The Internet we thought we were building

In the beginning …

there was a period where the internet felt almost unquestionably hopeful.

I was an undergraduate at Victoria University of Wellington in the late 1980s and graduated into the early 1990s, right as the networked world was beginning to emerge from universities and research institutions into public consciousness.

Back then, the internet did not feel like it does today, not a polished commercial platform, but a loose collection of systems, protocols, terminals, academics, hobbyists, engineers, and curious people experimenting in real time with what global connectivity might become. It felt like a frontier, a new frontier.

You could sit at a terminal in windy Wellington and suddenly participate in conversations, systems, and communities that spanned the planet. For someone growing up in New Zealand … geographically distant from much of the world’s technical and cultural infrastructure … this felt profound.

The early online world was fragmented and messy. There was no singular “internet” with a capital ‘I’. There were overlapping ecosystems:

Long before feeds and platforms and apps, there were shared systems humming away in university machine rooms. I still remember the arrival of an SGI HPC system at VUW – touted as the fastest in the southern hemisphere (well … that title lasted a long time … not … thanks to Moore’s Law).

I still remember network games running on the VAX 760 during my first year, Adventure, Dungeon … MUD something like that. I’ll never forgot at one point having my spaceship dragged around by another player’s tractor beam in an early proto – Star Trek game, Netrek I think. (Ironically you can’t do that in STO).

The internet still felt playful then … experimental … communal.

Usenet in particular carried an energy that is difficult to fully explain to people who only experienced the modern social web.

It was argumentative, chaotic, technical, occasionally FLAME-able, and intellectually alive.

You could move from discussions about operating systems to philosophy, networking protocols, science fiction, politics, distributed systems, astronomy, or obscure hobbies in the space of minutes.

After graduating, I continued working/studying at VUW … around that same period I set up the earliest form of the-bach. Like many early personal sites, it was part profile, part experiment, part declaration that you existed online at all. (In fact, my first ever ‘real’ professional job interview was as a direct consequence of that profile … which I completely trashed ofc, the interview that is.)

That was one of the defining characteristics of the early web: people built things because they wanted to participate in the architecture of the internet itself. The web still felt handcrafted. Personal pages were not heavily curated identity brands.

They were often collections of interests, links, notes, experiments, and fragments of personality assembled by technically curious people exploring a new medium. That distinction matters now more in hindsight than it did at the time. Back then people were primarily judged by what they wrote;

Not by follower counts.
Not by personal branding.
Not by algorithmic visibility.

The culture rewarded curiosity and competence more than performance. I think hats a really important point.

The great wyrm …

NCSA Mosaic: suddenly information became visual and navigable. Hyperlinks stopped feeling abstract and started feeling spatial … explorable. At first the web was small enough that manually curated directories were still useful. Then search engines appeared. AltaVista felt almost magical, an information universe you could dynamically query and the Internet earnt its capital ‘I’. Then Netscape and Navigator transformed everything.

One of the strongest memories from that era was the emergence of Amazon. Which initially mostly meant buying books. Technical books (and maybe T-shirts).

For me — and for many others — that often meant endless streams of titles from O’Reilly Media: UNIX, networking, TCP/IP, perl, CGI scripting, system administration best practices – many of which still sit proudly holding their own on my bookshelves. The famous animal covers became part of techno-pop culture itself.

Suddenly you could discover knowledge globally and have it shipped directly to your door. Again, it is difficult to overstate how revolutionary this felt at that time.

There was a genuine belief that global connectivity would democratise knowledge, flatten hierarchies, and broaden human understanding. The internet genuinely enabled extraordinary collaboration, learning, open-source development, publishing, communication, and access to information.

Utopia denied.

But there was also an assumption … perhaps an overly hopeful one … that increased access to information would naturally produce better personal and societal outcomes.

The internet we thought we were building was often framed around:

Reality turned out to be more complicated … the same systems that democratised knowledge also democratised influence … and influence behaves different. Many modern systems optimise instead for:

Return to The Garden of Eden

Perhaps that is why I feel a strange mixture of nostalgia and unease today. I remember a period when the network still felt human-scaled. When understanding the machinery was normal.

When personal websites felt more important than platforms.

When the internet seemed less like a commercial optimisation engine and more like a shared intellectual frontier … not perfect … not utopian … but hopeful.

Maybe that is also part of why I find myself rebuilding the-bach now, despite it feeling slightly parochial or antiquated in an era dominated by platforms, feeds, and algorithmic visibility.

Early personal sites were not merely content surfaces, they were places.

Small pieces of independently owned digital territory connected together by curiosity, hyperlinks, and shared exploration.

And while we probably cannot — and should not — simply return to the early internet, I do think the emergence of AI presents a new opportunity to correct course. Potentially a very important one because AI will almost certainly amplify whatever incentives and systems we build around it.

If we attach it purely to advertising economics, engagement maximisation, outrage amplification, and behavioural extraction, we may accelerate many of the very dynamics already fragmenting society.

Instead, if we use it to augment human capability, improve synthesis, democratise expertise, reduce cognitive overload, strengthen communities, and return agency to individuals, then perhaps this next phase of the internet does not need to end in further centralisation and manipulation.

Perhaps it becomes an opportunity to rediscover some of the original aspirations that made the early web feel so hopeful in the first place.

Not a return to the past, but a chance to carry forward some of the better angels of our nature.