Grok-Powered Stylometry Study Ranks Elon Musk 30th of 66 in Satoshi Writing Test

Three Writing Fingerprints, Three Failed Tests

An open-source research project has run one of crypto's oldest conspiracy theories through a statistical filter, and Elon Musk does not survive it. The satoshi-stylometry repository, published on GitHub, paired Grok, the AI chatbot built by Musk's own company xAI, with stylometry, the statistical study of writing habits that researchers use to compare or identify authors. Such analysis measures habits writers rarely notice in themselves: spelling choices, punctuation spacing and how often common words appear. Because those patterns are hard to fake deliberately, they serve as a standard tool wherever authorship is disputed. The target here was the persistent claim that Musk is Satoshi Nakamoto, the pseudonymous creator of Bitcoin, a theory that gained traction after a former SpaceX intern floated it in 2017, despite Musk's repeated denials. The researcher tested the three fingerprints supporters cite most often. Double spacing came first. Satoshi placed two spaces after a period 95.3% of the time, while Musk's raw X posts contained not a single instance of the habit. British spelling was the second marker. Satoshi's writing was mixed, with British forms appearing in 32.7% of cases; Musk used them in only 0.65% of his. The third test measured function-word similarity, where Musk ranked 30th of 66 writers, statistically no closer to Satoshi than a typical control author. There is a certain irony in the tooling, since Grok produced the evidence against a candidacy promoted on the back of its own builder's fame. The dataset also undercuts the idea that Musk once wrote like a Briton, because his earliest testable texts, dating from 2006, already use American spelling. To give the negative result weight, the author validated the method itself: it attributed Satoshi's own text samples correctly in 27 of 30 attempts.

Musk's 2017 Denial and Closer Names

Musk has denied the story himself, and in blunt terms. In a reply posted on November 28, 2017, he wrote that the claim was “Not true. A friend sent me part of a BTC a few years, but I don’t know where it is.” That message, preserved on his X account, remains his clearest public statement on the question. The new data adds texture around the edges of the debate. Small quirks separate the two writers as well: Satoshi wrote “e-mail” with a hyphen 17 times against three uses of the unhyphenated form, while Musk wrote “email” 65 times and never once used the hyphen. Every count is documented in the repository, so readers can check the arithmetic themselves. The same tests were run on Brad Mills, a Bitcoin advocate whose posts have been swept into the speculation, and his writing ranked 52nd of 66, so the data does not support his candidacy either. The comparison set produced one genuine surprise: cypherpunk Hal Finney scored closer to Satoshi’s style than Musk did, while Nick Szabo scored further away. That ordering runs against much of the online speculation of the past decade, which has tended to favor Szabo. The author is careful about limits. Satoshi wrote emails and forum posts, while Musk mostly publishes short social media posts, a genre gap that distorts any direct comparison. Smartphones often collapse double spaces automatically, which weakens the spacing test, and the study located only eight Musk spelling examples from before 2013 to work with. The repository’s README describes the exercise as a hobby-scale project, not forensic linguistics, and therefore as work that cannot prove or disprove who Satoshi really is.

@elonmusk · X post

Preserved on his X account.

View on X

Where the Evidence Stops

Measured against the repository itself, the exercise matters as much for its method as for its verdict. A 90% self-attribution rate gives the comparison statistical weight that earlier Satoshi theories never had, yet the limits the author documents, from the genre gap to the eight pre-2013 spelling samples, cap what any single test can settle. The one name the data favors over Musk is Hal Finney, and even that margin is modest. Whoever the creator was, the designer of Bitcoin’s issuance rules and its Bitcoin Halving schedule still has no confirmed name, and a hobby-scale study will not change that. What it does show is that open AI tooling can turn an internet legend into a testable dataset, a practical case of blockchain and AI synergy that anyone can audit line by line.