The Rise of Platform Dependency and What It Means for Freedom

Email works the way it does for a specific, almost old-fashioned reason: nobody owns it. Anyone can run a mail server, using an open protocol standardised in the 1980s, and it will interoperate with everyone else's mail server without asking permission first. A Gmail user can email a Yahoo user can email someone running their own server in a spare room, and none of them need the others' platform's approval for it to work.

This used to be the default architecture of the internet, not the exception. Understanding how that changed — not just that it changed — is the more useful history to know, because the mechanism explains today's platform dependency far better than any single company's specific choices do.

The internet's original design was deliberately decentralised

The foundational protocols of the internet — email, the web itself, Usenet, IRC — were built by engineers and researchers with an explicit design philosophy: no single point of control, interoperability by default, and open standards anyone could implement without asking permission. This wasn't naivety about business models; several of the internet's architects wrote explicitly about avoiding centralised control as a design goal, not an accident.

For roughly two decades, this architecture mostly held. Websites linked to each other freely. Anyone could set up a blog, a forum, a mailing list, using open tools, reachable by anyone else using different open tools. The web was, in a real technical sense, a commons — nobody's permission was required to build on it.

Why centralisation won anyway

The shift toward platforms wasn't a conspiracy or a single decision — it was the accumulated result of several genuine advantages centralised services offered, each individually reasonable, that compounded into today's landscape.

The first was simplicity. Running your own mail server, or your own website, required technical skill most people didn't have and didn't want to acquire. A centralised platform that handled the technical complexity in exchange for some control was, for most people, a completely rational trade at the individual level.

The second was network effects, which reward concentration almost mathematically: a social platform becomes more valuable to you as more people you know join it, which makes it more valuable to the next person considering joining, which accelerates further concentration. Open protocols don't generate this same gravitational pull in the same way, because their value doesn't depend on which specific server you're using — which paradoxically made them less "sticky" for any single company to build a business around, even as it made them more resilient.

The third was capital. Venture-backed platforms could operate at a loss for years to acquire users, subsidised growth that open, non-profit-motivated protocols simply couldn't match. A free, well-funded, centralised alternative reliably out-competed a technically superior but unfunded open one, again and again, across nearly every category of online service.

What was actually lost in the trade

None of these individual trades were obviously wrong at the time they were made — simplicity, funded growth, and network effects are all genuinely valuable, and centralised platforms delivered real user benefits that the earlier open protocols often struggled to match in polish and ease of use.

What was lost, cumulatively, is harder to see because it never announced itself as a single loss: the internet went from an environment where switching providers meant changing a setting to one where switching providers means abandoning years of social context, content, and reputation. Interoperability — the quality that let a Gmail user reach a Yahoo user without either company's cooperation — quietly stopped being the default assumption for anything built after roughly the mid-2000s, replaced by walled gardens that interoperate poorly or not at all with each other, by design rather than oversight.

Nobody voted to trade an open internet for a centralised one. It happened one individually reasonable convenience at a time.

Why this history matters more than any single platform's behaviour

Understanding platform dependency as the predictable output of specific structural incentives — rather than as a story about any one company being unusually greedy — matters because it points toward the actual lever available for change. Blaming a specific platform for specific behaviour treats the problem as a matter of that company's ethics. Understanding the underlying incentive structure — network effects rewarding concentration, capital subsidising growth, complexity favouring centralised solutions — points instead toward structural interventions: interoperability requirements, funding models for open protocols that can compete with venture capital's patience, and genuine portability standards.

This reframing also changes what "freedom" would require going forward. It's not primarily about individual willpower to choose harder, less convenient alternatives — that fight was already lost once, for entirely rational reasons, and would likely be lost again on the same terms. It's about whether the next generation of infrastructure gets built with interoperability as a design requirement from the start, the way the original internet was, rather than as an afterthought fought for after the fact.

Frequently asked questions

Was the internet ever actually decentralised, or is this nostalgia?
It genuinely was, for its first two decades or so — email, the early web, and Usenet all ran on open protocols nobody owned, with interoperability by design rather than by exception.

Why did centralised platforms win if open protocols came first?
Mainly three compounding advantages: simplicity for average users, network effects that reward concentration, and venture capital that could subsidise growth in a way open protocols couldn't match.

Is this just about any one company being greedy?
Not primarily — the pattern shows up across nearly every category of online service, suggesting it's a structural outcome of the incentives involved rather than a story about any single company's specific ethics.

Can the internet become more decentralised again?
It's an active area of work — interoperability regulation and new funding models for open protocols are both attempts to correct the imbalance, though they're working against the same structural incentives that caused the original shift.

Does this mean using open protocols today is impractical?
Not impractical, but it requires more deliberate effort than defaulting to a centralised platform, precisely because the mainstream ecosystem has been built around centralisation for two decades now.