TL;DR:
Vitalik Buterin has confirmed that AI-assisted analysis identified his anonymous contribution to Ethereum’s research process, ending a two-week challenge meant to test whether tools can pierce online anonymity. The Ethereum co-founder had hidden a document of “medium importance” under another name, then asked the public to find it. Co-Invest CEO Franklyn Wang’s submission pointed to an anonymous rewrite of EIP-7503 by analyzing how the document explained mathematical and technical ideas. The surprise was not that prose style leaked, but that reasoning patterns did, even after Buterin tried to mask the writing.
Using AI, I doxxed Vitalik in about 2 hours.
He confirmed it — "we have a winner." The doc was an anonymous EIP-7503 rewrite he'd hidden by writing it in Chinese and machine-translating it. Airtight prose. Still found him — because I ran it through @coinvestai, and the tell… https://t.co/5jn8AvAJ8b
— Franklyn Wang (@frank_liquid) July 6, 2026
Buterin later revealed that he wrote the anonymous EIP-7503 rewrite in Chinese, translated it into English with Qwen 2.5 and manually corrected the output to disguise his prose. Wang said the clue was not the words, but the reasoning. His system ranked Buterin as the most likely author with about 20% confidence, roughly 10 times higher than the next candidate, after reviewing 27 documents. The experiment exposed an uncomfortable authorship signal, where intellectual habits around math and algorithm explanation survived language switching, translation and manual editing.
The test landed inside a broader debate over whether pseudonymous technical work can remain durable in an AI-heavy internet. Buterin’s June 22 challenge referenced claims that AI text analysis could make online anonymity untenable. A February paper from ETH Zurich and Anthropic argued that large language models have made online deanonymization practical at scale by extracting identity clues from unstructured text, searching for possible matches and reasoning over likely candidates. Ethereum’s open research culture now has a privacy problem, because contributors often rely on pseudonyms to separate identity from ideas.
The result has limits. Lighter CEO Vladimir Novakovski said he worked with Wang in 2023 on a GPT-4-based effort to identify Bitcoin creator Satoshi Nakamoto by matching writing style in cryptography research, but that project did not produce a high-confidence result. That comparison matters because Buterin’s case shows progress, not omniscience. AI identification remains probabilistic, not magical, yet even a partial signal can reshape how open-source communities handle drafts, authorship, governance and anonymity. For Ethereum, the lesson is clear: anonymous input may still be possible, but hiding style alone no longer looks sufficient.
