You don’t have to be an engineer to feel the difference. Open a messaging app, and you’re either in a walled garden or on a public footpath. The garden is pretty, well-maintained, and has a gatekeeper. The footpath is a little rougher, but nobody can fence it off. That’s the gap between a platform and a protocol, and it’s not just a technical detail. It’s a decision about who gets to make the rules. Maren Solberg has spent twenty years watching these decisions get baked into the systems we live inside, and the pattern never changes: platforms hoard power, protocols spread it around.

What a Protocol Actually Is

A protocol is a handshake. It’s a shared set of rules that lets different systems talk to each other without anyone asking for permission. Email runs on SMTP, a protocol hammered out in 1982. The web runs on HTTP. These aren’t products you buy from a company. They’re specifications—documents that say, “If you format your data this way, everyone else who follows the spec will understand you.” You can build software that speaks SMTP, and it’ll exchange messages with any other SMTP-speaking software on the planet. Nobody owns the protocol. The Internet Engineering Task Force keeps the documents tidy, but the protocol itself belongs to whoever uses it.

That’s why your Fastmail account can chat with a Gmail account without a second thought. The two companies compete, but they interoperate because they both follow the same rules. A protocol doesn’t care about your business model. It doesn’t care about your market share. It just defines the grammar of the conversation and gets out of the way.

What a Platform Is

A platform is a fenced-in field. It’s a piece of software—or an entire ecosystem—owned and operated by a single entity. Facebook is a platform. Twitter is a platform. Uber is a platform. When you step onto a platform, you’re not just using a tool. You’re entering a jurisdiction where the owner writes the laws, holds the keys to your data, and can redraw the boundaries whenever they like. Platforms often start by piggybacking on open protocols. Facebook leaned on HTTP and email notifications to pull people in. But once they reach a certain size, they brick up the exits. They stop playing nice with competitors. They make leaving feel like a punishment.

The real question is who holds the power. In a protocol-based system, power is scattered across everyone who participates. Run a server, build a client, spin up a new service—nobody can stop you. In a platform, power pools at the top. The owner decides what’s allowed, who gets silenced, and how your information gets used. If you don’t like it, your only move is to walk away—and lose everyone who stays behind.

The Email Story: A Protocol That Refused to Die

Email is the old warhorse that keeps proving the point. SMTP was born in 1982, and it still carries the world’s messages. You can pick from thousands of providers. You can run your own server in a closet. If you own your domain, you can switch providers without changing your address. The protocol doesn’t care who you are or what client you use. It just moves the mail.

But email also shows how platforms try to colonize open ground. Gmail is the biggest player, and Google has spent years grafting on proprietary features that only work inside Gmail. Smart replies, nudges, certain spam-filtering tricks—none of that is in the SMTP spec. Those are platform features, designed to make Gmail stickier. The protocol still hums along underneath, but the platform layers on top create a quiet lock-in. Most people don’t notice until they try to pack up and leave.

Social Media: The Platform Counter-Revolution

Social media could have grown up on protocols. For a while, it did. RSS let anyone subscribe to updates from any blog or news site. XMPP and IRC ran open, decentralized chat. But these protocols didn’t have venture capital pouring gasoline on them. They didn’t have growth hackers or engagement algorithms. They were slow, steady, and boring—exactly what good infrastructure should be.

Then Facebook, Twitter, and Instagram arrived. They traded convenience and network effects for control. They made it dead simple to find friends and share photos, but they also made it impossible to take your social graph with you. Your followers, your posts, your identity—all locked in a company’s database. The platform model didn’t win because it was better technology. It won because it was politically convenient. It concentrated power in a way that made rapid growth and monetization easy.

Why Protocols Are a Slog to Build

Protocols demand coordination. You need a bunch of independent parties to agree on a standard, implement it, and keep things compatible. This is slow, messy work. It means committees, mailing lists, and endless arguments about weird edge cases. Platforms, by contrast, can sprint. A single company can design, build, and ship a feature in weeks. They don’t need consensus. They just need users.

But that speed has a dark side. When a platform screws up, everyone on it pays the price. Facebook tweaks its algorithm, and millions of small businesses watch their reach evaporate overnight. Twitter suspends an account, and there’s no real appeal. The platform’s word is final because there’s no other court to go to. Protocols, by spreading power around, also spread the risk. If one email server goes down, the rest of the network keeps humming. If one XMPP server starts censoring, you can move to another and keep your contacts.

The Political Layer Beneath the Code

Every piece of infrastructure has a political philosophy baked into its bones. Protocols are federalist. They set up shared rules with local autonomy. Platforms are authoritarian. They centralize authority and expect obedience. This isn’t a metaphor. It’s a structural fact. When you build on a protocol, you’re choosing a governance model. When you build on a platform, you’re choosing a boss.

Maren Solberg has watched this same story play out in city planning, energy grids, and transportation networks. The dynamics are identical. A public road is a protocol: anyone can use it, and no single entity controls access. A private toll road is a platform: you pay for entry, and the owner sets the terms. The technical layer is just the surface. The real question is always about power.

When Platforms Swallow Protocols

There’s a pattern that repeats across the tech industry like a bad song. A new technology shows up as an open protocol. Early adopters build tools and services on it. Then a well-funded company steps in, builds a slicker user experience on top, and starts adding proprietary extensions. Over time, the protocol becomes a commodity, and the platform becomes the tollbooth.

This happened to the web itself. HTTP is an open protocol, but Google’s search engine became the de facto index of the web. Chrome became the dominant browser. AMP, a Google-controlled framework, tried to replace HTML for mobile pages. Each move pulled more of the web’s infrastructure under Google’s roof. The protocol still exists, but the platform has captured so much traffic and tooling that the openness is increasingly theoretical.

The Exit Strategy Problem

Here’s a quick test to tell a platform from a protocol: Can you leave without losing everything? With email, the answer is yes—if you own your domain. With social media, the answer is almost always no. Your data, your connections, your reputation are all bolted to the platform. This is intentional. Platforms spend heavily to make exit painful. They call it “retention,” but it’s really a form of lock-in.

Protocols, on the other hand, have low exit costs. You can switch email providers, DNS registrars, or web hosts without breaking the underlying system. The protocol guarantees continuity. That’s why protocols are tougher. They don’t depend on any single company staying alive. When a platform dies—remember MySpace, Google+, or Vine?—everything built on it vanishes. When a protocol-based service shuts down, the network keeps going.

Building for the Long Haul

If you’re an engineer or a founder, the choice between building on a platform or a protocol is one of the biggest decisions you’ll make. Building on a platform is faster and often more profitable in the short run. You get access to existing users, payment systems, and distribution channels. But you’re also building on rented land. The platform can change the rules, jack up prices, or compete with you directly. Amazon has done this to countless third-party sellers. Apple has done it to app developers.

Building on a protocol is slower and harder, but it creates value that lasts. You’re contributing to a commons that can outlive any single company. The Linux kernel, the HTTP protocol, and the Bitcoin network are all examples of protocol-based infrastructure that has survived for decades. They’ve weathered corporate attacks, government pressure, and internal fights. They persist because no single entity can pull the plug.

FAQ

Is Bitcoin a platform or a protocol?

Bitcoin is a protocol. It defines a set of rules for a decentralized ledger, and anyone can implement those rules in software. There are multiple Bitcoin clients, multiple mining pools, and multiple exchanges. No single entity controls the network. However, some services built on top of Bitcoin—like certain exchanges or custodial wallets—are platforms. They add a layer of control on top of the open protocol.

Can a platform become a protocol?

Rarely, and usually only under pressure. When a platform becomes too dominant, regulators or market forces may force it to open up. Microsoft’s move to support open document formats was partly a response to antitrust pressure. But true protocols are designed to be open from the start. Retrofitting openness onto a platform is difficult because the business model depends on control.

Why do users choose platforms over protocols?

Convenience and network effects. Platforms invest heavily in user experience, making their products easy to use and full of features. Protocols often have clunkier interfaces because they’re built by volunteers or small teams. Platforms also benefit from the fact that everyone else is already there. It’s hard to switch to a protocol-based alternative when your friends, family, and customers are all on the platform.

Are there any successful protocol-based social networks?

The Fediverse—a network of servers running software like Mastodon, Pleroma, and Pixelfed—is a protocol-based social network built on the ActivityPub standard. It’s small compared to Facebook or Twitter, but it’s growing. The key difference is that no single company owns the Fediverse. You can choose a server, move between servers, and still interact with anyone else on the network. It’s social media the way email is email.

The Ghost in the Machine

Back in the early 1970s, Vint Cerf and Bob Kahn weren’t dreaming of marketplaces. They were wrestling with a military logistics puzzle: how to make a bunch of incompatible computer networks talk to each other without a central switchboard. The answer was TCP/IP, a protocol suite that didn’t care about the content of your message, only that the packets got where they were going. It was a design born of a specific political philosophy—decentralization as resilience. A protocol is a set of rules, a common language. It’s a sidewalk, not a shopping mall. It doesn’t matter if you’re walking a dog or staging a protest, as long as you stay on the path.

Fast forward half a century, and the digital world isn’t run by protocols. It’s dominated by platforms. Facebook, Amazon, iOS—these aren’t neutral paths; they’re gated communities with strict HOAs. The difference isn’t just technical jargon. It’s a fundamental question of power. A protocol defines a how. A platform dictates a who and a what. Understanding this shift is the first step in recognizing why the internet feels less like a frontier and more like a company town.

The Unseen Hand of the Protocol

A protocol is an agreement. It’s a specification for how data is formatted, transmitted, and received. The key political feature is that no single entity owns it. HTTP, SMTP, TCP/IP—these are standards, not products. They’re governed by bodies like the Internet Engineering Task Force, whose mantra, famously coined by network engineer David Clark, is: “We reject kings, presidents, and voting. We believe in rough consensus and running code.” It’s a deeply anarchic, meritocratic approach to infrastructure. If your code works and others adopt it, it becomes the standard.

This architecture has radical consequences. Because no one owns the protocol, no one can charge rent. The Simple Mail Transfer Protocol (SMTP) lets anyone set up an email server. You don’t need a license or permission from a corporate landlord. The barrier to entry is technical skill and a server, not a fee. This is why email, at its core, is federated and decentralized. A Gmail user can email a Fastmail user without a problem because they’re both speaking SMTP. The protocol is a public utility, even if we mostly access it through private interfaces.

The Walled Garden as a Business Model

A platform, on the other hand, is built on top of protocols, but its entire purpose is enclosure. Facebook uses HTTP; it’s not a protocol. It’s a platform. The difference lies in your relationship to the infrastructure. On a protocol, you’re a peer. On a platform, you’re a tenant. The platform’s Terms of Service replace the protocol’s technical specifications as the governing law. You can build a new email client that speaks SMTP without asking anyone’s permission. Try building a new client for Facebook’s social graph without their API key. You can’t. They control the door, and they can revoke your key on a whim.

This is a massive transfer of sovereignty. In the protocol era, innovation happened at the edges—anyone could build a new web browser or email client. The platform era centralizes innovation. The platform owner decides which features are allowed, which data you can access, and which business models are permitted. That’s why platforms are so sticky and so profitable. Apple’s iMessage is a classic example. The green bubble versus blue bubble isn’t a technical limitation; it’s a deliberate choice to make communication outside the iOS ecosystem feel inferior. It’s soft coercion, nudging users to stay inside the walls to maintain social status and a fluid chat experience. This isn’t a bug. It’s the core logic of platform capitalism. The goal is to make the cost of leaving so high that users become serfs, not customers.

When Protocols Become Platforms

The story gets messier when protocols themselves become the foundation for platform power. Google’s dominance started with a protocol: the PageRank algorithm for crawling and indexing the web. The web itself is a protocol (HTTP) and a format (HTML), open and unowned. But Google built a platform on top of it, becoming the de facto gateway. Today, the search engine doesn’t just index the open web. It answers questions directly, hoarding traffic inside its own walled garden of knowledge panels and featured snippets. The protocol is still there, but the platform has made it nearly invisible—a substrate to be mined, not a space to be inhabited.

This pattern repeats. Email is a protocol; Gmail is a platform that uses the protocol but adds proprietary layers of spam filtering, categorization, and ad insertion. The underlying SMTP standard ensures delivery, but the user experience is entirely mediated by Google’s business logic. The platform doesn’t break the protocol; it subsumes it, making the open standard feel like a feature of the corporate product. It’s a subtle but effective form of enclosure, turning a commons into a shopping mall without ever tearing down the public square—just building a roof over it and charging for the escalators.

The Governance Gap

The fundamental difference between a platform and a protocol is governance. A protocol is governed by a transparent, if often slow and bureaucratic, process of standard-setting. The Request for Comments (RFC) system, for all its flaws, is a public record of debate and decision-making. A platform is governed by a Terms of Service document, written by corporate lawyers, and enforced by opaque, algorithmic moderation systems. There is no due process on a platform. Your account, your content, your digital identity can be erased by a single automated decision with no meaningful appeal.

This governance gap is the central political crisis of the modern internet. We’ve moved our public sphere—our town squares, our newspapers, our postal services—onto private platforms. The result is a form of digital feudalism, where a handful of corporations exercise sovereign power over speech, commerce, and identity. The solution isn’t to demand nicer kings. It’s to rebuild the infrastructure using protocols that distribute power. The blockchain and cryptocurrency movement, for all its excesses, is fundamentally a reaction to this crisis: an attempt to create money and contracts as protocols, not platforms.

Frequently Asked Questions

What is a real-world example of a protocol outside of tech?

The postal system is a classic protocol. A letter’s format—address on the front, stamp in the corner—is a standard. The system doesn’t care about the content of the letter. It is a neutral carrier. A shopping mall, by contrast, is a platform. It controls who can set up a store, what they can sell, and what the environment looks like. The mall owner can evict a store for violating its rules, just as a social media platform can de-platform a user.

Why did platforms win over protocols?

Platforms won because they offer a better user experience and solve the “last mile” problem of trust and discovery. A protocol like SMTP can deliver email, but it can’t filter spam, suggest contacts, or provide a polished mobile app. Platforms bundle these services, creating a smooth experience in exchange for user data and lock-in. The convenience is real, but the cost is a loss of user agency and the creation of massive centralized points of control and failure.

Can a platform ever become a protocol?

It’s rare, but possible. The key is for the platform owner to relinquish control and allow the underlying system to be governed by an open standard. This happened with the web itself. Tim Berners-Lee convinced CERN to release the World Wide Web technology into the public domain without royalties. This act of deliberate non-ownership turned a project into a universal protocol. For a modern platform to do the same, it would require a business model not based on extracting rent from user data and attention—a near-impossible ask for a publicly traded company.

The Architecture of Control

Back in the early 1990s, the internet wasn’t a place you visited. It was a set of agreements you plugged into. Email ran on SMTP—a protocol nobody owned. The web used HTTP, which no government could revoke. These weren’t products. They were languages. And because they were languages, they were generative. Anyone who learned to speak them could build something new.

Now look around. Most of what people call “the internet” is actually a handful of private platforms. Facebook, Twitter, YouTube—these aren’t protocols. They’re walled gardens that mimic the behavior of open networks while quietly centralizing control. The difference between a platform and a protocol isn’t just technical. It’s political. It shapes who gets to speak, who can build, and who can walk away.

A protocol is a shared language for machines. It sets the rules for interaction but doesn’t own the conversation. SMTP doesn’t care if you use Gmail or a server in your closet. It just delivers the message. A platform, on the other hand, is a governed territory. It often runs on protocols underneath, but it mediates access, enforces terms of service, and monetizes your attention. The platform decides what’s visible, what’s acceptable, and what’s profitable.

When we blur this line, we hand over the keys to our digital public squares. We mistake a landlord for a city planner.

The Political Economy of Protocols

Protocols aren’t neutral. They encode assumptions about power, trust, and participation. TCP/IP, for instance, was built on the idea of cooperative peers. It never anticipated massive content delivery networks or state-level surveillance. Yet its openness let the internet scale in ways no centralized system could have planned.

Platforms flip this model. They start with a centralized core and then offer APIs—application programming interfaces—that simulate openness. But these APIs are revocable. They’re permissions, not rights. When Twitter shut down third-party clients in 2023, it wasn’t a technical failure. It was a demonstration of sovereignty. The API was never a protocol. It was a leash.

This pattern echoes through history. The Roman roads were a kind of protocol for empire: anyone could use them, but only on Rome’s terms. The British railway gauge became a standard that locked colonies into specific industrial paths. Today’s digital platforms are the roads and rails of the twenty-first century. The question is who gets to lay the tracks.

Email: The Last Open Protocol Standing

Email is still the most successful open protocol we have. SMTP, POP3, and IMAP are specifications anyone can implement. You can run your own mail server, pick your own client, and reach anyone regardless of their provider. That interoperability isn’t a bonus feature. It’s the definition of a protocol.

But even email is under siege. Gmail’s dominance means Google effectively controls deliverability. If your independent server gets blacklisted by Google’s spam filters, you’re cut off from a huge chunk of users. The protocol is open, but the implementation is concentrated. That’s platform logic creeping into protocol territory—a slow enclosure of the commons.

Why Platforms Keep Winning

Platforms win because they solve problems protocols can’t. Protocols evolve slowly. They need consensus from a messy coalition of stakeholders. Platforms can ship features overnight, fix bugs, and polish the user experience without waiting for a committee. That speed is intoxicating.

Platforms also solve discovery. RSS can syndicate content, but it won’t tell you what to read. Facebook and Twitter use algorithms to surface posts, creating a feedback loop that keeps you scrolling. The trade-off is that the algorithm isn’t impartial. It’s tuned for attention, which often means amplifying outrage, sensationalism, and division.

The money doesn’t line up either. Protocols are public goods. They’re maintained by volunteers or funded by grants. Platforms are businesses. They extract rent from every interaction they host. That rent-seeking pushes them to enclose what was once open. Google’s AMP project was sold as a way to speed up the mobile web. In reality, it gave Google more control over how content gets served and monetized.

The Sovereignty Question

When a society builds its communication systems on private platforms, it outsources sovereignty. The platform becomes the de facto public square, but it’s governed by terms of service, not law. Content moderation is handled by trust and safety teams, not courts. This isn’t a conspiracy. It’s a design choice we’ve collectively accepted.

Protocols distribute sovereignty. No single entity controls SMTP or HTTP. If you don’t like how your email provider handles your data, you can switch to another one and still reach everyone. That’s the power of interoperability. It’s also why platforms resist it. Interoperability threatens their business model.

The European Union’s Digital Markets Act tries to force platforms to behave more like protocols. It mandates that messaging services be interoperable—so a Signal user can message someone on WhatsApp. That’s a step toward unbundling platform control over user bases. But it’s a regulatory patch, not an architectural fix. The underlying systems stay proprietary.

The Fediverse: A Protocol Revival

The fediverse—a network of interconnected servers running open-source software like Mastodon—is a real attempt to build social media on protocols. ActivityPub, the protocol behind it, is a W3C standard. It lets users on different servers follow, like, and reply to each other. No single company owns the network.

It’s not a utopia. The fediverse has its own governance headaches. Server admins hold significant power over their local communities. They can block other servers, moderate content, and set their own rules. But that power is distributed. If you don’t like how your server is run, you can move to another one—or start your own—without losing your social graph. That’s a fundamental break from platforms, where leaving means losing everything.

The Price of Convenience

Platforms sell convenience in exchange for agency. They remember your preferences, suggest connections, and filter content so you don’t have to. This isn’t a gift. It’s a bargain. You hand over control of your data, your attention, and your social graph for a frictionless experience. The platform learns from your behavior and uses that knowledge to keep you hooked, sell ads, or shape your worldview.

Protocols are less convenient. They demand more effort to set up and maintain. But they give you control. You decide where your data lives, who can access it, and how it’s used. This isn’t a technical distinction. It’s a political one. Choosing between a platform and a protocol is choosing between renting and owning, between being a user and being a citizen.

Building for the Long Now

If we want an internet that serves the public interest, we need to invest in protocols. That means funding open-source development, supporting standards bodies, and choosing protocol-based services over platform-based ones. It also means designing new protocols that address the shortcomings of the old ones. We need protocols for identity, payments, content moderation, and governance that don’t rely on centralized middlemen.

This isn’t a call to abandon platforms entirely. They have their place. But we should be clear-eyed about what they are: private infrastructure with public consequences. The next generation of the internet should be built on protocols that are as open, resilient, and generative as the ones that came before. The alternative is a digital world where a handful of corporations decide what’s possible.

Frequently Asked Questions

What is the main difference between a platform and a protocol?

A protocol is an open standard that defines how systems communicate, allowing anyone to build compatible software without seeking permission. A platform is a centralized service that controls access, data, and interactions, often built on top of protocols but adding proprietary layers of control.

Why do platforms tend to dominate over protocols?

Platforms offer faster innovation, better user experience, and built-in discovery mechanisms. They solve coordination and monetization problems that protocols, as public goods, struggle with. This convenience attracts users, creating network effects that reinforce the platform’s dominance.

Can a platform become a protocol?

Not easily. A platform would need to relinquish central control, adopt open standards, and allow interoperability with competitors. This runs counter to the business model of most platforms, which rely on user lock-in and data ownership. Regulatory pressure or market shifts could force such a transition, but it is rare.

What are some examples of protocols still in use today?

Email protocols like SMTP, IMAP, and POP3 remain widely used. HTTP and HTTPS underpin the web. DNS translates domain names to IP addresses. ActivityPub powers the fediverse. These protocols enable decentralized communication without a single controlling entity.

The Ghost in the Machine

Most people don’t see the difference. They tap an app, fire off a message, and assume the digital world just hums along. But underneath every swipe and click is a choice that shapes who holds power, who gets heard, and who takes a cut. That choice is between a platform and a protocol. I’ve spent years tracing this line, and it’s never been just a technical one—it’s political infrastructure, plain and simple.

A protocol is a handshake. A set of rules that lets different systems talk to each other without asking permission. Email runs on SMTP. The web runs on HTTP. Nobody owns these. They’re not products; they’re languages spoken by machines. A platform, on the other hand, is a walled garden. Facebook, Uber, Airbnb—they set the terms, broker the interactions, and take their pound of flesh from every deal. The difference isn’t about code elegance or user interface. It’s about who holds the keys.

The Shape of Control

When you send an email, you can use Gmail, Outlook, or a server you run in your closet. The protocol—SMTP—doesn’t care. It’s an open standard, hammered out by the Internet Engineering Task Force, a loose collective of engineers. No single entity can revoke your ability to send mail. The protocol is indifferent to your identity, your location, or your message. It just routes the packets.

Now look at a platform like Twitter. You can post, reply, and retweet, but only through Twitter’s interface and under Twitter’s rules. The company can shadowban you, demonetize you, or boot you off entirely. The platform isn’t a neutral pipe; it’s a curated space with a business model. The drift from protocols to platforms over the last two decades is one of the most consequential—and least discussed—power shifts of our time.

How Protocols Build Commons

Protocols create commons. TCP/IP, the internet’s backbone, doesn’t charge per packet. It doesn’t demand a login. It just defines how data should be wrapped and addressed. This wasn’t a happy accident. The early internet architects, many funded by public research grants, deliberately built a permissionless system. They knew a network’s value grows with every new participant, and that gatekeeping would choke it.

That openness unleashed a torrent of creativity. Anyone with the skills and a server could build a website, launch an email newsletter, or invent a new application-layer protocol. The barriers were technical, not corporate. The commons was messy, chaotic, and wildly generative. It gave us the web, VoIP, and countless other tools that platforms later repackaged and monetized.

The Platform Playbook

Platforms didn’t invent new human needs. They captured existing behaviors and inserted themselves as middlemen. Before Facebook, people shared updates via email, forums, and personal blogs. Facebook offered convenience—a single feed, a unified identity, a frictionless interface. The hidden cost was that you now shared on Facebook’s terms, and your social graph became their asset.

This playbook repeats across industries. Uber wedged itself between drivers and riders. Airbnb between hosts and guests. Spotify between musicians and listeners. In each case, the platform provides real value—discovery, trust, ease of payment—but it also extracts a toll and hoards the data. The underlying activity—transportation, lodging, music—existed long before the platform and could exist without it. But the platform makes leaving expensive.

The architecture reinforces this lock-in. Platforms are centralized databases with proprietary APIs. They can rewrite the rules overnight. A protocol like email is distributed; no single server failure can take down Gmail and Yahoo at once. A platform can deplatform you completely. That’s not a flaw. It’s the business model.

When Protocols Push Back

History offers examples of protocols challenging platform dominance. The web itself was a protocol victory—HTTP and HTML let anyone publish without asking AOL or CompuServe. More recently, ActivityPub, the protocol behind Mastodon, tries to bring federation to social networking. Instead of one company owning the social graph, different servers agree on a common language. You can move your account between servers, or run your own, without losing your connections.

Matrix does similar work for messaging. It’s an open protocol for decentralized, encrypted communication. You can use the Element app, or build your own client. The protocol guarantees interoperability. If a server goes down or kicks you off, you switch. The network survives individual failures.

These efforts face an uphill battle. Platforms have network effects, deep pockets, and user inertia. Most people don’t care about protocols until something breaks. But the existence of these alternatives matters. They prove the current infrastructure isn’t inevitable. It’s a choice.

The Political Layer

Calling infrastructure political isn’t a metaphor. When a platform sets speech rules, it acts as a legislator. When it resolves user disputes, it acts as a court. When it takes a cut of every transaction, it acts as a tax collector. These are functions of a state, performed by private companies with no democratic mandate.

Protocols push governance to the edges. The core rules are minimal and hard to change. Decisions about content moderation, monetization, and community norms happen at the server or client level. This doesn’t erase politics; it distributes it. One Mastodon server can enforce strict anti-harassment rules, while another permits almost anything. Users pick their jurisdiction. It’s messy, but it’s also accountable in a way that a platform’s opaque moderation can never be.

The economic incentives are fundamentally different. A protocol is usually maintained by a nonprofit or standards body, funded by dues or grants. Its goal is stability and interoperability. A platform is funded by venture capital and advertising; its goal is growth and lock-in. The protocol wants to become boring infrastructure, like the electrical grid. The platform wants to become indispensable, like a utility you can’t switch off.

Why This Matters Now

We’re at a fork in the road. The platform era is mature, and its costs are hard to ignore: surveillance capitalism, algorithmic radicalization, fragile supply chains controlled by a handful of firms. At the same time, the tools to build protocol-based alternatives have never been cheaper. Cryptographic identity, distributed storage, and consensus mechanisms enable new kinds of networks that don’t need a central operator.

But protocols aren’t a magic fix. They’re slower to develop, harder to fund, and often deliver a rougher user experience in the short term. The challenge isn’t just technical; it’s social. Protocols demand collective maintenance. They ask users to shoulder some responsibility for the infrastructure they depend on, rather than outsourcing it to a platform’s terms of service.

The choice between platform and protocol isn’t about pining for some decentralized past. It’s about what kind of digital society we want to build. Do we want a world where a few companies own the pipes and the content, or one where the pipes are public and the content belongs to communities? The answer won’t be decided by technology alone. It’ll be decided by the political choices we make about who gets to control the infrastructure of everyday life.

Frequently Asked Questions

What is the main difference between a platform and a protocol?

A protocol is an open standard that lets different systems communicate without a central authority. A platform is a controlled environment where a single company sets the rules, mediates interactions, and often extracts value from transactions. The key distinction is who holds power: protocols distribute it, platforms concentrate it.

Can a platform be built on top of a protocol?

Yes, and it happens all the time. Gmail is a platform built on the SMTP email protocol. You can use Gmail’s interface and features, but you can also send email to someone using a different provider. The protocol layer ensures interoperability, while the platform layer adds convenience and, often, lock-in. The tension is that platforms have incentives to extend protocols in proprietary ways that erode the commons.

Why do protocols struggle to compete with platforms?

Protocols are typically developed by standards bodies or open-source communities with limited resources. They prioritize stability and consensus, which slows iteration. Platforms, backed by venture capital, can move fast, subsidize user acquisition, and build polished experiences. Network effects also favor platforms: users go where other users already are, and platforms can lock in those users with proprietary features that make leaving costly.

Are there any successful modern protocols that rival platforms?

ActivityPub, the protocol behind Mastodon and other federated social networks, is a notable example. It enables decentralized social networking where users on different servers can interact. While it hasn’t displaced Twitter or Facebook, it has grown significantly and demonstrated that protocol-based social media is viable. The web itself, built on HTTP, remains the most successful protocol ecosystem, though it is increasingly mediated by platform gatekeepers like search engines and app stores.

We talk about the internet as if it’s a single, coherent thing. It’s not. It’s a stack of decisions, piled one on top of the other, some open and some locked tight. The most important split in that stack isn’t between hardware and software, or code and content. It’s between platforms and protocols. Get this distinction wrong, and you’ll misunderstand who holds power in the twenty-first century.

Protocols are agreements. They’re the rules of the road that let different systems talk to each other. Take TCP/IP, the bedrock of the internet. It doesn’t care what kind of computer you’re on, what operating system you’re running, or what you’re trying to say. It just defines how packets of data should be addressed and routed. It’s a neutral standard, a common tongue. Email runs on protocols too—SMTP for sending, IMAP or POP3 for receiving. Anyone can set up an email server that speaks these languages, and it will work with any other. The power is scattered. The rules are public. No one needs permission to join.

Platforms, on the other hand, are walled gardens with a gift shop at the exit. A platform is a controlled environment, owned and operated by a single entity, where interactions happen according to the owner’s rulebook. Facebook is a platform. Uber is a platform. The Apple App Store is a platform. You can build a business on them, sure, but you’re a tenant farmer. The landlord can change the rent, alter the rules, or evict you at will. That’s not a flaw in the platform model; it’s the whole point.

The Political Economy of a Stack

The early internet was a messy bazaar of protocols. HTTP, the protocol for fetching web pages, was open. Anyone could build a server or a browser. This openness sparked a wildfire of innovation in the 1990s. But the real money, it turned out, wasn’t in the open protocol layer. It was in the proprietary platforms that sat on top. Google built a platform on the open web. Facebook built a platform on the open social graph. Amazon built a platform on open commerce. They took the public infrastructure and built private shopping malls on it.

This is the bitter irony of the modern internet. The most valuable companies in the world owe their existence to the radical openness of protocols, yet their business models depend on closing that openness down. They capture users, data, and interactions inside silos. A protocol like email is decentralized; no one owns it, and no one can charge you rent for using it. A platform like WhatsApp is centralized; Meta owns the pipes, the data, and the relationship. The shift from protocol to platform is a shift from a digital commons to a feudal estate.

Interoperability as a Political Weapon

Protocols are political because they govern relationships. A protocol that mandates interoperability is a direct check on monopolistic power. The early internet’s design philosophy, hammered out in the rough consensus of the IETF, was to solve technical problems, not to create business moats. The result was a system where no single player could lock out the others. If you didn’t like your internet service provider, you could switch and still reach the same web, the same email, the same everything. The protocol layer guaranteed your freedom of movement.

Platforms reverse this. They make leaving painful. Your data, your social graph, your reputation—all of it is held hostage. You can quit Facebook, but you can’t take your network with you. This isn’t an accident. It’s a deliberate lock-in strategy. The platform’s interface is designed to be sticky, to make the cost of exit feel insurmountable. The protocol layer, by contrast, is designed to be invisible. It just works, quietly, without demanding your loyalty or your data.

The Siren Song of the Platform

So why do we keep choosing platforms? Convenience, plain and simple. Protocols are raw. They demand you bring your own interface, your own storage, your own identity. Platforms hand you a polished, all-in-one package: a profile, a newsfeed, a camera, a payment system, all bundled up neatly. This convenience isn’t a gift. The price is your autonomy. You swap the messy, open-ended potential of a protocol for the smooth, predictable rails of a platform. You become a sharecropper on land you’ll never own.

This trade-off is rarely presented honestly. Platforms market themselves as communities, public squares, digital utilities. But a real utility is regulated and interoperable. You can plug any toaster into any wall socket because the electrical grid is a protocol, not a platform. Imagine if your blender only worked with Con Edison electricity. That’s the platform model. The cozy language of “community” is a smokescreen for a proprietary data-extraction machine.

Identity and the Protocol Problem

Platforms won the first round of the internet largely because protocols failed to solve identity. TCP/IP gives a device an address, but it doesn’t tell you who’s behind the keyboard. Email protocols let you send messages, but they don’t verify you are who you claim to be. Platforms rushed into this vacuum. Facebook became the de facto identity layer for billions. “Log in with Facebook” is a confession that the open web couldn’t solve a basic human need: proving you’re a unique, real person without handing all your data to a central authority.

That’s starting to change. Cryptographic identity systems, built on public-key infrastructure, are making protocol-based identity workable. You can own your identity as a set of keys, not as a row in some corporate database. You can prove things about yourself—your age, a group membership, a professional credential—without exposing everything else. This is the architectural shift that could finally unbundle the platform monopolies. When identity becomes a protocol, not a product, the platform’s strongest lock-in mechanism starts to dissolve.

Data as a Liability, Not an Asset

Platforms hoard data. Their entire business model depends on it. Every click, every pause, every scroll is logged, analyzed, and monetized. They treat this data as a proprietary asset, a deep moat against competitors. But from the user’s perspective, that data is a ticking time bomb. It’s a honeypot for hackers, a dossier for advertisers, a record that can be subpoenaed, leaked, or misused. The platform’s prized asset is the user’s unacknowledged liability.

Protocols flip this relationship. In a protocol-based system, data can be portable and user-controlled. You store your data where you want—on your own device, on a server you trust, or encrypted on a public network. The protocol just defines how that data is formatted and shared when you choose to share it. This is an architecture of dignity. It treats data not as a resource to be strip-mined, but as a personal record to be guarded.

The Governance Gap

Protocols aren’t a magic fix. They have their own pathologies, and the most glaring one is governance. Who decides how a protocol evolves? The IETF’s “rough consensus and running code” model worked well enough for technical standards, but it breaks down when protocols become the substrate for social and economic life. Without a clear governance mechanism, protocols can be captured by powerful interests or simply stagnate. The history of RSS, the open standard for web feeds, is a cautionary tale. It was torn apart by competing corporate interests and a lack of effective coordination, leaving a vacuum that platforms like Twitter and Facebook happily filled.

Platforms, for all their faults, can make decisions quickly. They can ship features, fix bugs, and respond to market pressures. Protocols are slow by design. This is both a strength and a weakness. The slowness prevents rash changes that could break the network, but it also makes it hard to adapt to new user needs. The challenge for the next generation of internet infrastructure is to build protocols that are both sturdy and governable, that can evolve without being captured by a single interest.

The Return of the Protocol

We’re in the middle of a quiet resurgence of protocol-based thinking. The blockchain and cryptocurrency space, for all its noise and fraud, is fundamentally an attempt to build economic and social protocols that don’t rely on centralized platforms. Bitcoin is a protocol for value transfer. Ethereum is a protocol for programmable contracts. These systems are far from perfect. They’re clunky, slow, and often user-hostile. But they represent a genuine architectural alternative to the platform model. They ask: what if the rules of the system were transparent and enforceable by code, not by a corporate terms of service?

This isn’t just about money. The same principles are being applied to social networks, identity systems, and data storage. Projects like Bluesky and Mastodon are building social protocols where users can choose their own algorithms and move their data between providers. This is the opposite of the Facebook model. It’s an attempt to rebuild the social web on a protocol foundation, where the platform is just one of many possible interfaces, not the entire universe.

Infrastructure Is Destiny

The choice between platforms and protocols is a choice about who holds power. Platforms concentrate power in the hands of the owner. Protocols distribute power among the participants. This isn’t a technical distinction; it’s a political one. The architecture of our digital systems shapes our social and economic relationships. A world built on platforms is a world of landlords and tenants. A world built on protocols is a world of peers.

This is why the distinction matters. When we talk about “regulating big tech,” we’re often talking about making platforms less exploitative. But the real question is whether we want to live in a platform-dominated world at all. Regulation can make a landlord slightly less rapacious. It cannot turn a landlord into a neighbor. For that, you need a different architecture. You need protocols.

Frequently Asked Questions

What is a simple example of a protocol I use every day?

Email is the classic example. SMTP (Simple Mail Transfer Protocol) is the protocol that allows you to send an email from a Gmail account to a Yahoo account, or from your work’s custom email server to any other. The protocol is open, so many different companies and individuals can build email clients and servers that all work together. You are not locked into one provider.

Why can’t a platform just be built on open protocols?

Some platforms do use open protocols underneath, but they add a proprietary layer on top that captures value. For instance, Google’s search engine uses the open HTTP protocol to access web pages, but the search algorithm, the ad auction, and the user data are all proprietary. The platform uses the protocol as a raw material source, then builds a walled garden around the refined product. The economic incentive is to close off what was open.

Are protocols always better than platforms?

Not always. Platforms can provide a more coherent user experience and can innovate faster in the short term. A well-run platform can solve coordination problems that a protocol cannot. The danger is when a platform becomes so dominant that it eliminates the possibility of a protocol-based alternative. The goal should not be to eliminate platforms, but to ensure that protocols remain a viable and competitive option, so that users have a genuine choice between convenience and control.