Two Bits Per Beacon: Auditing the Bitcoin-Over-Find-My Experiment

The ledger was clean, but the vision was fragile. Most breakthroughs in this industry are marketing wearing a technical costume, and after enough years reading contract bytecode and transaction logs, I can smell the difference at ten paces. So when word reached me that a developer had broadcast a Bitcoin bearer token across Apple's Find My network — no internet, no Lightning channel, no cell tower, just a two-dollar ESP32 chip impersonating a lost AirTag and a thousand strangers' iPhones unknowingly doing the relaying — my first instinct was to file it under party tricks. Then I ran the bandwidth numbers. Two bits per signal. More than a thousand broadcasts to move a single token. And I stopped laughing. Not because it works as a payment system. It does not, and its author says so first. But because of what it quietly proves about the physical layer sitting beneath everything we have agreed to call infrastructure. To understand why this matters, you have to understand two systems that had nothing to do with each other until someone welded them together. The first is Cashu. Cashu is an implementation of Chaumian ecash running on Bitcoin — an idea with real intellectual lineage, tracing back to David Chaum's blind-signature work in 1982, years before anyone put the word blockchain on a whitepaper. The mechanics are deceptively plain. A mint, a server operator, holds Bitcoin in custody and issues bearer tokens representing claims on that Bitcoin. Hold the token, hold the money. No identity, no account, no visible chain of custody, because the signatures are blinded. The mint cannot see which token went to whom. That is the privacy pitch, and it is genuine. It is also — and this is the clause the marketing deletes — a complete surrender of self-custody. You are trusting the mint not to freeze you, not to over-issue, not to run with the reserve. Chaumian ecash gives you privacy by taking away your keys. It is the oldest trade in finance, dressed in new clothes. The second is Find My. Apple's Find My network is one of the largest crowd-sourced sensor networks ever deployed, and almost nobody thinks about it as infrastructure because it is invisible. When an AirTag drifts out of Bluetooth range of its owner, it broadcasts a rotating Bluetooth Low Energy beacon. Any nearby iPhone — not just yours, any iPhone — picks up that beacon, encrypts its own location, and relays it through Apple's servers back to the tag's owner. Apple's phone fleet becomes a global mesh, and the user pays nothing. Elegant. Enormous. Entirely closed. The developer in question, known on X as calle, is a recognized name in the Cashu ecosystem. He took these two systems and asked a question that is either brilliant or insane depending on your risk tolerance: what if the Find My beacon could carry a payment instead of a location? The answer, published as open-source code on GitHub, is yes. Technically. And that is exactly where the interesting part ends and the audit begins. The Architecture of an Absurdity Strip away the novelty and the construction is straightforward in the way that only deeply clever things are. An ESP32 — a cheap, Wi-Fi-and-Bluetooth-capable microcontroller that every hardware hobbyist has a drawer full of — is programmed to emit a Bluetooth Low Energy beacon formatted to look like an AirTag. Nearby iPhones, doing exactly what Apple designed them to do, pick up the beacon and relay it into the Find My network. On the other end, a receiver uses an unofficial interface into Apple's system, ties the signal to an Apple account, and reconstructs the payload. The payload, in this case, is a Cashu ecash token: a bearer instrument backed one-to-one by Bitcoin held at a mint. In other words, the payment does not travel over the internet. It does not travel over Lightning. It travels as a fake lost-luggage signal, hitchhiking on the goodwill of a trillion-dollar company's mesh network and the passive cooperation of every iPhone within Bluetooth range. I want to be fair to the engineering, because it deserves it. The integration is not trivial. You are coupling an embedded radio stack, a proprietary consumer mesh, a Chaumian mint protocol, and a Python-based reconstruction pipeline. That is a cross-stack build that most developers could not complete, and the fact that it runs at all is a genuine demonstration of skill. In 2018, I spent six months manually auditing the smart contracts behind Power Ledger's token sale and found a reentrancy flaw in their distribution mechanism that the team ignored for speed. That experience taught me something I have never unlearned: technical elegance without battle-testing is fatal, and the elegance here is real even as the battle-testing is absent. The code is open, but it is a single developer's experiment, unaudited, unreviewed, and never stress-tested against an adversary. Elegance, again, is not the same as safety. So let us audit it the way it deserves to be audited — not as a product, but as a claim. Two Bits Per Signal Here is the number that reframes the entire story. Each Find My beacon carries roughly two bits of usable information. Two bits. That is enough to encode four states — the equivalent of a two-question multiple choice. A single Cashu token, by the developer's own accounting, requires more than a thousand broadcast cycles to transmit completely. Sit with that for a moment. To move one unit of value — and I use the word value loosely, because the documentation itself recommends sending only gift-level amounts — the system must emit over a thousand separate radio pulses, each one dependent on an iPhone being physically nearby at the right moment to catch it. This is not a latency problem you can engineer away with a bigger antenna. It is a fundamental property of the channel. Find My was designed to carry a rotating identifier and a rough location, not a payment. You are using a firehose to move a single drop of water, and the firehose does not belong to you. Compare this to what we already have. Lightning settles in seconds at sub-cent fees, with point-to-point encryption and hash time-locked contracts. Even a mediocre on-chain transaction confirms in minutes. The Find My path does not confirm; it dribbles. It is the slowest, lowest-throughput payment rail ever demonstrated on Bitcoin, and it is only impressive if you forget that the alternative already exists. This is why I refuse to call it a payment innovation. It is a transport-layer micro-trick. The genuinely new thing here is not the payment protocol — Chaumian ecash is forty years old and Cashu is not new. The new thing is the discovery that a physical data channel previously reserved for lost-luggage beacons can be repurposed to carry value. That is a clever observation about a network nobody was looking at. It is not a breakthrough in how money moves. And note what the two-bit ceiling implies about the future. If anyone ever wanted this to carry more than a novelty, they would have to bolt on compression and error-correction coding, because the raw channel cannot even carry a modest payload. The developer did not build that, because building it was never the point. The point was the demonstration. Which, again, is fine. But do not confuse a proof of concept with a product roadmap. Cash on a Public Square Now the part that turns this from a curiosity into a genuine architectural failure. The payload is unencrypted. The identifier is public. And the token is a bearer instrument. Each of these is survivable alone. Together, they are fatal. A Cashu ecash token works like physical cash: whoever holds it, owns it. There is no signature check at redemption, no account lookup, no way to claw it back. If you hand someone a token, they spend it. If someone else takes it, they spend it. This is not a bug; it is the entire design philosophy of Chaumian ecash. But this experiment broadcasts that bearer token in plaintext, over a public radio channel, with a public identifier, to any device in range. The consequence is brutal and simple. Any third party who sniffs the signal can grab the token and spend it before the intended recipient ever reconstructs it. The money is not just exposed; it is exposed as literal, anonymous cash lying on a public plaza. The security model has degraded from trusting a mint not to steal from you, to trusting nobody within Bluetooth range of your fake AirTag not to steal from you first. I keep coming back to the sentence I wrote years ago and never retired: code does not lie, but people certainly do. Here the code does not even need to lie. The architecture itself is the lie — it wears the costume of a private transfer while broadcasting your money in the open. Could you encrypt it? This is the objection every defender reaches for, and it collapses under its own weight. To encrypt, you need a shared key. To share a key, you need a key exchange. To run a key exchange, you need bandwidth — and the channel delivers two bits per signal, over more than a thousand cycles, best-effort, non-real-time. You cannot run a Diffie-Hellman handshake through a channel that can barely carry a token, and even if you could, you would then need the receiver to be online at precisely the right moment, which Find My does not guarantee. The absence of encryption is not laziness. It is the physical constraint of the channel expressing itself as a security hole. This is a compromise, not an oversight, and it is precisely the kind of compromise that gets people's money stolen. The Parasite and Its Host Every dependency in a system is a point of control for whoever owns it. This is the lesson of the 2024 ETF cycle, when I advised a mid-sized fund in Bogotá on integrating crypto into a traditional portfolio and insisted on strict risk parameters over the objections of people who underestimated volatility. My data-driven framework preserved ninety percent of capital while competitors lost thirty. The reason it worked was not cleverness. It was the refusal to pretend that dependencies do not exist. Every system has a throat, and you must know where it is. The throat of this experiment is Apple. The entire mechanism rests on unofficial access to the Find My interface. There is no partnership, no API key, no license. It is parasitism on Apple's infrastructure, dressed as interoperability. And the host can sever the parasite at any moment, for any reason, without warning or appeal. Apple has already tightened Find My's rotating-beacon privacy model once; a single policy adjustment or a server-side change would render this entire tool inert overnight. This is not a technical flaw you can patch. It is an architectural fact about who holds the leverage. The experiment's greatest apparent strength — that it leverages billions of already-deployed iPhones — is also its deepest fragility. The larger and more valuable the infrastructure you borrow, the more completely its owner controls you. You do not get the network without the landlord. There is a real-world cost hidden in this, and it is not borne by the developer. If this class of parasitic reuse scales — if Apple's mesh becomes a recognized vector for moving money outside any framework Apple controls — the most likely corporate response is to tighten the interface, and the collateral damage lands on the legitimate third-party ecosystem: the certified Find My accessory makers, the tracking-tag startups, the ordinary users. This is the most under-discussed externality in the whole story. A clever hack, repeated often enough, can poison the commons for everyone who used it honestly. And it is worth saying plainly, because the cheerleaders will not: this is not a Bitcoin Layer 2. It has no relationship to the Bitcoin community's actual scaling work. It is a transmission side-channel that happens to carry a token whose value is denominated in Bitcoin. The claim that it expands Bitcoin's capabilities is the same category error I see every cycle — a project borrowing Bitcoin's brand to sell a transport gimmick. The real Bitcoin builders, the ones who have spent a decade on Lightning and covenants and Schnorr, do not recognize this as infrastructure. They should not. Ninety percent of what markets call Bitcoin Layer 2s are Ethereum patterns wearing a Bitcoin flag, and this is one more mask in the drawer. The Mint You Cannot See We have to talk about what a Cashu token actually is, because the experiment's framing obscures it. This is not a new asset. It is a claim on a custodian. When you hold a Cashu ecash token, you are holding a promise from a mint. The mint received Bitcoin and issued you a blinded token against it. Redemption depends entirely on the mint honoring that promise. If the mint's reserve is short, you find out at the worst possible time. If the mint freezes your funds, you have no recourse, because the whole point was that it never knew who you were — which means it also has no obligation to know how to give your money back. Privacy and recourse are in direct tension, and Chaumian ecash resolves it by choosing privacy and abandoning recourse entirely. This is why the standard tokenomics playbook — TGE schedules, vesting cliffs, unlock pressure, FDV — is simply inapplicable here. There is no team allocation, no investor round, no governance token, no emission curve. Forcing those frameworks onto a bearer-payment instrument produces noise, not signal. The only economic question that matters is the integrity of the mint, and the integrity of the mint is an off-chain trust assumption that no amount of on-chain analysis can verify. The mint operator's real business, when you squint, is Bitcoin custody plus note issuance. That is a business that, in most jurisdictions, touches money-transmission licensing. Most Cashu mints have not confronted this. Most operators are hobbyists. Which means the systemic risk is not in the token; it is in the custodian, and the custodian is often a stranger running a server. This experiment stacks that custodial risk under a second layer of exposure. You are trusting the mint not to betray you, and simultaneously trusting that no nearby stranger with a radio sniffer betrays you first. Two independent trust assumptions, both of which can fail, neither of which is disclosed in the breezy novelty of the demonstration. That is not a payment system. That is a trust-fall with two blindfolds. The Bearer Token as an Economic Object Step back and think about what bearer cash does to incentives, because this is where the design reveals its intended scale. The developer's own documentation recommends sending only gift-level amounts. Not because the tool is unfinished, but because its economics forbid anything else. An unencrypted, interceptable, non-real-time bearer transfer is a terrible way to move meaningful value and a perfectly fine way to move a novelty. The architecture and the recommended use case agree: this is designed for micro-payments, tips, curiosities. There is no arbitrage here, no yield, no reason for capital to flow in. The economic incentive structure is deliberately flat. This matters because it closes off the one story that would make the news dangerous. There is no token to pump. There is no vesting cliff to front-run. There is no TVL to chase. Whatever attention this draws is pure narrative attention, disconnected from any instrument that could absorb it. If this story appears during a sensitive period for privacy-payment sentiment, it can be meme-ified into a speculation theme, but there is no fundamental underneath it — just the afterglow of a clever trick. And this is where my long-standing irritation with this sector's storytelling returns. Liquidity fragmentation is not a real problem. It is a manufactured narrative that venture capital funds use to justify funding yet another product that solves a problem nobody had. The Find My experiment is not that kind of manufactured narrative — it is honest about being a novelty — but the reaction to it will be. Watch how quickly the think-pieces appear arguing that this proves the need for some new interoperability layer, some new bridging standard, some new fragmentation solution. It proves no such thing. It proves that a clever person with a two-dollar chip can move two bits per signal through a mesh network. That is the entire claim. Everything built on top of it will be narrative, not engineering. Against the Fragmentation Sermon Let me press on this, because it is the trap the market always falls into and I refuse to let it pass quietly. The reflexive response to any novel transport mechanism is to declare that the ecosystem needs a new abstraction. The Find My demonstration will be recruited into exactly this argument: look, another rail, another channel, another fragment — surely we need a unified layer to stitch it all together. This is backwards. The problem was never that Bitcoin payments lack rails. Lightning exists. Liquid exists. Cashu and Fedimint exist for privacy and small-value use. There is a mature, layered stack that handles daily payments, settlement, and private micro-transfers. The Find My path occupies none of these layers, and it does not need to, because none of them are missing. The people who will argue otherwise have a product to sell. That is the tell. When someone tells you a new fragmentation has appeared and therefore requires a new solution, check who benefits from the solution. Almost always, the fragmentation was never the disease. It was the sales pitch. What the Find My experiment actually demonstrates is the opposite of what the fragmentation merchants will claim. It demonstrates that the existing payment stack is so complete that a novelty transport has to reach into an unrelated consumer network to find anything new to say. It is a symptom of maturity, not a cry for more infrastructure. Why You Cannot Simply Encrypt It I want to linger on the encryption objection one more time, because it is the first refuge of every enthusiast and it deserves a full, unsentimental demolition. The instinct is: the developer just did not bother to encrypt the payload, and a better version would. This misunderstands the channel. Encryption does not create bandwidth. It consumes it. Every bit you spend on a nonce, a tag, a key identifier, or a signature is a bit you cannot spend on the token itself, and the token already strains against a two-bit-per-signal ceiling that requires a thousand broadcasts. You are trying to fit an encrypted envelope into a channel sized for a postcard's worth of entropy stretched over thousands of pulses. The deeper problem is the key exchange. Symmetric encryption requires both sides to share a secret. Getting them that secret requires either a pre-shared key — which defeats the purpose of a spontaneous transfer — or an asymmetric handshake, which requires bidirectional, low-latency, reliable communication. Find My is none of those things. It is unidirectional, high-latency, best-effort. You cannot complete a handshake through it any more than you can have a conversation through a series of postcards delivered by strangers over several days. The channel's geometry forbids it. So the plaintext broadcast is not an oversight. It is the channel telling you what it is. When the architecture reveals its own constraints this clearly, you should believe it. This is why I keep returning to the discipline of audit over enthusiasm. The 2022 collapse of Terra and Luna taught an entire market that algorithmic promises fail at exactly the moment the mechanism is stressed, and the mechanism here is stressed at the architectural level before a single dollar moves. The fragility is not hidden in an edge case. It is the front door. In the void, we found the edge no one else saw. That was true of the DeFi arbitrage I ran in 2020, when my team extracted value from spreads that most traders could not even see because they were staring at the wrong chart. But the edge was real because the inefficiency was real, and it survived stress. This Find My edge is real in a different way: it reveals a channel nobody was watching. But it does not survive stress, and a channel that fails under stress is a demonstration, not an edge. The Rebranding Problem There is a broader pattern this story sits inside, and it is worth naming. Bitcoin's brand is valuable, and therefore it gets borrowed constantly. Projects that are functionally Ethereum-native or protocol-agnostic wrap themselves in Bitcoin's name because that name carries ideological weight with a community that does not want to be sold something foreign. This experiment is a mild case — its author is a genuine Cashu developer, not a rebrander — but the reception will be a strong case. Expect the coverage to frame this as a Bitcoin story, a Bitcoin capability, a Bitcoin breakthrough. It is none of those things in any rigorous sense. It is a Cashu story and a Find My story that happens to carry a Bitcoin-backed token. The distinction matters because it is the difference between understanding the actual dependency graph and being sold a brand. The real Bitcoin Layer 2 work — the channels, the covenants, the careful engineering that has absorbed a decade of adversarial pressure — does not recognize this as part of itself. And it should not. A transmission side-channel through a closed consumer mesh is not Bitcoin infrastructure. It is a fascinating footnote to a different story. What I keep watching for is whether the market can hold two ideas at once: that this is a genuinely clever engineering demonstration, and that it is a genuinely unusable payment system. Most coverage will collapse them into one, and in the collapse, the useful signal — that a mesh network can be repurposed as a covert data channel — will be lost under the noise of a payment narrative that does not survive contact with the numbers. Audit the Soul, Then Audit the Contract The honest way to read this is as a study in why architecture beats intention. The developer did not intend to build something insecure; he intended to build something clever, and he succeeded at that. The insecurity is not a moral failing. It is the geometry of the channel and the nature of bearer cash, expressed through code. You can intend a private transfer and ship a public broadcast. You can intend a payment rail and ship a novelty. The gap between intention and architecture is where most of this industry's casualties live. I learned that in the Colombian Andes in 2022, after the Terra collapse drained me to the point where I withdrew from every trading group for three months and wrote a technical paper on the fragility of algorithmic stablecoins. Solitude clarified what noise obscured: that the mechanisms do not care about our hopes for them. A bearer token does not care that you meant to transfer it privately. A two-bit channel does not care that you wanted to send more. The mechanism simply does what its geometry dictates. So when I look at this experiment, I do not ask whether it is impressive. It is. I ask whether its architecture can survive an adversary, and the answer is no. The plaintext broadcast cannot survive a sniffer. The Apple dependency cannot survive a policy change. The mint trust cannot survive a dishonest operator. Three throats, all exposed, none guarded. The Contrarian Angle Here is where I part ways with both the cheerleaders and the dismissers, because both are reading the wrong layer. The cheerleaders see a Bitcoin payment breakthrough and miss that it is a transport trick. The dismissers see an absurd novelty and miss that it is a genuine discovery about a network nobody was auditing. Both are wrong for the same reason: they are arguing about whether this is a good payment system, when the interesting claim is that it is not a payment system at all. It is a covert channel. The most valuable thing this experiment demonstrates is that Apple's Find My mesh can be repurposed as a data channel outside its intended purpose, at extremely low bandwidth, with no authorization. That finding transcends Bitcoin. It belongs to the study of steganography and covert channels, and it should alarm anyone thinking about the security of crowd-sourced sensor networks, not just anyone thinking about payments. The payment narrative is a sideshow. The channel discovery is the headline, and almost nobody is writing that headline. There is also a delicious contradiction buried in the privacy pitch. Cashu's entire reason to exist is privacy — that is the value proposition. Yet this path requires the receiver to bind an Apple account to reconstruct the payload. A privacy-preserving payment that demands you tie your identity to the most surveilled consumer ecosystem on earth is not a privacy payment. It is a privacy payment wearing a tracking device. The value proposition contradicts the implementation at the exact point where it matters most. And then there is the deepest contrarian read of all: this may be best understood as a censorship-resistance prototype, not a payment product. Imagine an internet blackout — a natural disaster, a regime blocking connectivity. The romantic appeal of this mechanism is that it needs no internet, only nearby iPhones. That is a real and genuinely interesting direction. But romantic appeal is not robustness, and a mechanism that depends on a corporation's goodwill and a stranger's radio silence is not censorship-resistant. It is censorship-adjacent, borrowing the aesthetics of resistance without the architecture of it. The void has a floor, and this one is made of Apple's server policy. We bet on the pattern, not the hype. The pattern here is not adoption. The pattern is that novel channels get discovered, celebrated as products, and quietly abandoned when the numbers arrive. Two bits per signal is the number. Watch how fast the celebration fades when it does. What I Am Watching The forward-looking judgment is not about whether this tool ships — it will not, and it does not need to. It is about three signals that will tell us whether the interesting part of this story develops or dies. First, Apple's interface policy. If Apple tightens or blocks unofficial Find My access, this entire direction collapses overnight, and the collateral damage lands on the legitimate certified-accessory ecosystem. That is the single most consequential variable, and it has nothing to do with Bitcoin. Second, whether any real abuse case emerges. If this transmission path is ever used for money laundering or sanctions evasion, expect targeted regulatory attention on ecash mints and crowd-sourced transmission channels alike. The anonymity is a feature until a regulator decides it is a liability. Third, whether the underlying channel discovery attracts serious research — whether academic work on covert channels through consumer mesh networks picks this up, or whether it stays a footnote in a Bitcoin news cycle. The former would be the signal that the real story was noticed. The latter is the default. The ledger was clean, but the vision was fragile — and in this case, the fragility was never in the vision. It was in the physics. Two bits per signal, a bearer token in the open, a dependency on a landlord who never agreed to the deal. That is not a payment system waiting to mature. That is a channel telling us, with perfect honesty, exactly what it is. The only question left is whether we will listen to the numbers or to the narrative. We already know which one the market prefers. The interesting part is which one survives.