Ethereum Hard-Forks to Reverse The DAO HackEthereum Hard-Forks to Reverse The DAO Hack

This is a Bitcoin history account, and normally that means we stay in our own chain's story. Today we're making an exception, because ten years ago Ethereum ran the experiment Bitcoin has always refused to run. It stole back stolen money by rewriting its own ledger, and it did so in full public view, with a vote, a deadline, and a name for the philosophy on the losing side. Bitcoin has never done this, not for Mt. Gox, not for any exchange collapse since, no matter how large the loss. The DAO fork is the clearest real-world test of why that refusal is treated as load-bearing rather than stubborn.

The DAO and the ExploitThe DAO and the Exploit

The DAO was a decentralized venture fund built on top of Ethereum. Anyone could buy in with ether and vote on which projects the pooled funds would back, a structure that made it one of the more ambitious experiments the smart contract world had tried at the time. In June 2016 an attacker found a flaw in how the DAO's code handled withdrawals. A function could be made to call itself recursively before the contract updated the caller's balance, letting the attacker pull ether out of the fund far more times than they were entitled to and route it into a child copy of the DAO under their control.

The theft came to roughly sixty million dollars at the time. But the DAO's own rules included a waiting period, about 27 days, before the attacker's child contract could actually release those funds. That window is what turned a bad week into a month-long crisis for Ethereum's core developers and community: they had a hard deadline to decide whether to do nothing, attempt some kind of countermeasure, or intervene directly at the protocol level.

The Hard Fork Decision and ExecutionThe Hard Fork Decision and Execution

The option that won out was a hard fork. Rather than accept the theft as final, Ethereum's developers proposed rewriting the chain's rules so the drained funds would move into a new contract built to do one thing: let the original token holders withdraw ether at the DAO's original exchange rate, as though the theft had never settled. It was framed by supporters as a rescue operation for thousands of ordinary participants. It was also, unavoidably, an edit to a ledger that had been marketed as immutable.

The fork executed on July 20, 2016. As CoinDesk reported at the time, the China-based mining pool BW.com mined the new chain's first block, moving the stolen ether into the withdrawal-only contract. Christoph Jentzsch, the Slock.it co-founder who had authored most of the DAO's original code, wrote afterward that he'd witnessed "remarkable feedback and support from many different sides" during the scramble to get the fork built and shipped in time.

Ethereum Classic and the "Code Is Law" SplitEthereum Classic and the "Code Is Law" Split

Not everyone accepted the fork as legitimate. A minority of miners and users rejected it outright, arguing that a blockchain's history should never be altered regardless of the circumstances, a position that got shortened into the phrase "code is law." That minority kept validating the original, unaltered chain where the attacker's funds were never clawed back. That holdout chain became Ethereum Classic. The forked chain, the one that reversed the theft, became the Ethereum that exchanges, developers, and most of the public now simply call Ethereum.

A month later, Ethereum Classic was still running, still had real value trading against it, and had not collapsed the way some fork skeptics predicted. In Bitcoin circles that outcome got read less as a footnote and more as a data point: a chain that refused to reverse a hack could survive, and a chain that did reverse one had just demonstrated, in the most public way possible, that its immutability was negotiable under pressure.

What This Says About BitcoinWhat This Says About Bitcoin

Bitcoin has absorbed its own share of catastrophic losses, Mt. Gox chief among them, and has never once forked to undo one, no matter how sympathetic the victims or how large the sum. That stance isn't an accident or an oversight. It's treated inside Bitcoin culture as close to the whole point: a ledger that can be edited by consensus whenever the stakes get high enough isn't really a ledger, it's a database with extra steps. The DAO fork didn't invent that argument, but it gave it a real-world control group. Ten years on, it's still the case study Bitcoin people reach for first.

Part of an ongoing series on Bitcoin history. This event falls on July 20, 2016.