The internet is not private. Each time you visit a website, your IP address is logged. Your internet provider can see every domain you connect to. The websites themselves can see your approximate location. Advertising networks compile behavioural profiles of your browsing across sites. Governments and corporations perform deep packet inspection on backbone traffic. The default state of being online is to be watched. You can either accept this or push back.

The tools for pushing back are often presented as a menu: VPN, Tor, Freenet, each with its own features and limitations. But presenting them as independent options misses the point. They are not competitors. They are steps in a single logical progression, and understanding the logic of that progression matters more than memorising the details of each tool. The central question at every step is the same: who do you trust, and what would it take to trust no one?

A VPN is the first step. It creates an encrypted tunnel between your device and a server somewhere else. Your traffic exits from that server instead of from your home address. To the websites you visit, you appear to be wherever that server is. To your ISP, you appear to be sending encrypted data to a single destination. The benefit is real: your IP address is hidden from websites, your browsing is hidden from your provider, and simple geo-restrictions are bypassed. But the architecture has a fundamental limitation. The VPN provider can see everything — every website you visit, every connection you make, all of it passing through their server unencrypted at the exit point. You have not eliminated surveillance. You have moved it from your ISP to a company that may or may not be more trustworthy. Mullvad, Proton VPN, and IVPN have strong reputations and have been tested in court. Others have been caught logging and sharing data despite their marketing. A VPN is better than nothing. It does not offer anonymity. It offers improved privacy with an asterisk.

Tor takes the next logical step. If the problem with a VPN is that you must trust a single server operator, the solution is to distribute that trust across multiple operators so that no single one can betray you. Tor routes your traffic through three relays — a guard, a middle relay, and an exit relay — each operated by different volunteers around the world. Your traffic is encrypted in three layers. The guard relay knows who you are but not what you are accessing. The middle relay knows neither. The exit relay knows what you are accessing but not who you are. No single relay has both pieces of information. When used correctly, this design provides genuine anonymity. Your ISP sees only that you are using Tor. The websites you visit see only the IP address of the exit relay. The relays themselves cannot correlate your identity with your activity.

Tor is not perfect. Timing attacks, compromised exit nodes, and browser fingerprinting can all break the guarantee. But it represents a qualitative shift in architecture. Trust is no longer placed in a single entity. It is distributed across a network, and the network’s integrity depends on the assumption that not every node is colluding against you at the same time. Tor also provides hidden services — .onion sites that exist only inside the Tor network, their physical location concealed from everyone including their visitors. These enable whistleblower dropboxes, forums in repressive countries, and news outlets that cannot be taken down by any government. The same architecture that protects a dissident also protects a criminal, because the architecture cannot distinguish between them and was not designed to. This is the encryption paradox in network form: a tool that provides privacy for legitimate dissent necessarily provides privacy for illegitimate activity.

Freenet represents the logical endpoint of the progression. If a VPN shifts trust from your ISP to a single company, and Tor distributes trust across a network of volunteers, Freenet eliminates the need for trust by making surveillance structurally impossible. It is not a tool for accessing the regular internet anonymously. It is an entirely separate network — an encrypted data store where content is fragmented, encrypted, and scattered across the computers of everyone running the software. You retrieve content by its cryptographic key rather than by a server address. You do not know whose machine is storing the fragment you requested, and you do not know what fragments your own machine is storing for others. There is no server to shut down, no host to arrest, no central point to attack. Content survives as long as enough people request it to keep it replicated. Unpopular content fades away when people stop asking for it. This is censorship resistance not through policy but through architecture — the network is structured so that there is nothing to censor, no one to subpoena, no building to raid.

Freenet never achieved mainstream adoption. It is slow, difficult to use, and incapable of deleting anything once published. The possibility of unknowingly storing illegal content on your own hardware is an ethical burden most people will not accept. But Freenet demonstrated something that matters more than its adoption numbers: that a network can be designed so that surveillance is not merely difficult but structurally impossible. This idea has influenced every subsequent attempt to build censorship-resistant infrastructure.

Between Tor and Freenet lies a distinction worth making explicit. Tor answers the question: how do I read and publish on the existing internet without being watched? It is a privacy layer on top of the web we already have. Freenet answers the question: what would an internet look like if it were designed from scratch so that watching was impossible? It is a separate medium where anonymity is not a feature but a property, like wetness is a property of water. They were trying to solve different problems, and comparing them as competitors misses what each one teaches.

The progression from VPN to Tor to Freenet is interrupted by a caution worth stating plainly. Decentralisation and anonymity are distinct properties. BitTorrent is decentralised — files are distributed across peers rather than hosted on a central server — but it is not anonymous. When you download from a swarm, your IP address is visible to every other peer. Anyone monitoring the swarm can log who is downloading what. Decentralisation protects against a server being shut down. It does not protect against surveillance. Privacy requires specific architectural choices: encryption at rest, encryption in transit, and the careful separation of identity from activity. A fully peer-to-peer system can still expose every participant if it was not designed to hide them. This is the same lesson the encrypted messaging world learned: federation protects against one company controlling the network, but it does not protect against eavesdropping unless encryption is built into the protocol from the start.

Where you land on this spectrum depends on what you are protecting against. If you want to stop your ISP from selling your browsing history, a VPN is sufficient. If you need to protect a source or publish under a repressive regime, Tor is the minimum. If you believe that the ability to surveil is a problem in itself — that networks should be designed so that surveillance is impossible, not merely discouraged — then Freenet’s vision matters even if its implementation never reached the mainstream. The battle for privacy is fought on many fronts, and network anonymity is only one of them. But it is the front that protects the right to read, publish, and dissent without being watched. That right is worth defending, even when the tools are imperfect and the path is inconvenient.