Malcolm Reeves: February 2024. Someone receives fifty thousand dollars in Bitcoin. Three blocks have been mined on top of it. Their accountant asks: is it final? And the honest answer — the answer buried in Satoshi Nakamoto's 2008 whitepaper — is: define final.
Cole Brennan: Huh. Yeah that's — that's a little unsettling actually.
Malcolm Reeves: Because Bitcoin's immutability isn't a cryptographic absolute. It's an economic one. The chained block structure — SHA-256 hashing, each block's header locking in the hash of the one before — that's what makes rewriting history detectable. But detectable isn't the same as impossible.
Cole Brennan: Right, because — wait, so the reason it's hard to change is that any change to a block's data produces a completely different 256-bit SHA-256 output, which then breaks every block after it, and you'd have to redo all the proof-of-work on all of them?
Malcolm Reeves: Every subsequent block. And the further back the transaction, the more accumulated proof-of-work you're redoing. The cost doesn't add — it compounds.
Cole Brennan: So immutability isn't a wall. It's more like — I mean, it's a price tag that keeps getting bigger the longer you wait. And Satoshi apparently just... knew that. Built it that way on purpose.
Malcolm Reeves: And the mechanism is specific. A miner iterates through a nonce — just a number, really — until SHA-256 spits out a hash that falls below the network's difficulty target. That's the work. That's what costs electricity. Every block is essentially a lottery that burns real energy to win.
Cole Brennan: Wait, so the security is literally — it's electricity? Like, the thing protecting my transaction is that someone had to burn power to find that nonce?
Malcolm Reeves: Sunk cost, yes. And now here's where it compounds — an attacker going back, say, ten blocks doesn't redo ten blocks of work. They redo ten blocks *while the honest chain keeps growing ahead of them.* Every second they fall further behind. The cost isn't additive — it accelerates.
Cole Brennan: That's — I mean, nobody talks about that part. That's the thing that actually matters.
Malcolm Reeves: Eric Budish — economist at University of Chicago — he formalizes this. His argument is that security holds only as long as the flow cost of mining stays large relative to what you'd gain from a successful attack. That's an equilibrium condition. Not a law of physics.
Cole Brennan: Which means — okay, so back to that accountant. Three blocks. Is that actually safe? Like, everyone says six confirmations, but is that — where does six even come from?
Malcolm Reeves: There is no canonical answer. Six blocks against a miner running, say, Antpool-scale hashrate is completely different from six blocks against a nation-state. Confirmation depth is just a proxy — a rough one. It depends entirely on who's attacking and what they stand to gain.
Cole Brennan: Okay but — and this is the thing I can't get past — this isn't hypothetical anymore. 51% attacks have actually happened. On real chains. Not Bitcoin, but like, smaller proof-of-work chains got rewritten. Actual transaction history, gone. So when we say 'immutable,' we're describing a mechanic that has demonstrably broken on live networks.
Malcolm Reeves: And then consider — the Ethereum DAO fork, 2016. That's not a hashrate attack. That's governance. The community decided, collectively, to rewrite history to reverse a theft. They just... chose to.
Cole Brennan: Wait — so that's a totally different mechanism breaking the same promise.
Malcolm Reeves: Which is why some node-operator research lands on this term — practical immutability. The idea that rewrites are possible, they have occurred, but they require either overwhelming hashrate or overwhelming social consensus. That's conditional, not absolute.
Cole Brennan: So back to the fifty thousand dollar merchant — three confirmations — the answer actually depends on whether you're Foundry USA or Antpool sitting on a giant mining operation, or you're just... some guy. I mean, for Bitcoin's main chain, the hashrate concentration at that industrial scale is exactly what makes it resistant. But that's the thing, right? The security isn't distributed magic, it's those specific entities holding the line.
Malcolm Reeves: And 'immutable' isn't wrong — it describes the dominant trajectory. But it's not complete. The word is doing more work than the mechanism actually delivers.
Malcolm Reeves: And the part I keep turning over — not the DAO fork, not the smaller chains — it's what happens to Bitcoin itself, long-run. The block subsidy. It halves on a schedule. Eventually it approaches zero. And when it does, Budish's equilibrium condition becomes entirely dependent on transaction fees filling that gap. The flow cost of mining has to stay high enough to make attacks irrational. But if miner revenue collapses because fees don't scale to replace the subsidy... the cost of rewriting older transactions falls with it. Transactions that feel permanent today — I mean, that's the thing nobody has resolved. Whether a fee market actually sustains that security. Nobody knows.
Cole Brennan: And nobody really knows if fees fill that gap.
Malcolm Reeves: No. They don't. And that's — I mean, that's where the architecture Satoshi built in 2008 just... runs out of answer. The proof-of-work is sound. The chaining is sound. But the incentive that powers all of it? That's a question left open.
