In November 2008, Ray Dillinger reviewed the blockchain portion of Bitcoin’s source code, splitting the pre-release audit with Hal Finney, who took the scripting language, while Satoshi fielded their questions and asked his own. He’d built something related himself thirteen years earlier — a block-chain-based digital-cash design in a 1995 graduate networking paper — and found Satoshi’s cryptographic code tight. His two objections were structural, not stylistic: powerful actors could attack the early network, and bandwidth would strain as it grew.
The full essay is here. A few weeks after the review, Satoshi launched Bitcoin with Hal’s help, and Dillinger stepped aside. His main reason was history rather than code: he had followed digital-cash protocols since before 1995, had watched literally scores of launches fail, and saw no reason to expect a nonzero valuation. Read from 2017, that is the call the essay’s title points at, a reasonable judgment that missed what was actually starting.
What convinced him Satoshi wasn’t running a con was the design itself. Most of the people behind early digital-cash systems who went to jail, he wrote, had done it by abusing a position of being Trusted, and Bitcoin had no Trusted role at all:
Ray Dillinger (LinkedIn post, September 20, 2017)He built a highway with no toll bridge. People could use Bitcoin without creating any obligation to pay him anything ever. He wasn’t selling coins, he was giving them away for solving hashes. He reserved nothing for himself.
The line lands harder set against Dillinger’s own reasoning for stepping back: he judged the highway not worth building on, then watched everyone else build toll bridges on it anyway.
He added the coins Satoshi never cashed: “As far as we can tell he mined approximately a million Bitcoins and has never sold a single one of them.” Against that standard he set what came after: more than three thousand short-lived cryptocurrencies on the same protocol, launched by people of whom, in his words, “at least three quarters know that what they’re doing is ripping people off,” by means he “can’t really distinguish from blatant stock price manipulation and insider trading.” Selling all the coins at launch had acquired a new name, ICO, and he feared that the genuinely good ideas among the scams would burn with them. On the page that counts altcoins and compares their designs, his estimate sits beside the Satis Group’s later finding that roughly 78% of 2017-era ICOs were identified scams, a narrower population than the altcoins he was counting.
This piece was published on LinkedIn on September 20, 2017, and reached the front page of Hacker News on September 21, 2017, receiving 329 points and 86 comments.
Dillinger sketches the division of review labour with Hal Finney here; the deeper technical substance of that audit — the floating-point accounting discovery and the satoshi-precision design choice — is drawn out at length in his 2018 interview with Tim Swanson.
Dillinger’s claim here — that Satoshi mined roughly a million bitcoins and never sold any of them — is corroborated and quantified five years later by Jameson Lopp’s 2022 data analysis, which finds Satoshi’s actual mining output was only about half of what his hardware could have produced running at full capacity.
