, but not obviously historic.
The thing about incremental change
That is the part I keep coming back to. At the time, every one of those developments felt incremental. A new console. A new website. A slightly odd new way of storing files online. None of it carried the weight of “this changes everything,” because change rarely announces itself that way.
It is only with twenty years of distance that the shape of it becomes visible. The quiet launch of S3 and EC2 was the beginning of the infrastructure that would eventually make machine learning and AI computationally possible at scale. The launch of Twitter was an early step toward a world where information — true and false — would move faster than any institution could reasonably respond to it. YouTube’s acquisition was an early signal of where attention, and therefore value, was migrating.
And the archive was sitting there the entire time, accumulating decades of information that nobody yet had the tools or the imagination to fully use. The unglamorous warehouse full of boxes, built for legal compliance and nothing more. It existed purely to satisfy a legal obligation. The idea that this same information might one day be queried, analysed, and turned into operational insight wasn’t a deliberate omission. It simply wasn’t yet a question anyone knew how to ask.
The twenty-year-old in the room
My daughter was born into the world this article describes the beginning of. She has never lived through the “before.” She has never known a world where you couldn’t find an answer in seconds, where video wasn’t instantly available, where storage and computing power weren’t simply things you rented on demand without a second thought.
For her, none of this was a transition. There was no moment of adjustment, no “before and after” to compare. The world she was born into already worked this way. She has no memory of dial-up, of physical archives being the only option, of an internet that needed to be waited for rather than simply assumed.
I do. I remember the warehouse, the EDM conversations, the paperless office that never quite arrived as promised. The difference between us isn’t really about technology at all. It’s about expectation. She expects information to be instant, searchable, and available. I remember when that expectation didn’t exist yet. When the idea of an archive being more than a legal obligation would have sounded, frankly, a little strange.
That gap, between someone who grew up inside a fully digital world and someone who watched it being built, is not something either of us chose. It is simply a function of when you happened to arrive.
Twenty years on
I don’t think anyone in that warehouse twenty years ago could have predicted where we’d be now. Not really. Not the specifics. Nobody was sketching out generative AI on a whiteboard, or imagining that the archive itself would one day be discussed not as a legal obligation but as a strategic asset. The future rarely arrives in the shape anyone expects it to.
What strikes me most, looking back, is not how much the technology changed — although it obviously did, dramatically — but how much our relationship to data changed alongside it. Twenty years ago, data was something you stored and protected because the law required it. Today, the conversation has shifted to what that same data could be telling you, if only it were structured and accessible enough to ask it the question.
That shift didn’t happen overnight. It happened the way most real change happens: slowly, in small steps that didn’t feel particularly significant at the time, until enough of them had accumulated that the world looked entirely different from where it started.
Happy birthday, and thank you for twenty years of memories that have absolutely nothing to do with data, storage, or any of the above. Although I’ll admit the Wii has given us plenty of fun memories.