Messaging Without the Internet: Jack Dorsey’s Radical Experiment

I’ve been following Jack Dorsey’s career since his days at the helm of Twitter. After building the social network that redefined public discourse he walked away, in search of more decentralized forms of communication, in line with his beliefs about freedom in the digital environment.

His latest brainchild, Bitchat, illustrates the lengths he is prepared to go in pursuit of his ideas. An instant messaging app that does not require an internet connection, servers or user accounts, a phone number, email address or any personal information, instead, it works simply with a bluetooth-enable phone. The goal is to show that it’s possible to chat and even transfer bitcoins without relying on mobile networks or servers, using only the modest Bluetooth radio that we all carry in our pockets. And it works, at least to a certain extent.

The app was launched on July 10 as a limited beta and sold out its 10,000 TestFlight slots in a matter of hours. Now, it’s available in the App Store, and offers a somewhat spartan experience: you open the app and, if there’s someone nearby who also has it, you’re already in the channel, an almost romantic return to the IRC era, but without any common infrastructure. And it is precisely this precariousness that makes Bitchat a useful tool for festivals, excursions, or natural disasters, where infrastructure goes down and any form of communication, however rudimentary, is worth its weight in gold.

The premise behind Bitchat sounds almost like an ideological manifesto. Inspired by technologies such as Bluetooth Mesh, it allows messages to jump from device to device, bouncing between nearby phones as if they were a human neural network. The documentation itself boasts commands such as /join and /msg, message compression, and power-saving modes, but warns that the effective range rarely exceeds 300 meters. If two users are more than 300 meters apart, nothing happens: but with enough intermediate nodes, the message arrives, and, in theory, end-to-end encrypted with X25519 and AES‑256‑GCM. Messages self-destruct by default, leave no record in any cloud, and re only stored locally. Privacy? Total. Censorship? Impossible. Third-party dependency? None.

Don’t like ads? Become a supporter and enjoy The Good Men Project ad free

There’s more than a whiff of crypto-anarchism and anti-system resistence about all this, but then in a world where more and more governments shut down the internet at will or use it to surveil the population, it’s not such a crazy idea.

From the user’s point of view, the experience is radically different from systems such as WhatsApp, Signal, or Telegram: no identity verification; not a cloud in sight, and no “last seen”. There is no easy way to know if you are really talking to who you think you are talking to. It is a “trustless” system in the most literal and radical sense of the term. And that, as expected, has unleashed a storm among security experts.

Specialists such as Jim Miller didn’t take long to tear apart the app’s cryptography: the public key system is, in its current implementation, purely decorative. The “Favorites” feature allows you to store another user’s key, but with no guarantee that it has not been spoofed at some point. A man-in-the-middle attack is trivial and easy to automate. Added to this are optional signatures, unprotected headers and privileged commands that any user can hijack with a modified client. We are not talking about trivial bugs, but design flaws that could compromise the authenticity and integrity of conversations. In practice, Bitchat is not secure. Or at least, not yet.

Dorsey knows this: the repository itself warns that the software has not been audited and may contain vulnerabilities, a note that was added after closing (and subsequently reopening) the first critical issue on GitHub. Following the controversy, he announced the migration of the protocol to Noise and began accepting pull requests to strengthen the cryptography, acknowledging that his “weekend project” needed to mature before being used in less secure contexts.

And yet there is something deeply interesting about this experiment. Because Dorsey does not seem to be selling a definitive solution, but rather posing a provocative question: what if we could communicate without relying on centralized infrastructure? What if the phone in your pocket were, on its own, an autonomous message transmission station? What if we didn’t need Apple, Google, or any other authority to talk to whoever we want, whenever we want, and without leaving any trace?

The idea is not new: Bridgefy tried it in 2019 during the Hong Kong protests, and since then, other decentralized solutions such as Briar and Reticulum have proliferated. What is new here is the approach: a mixture of hacker spontaneity — Dorsey claims to have programmed it in a weekend — minimalist aesthetics, and a narrative as consistent with his personal evolution as it is provocative in technological terms. The most recent addition, the ability to send Bitcoin via the Lightning Network over Bluetooth without an internet connection, only reinforces the vision: communication and payments without intermediaries, without a trace, without supervision. It’s as if someone set out to design an app for a post-apocalyptic future or a radically free society.

Will it work? It’s hard to say. On a technical level, Bitchat is still in its infancy. The design flaws are too basic to recommend its use in sensitive situations. From a product perspective, its usefulness is limited by user density: without other devices nearby, there is no network, no messages, nothing. But as an idea, as a concept, as a statement of intent, it has immense value.

It forces us to rethink what we understand by “messaging” in a world saturated with surveillance and digital dependency. At its core, what Jack Dorsey is doing with Bitchat is not building a rival to WhatsApp, or even Signal. He is sketching out another category of app, a philosophy of communication that prioritizes autonomy over convenience, privacy over interoperability, resilience over ease of use. It’s imperfect, yes. It’s full of holes, too. In fact, the debate that has erupted illustrates the paradox of security: we demand Signal standards from a product that is only a week old, but at the same time we punish the transparency that makes that collective audit possible. Bitchat is not secure today, and anyone who uses it to coordinate a protest will expose themselves to fatal mistakes. However, the decision to open the code, acknowledge mistakes, and absorb pressure from the community is precisely what has allowed projects like Signal to achieve the robustness we now take for granted.

If Dorsey has proven anything in his career, it is that he knows where the world is going before others do. And if this experiment manages to find a community interested in strengthening, auditing, and adapting it, it would not be the first time that one of his ideas has ended up redefining how we communicate.

Maybe the problem isn’t that Bitchat is insecure; it’s that everything else we use is much more so, and it doesn’t bother us anymore: a few accept it as a lesser evil, and many others, the majority, don’t even consider it or are unaware of it. What does the future hold for Bitchat? If Dorsey manages to recruit enough developers and undergo rigorous audits, it could become the benchmark for post-internet messaging: a kind of “channel zero” that always works as long as there are a handful of phones nearby. If, on the other hand, it remains a weekend toy, it will end up like FireChat: a fond, discontinued memory of what might have been.

Knowing Jack, my bet is that he will persist until the prototype really supports the anonymity and verification requirements it currently lacks. And if he succeeds, in a few years we may remember this July as the moment when the first social network that doesn’t need the internet began to take shape.

(En español, aquí)

—

This post was previously published on MEDIUM.COM.

—