The main technological intrigue of recent weeks around Ethereum co-founder Vitalik Buterin has received an unexpected continuation. For those who missed the essence of the dispute, on June 22 Buterin challenged neural networks to test the popular thesis about the complete loss of privacy on the internet.
He admitted that during the current decade he had published several Ethereum documents under someone else’s name and invited AI to identify them by his individual writing style.
13 days.
So far no one has found it.
My only hint is that I would encourage people to somewhat broaden their search; I’ve seen quite a few searches and AI scripts that fail to include categories of documents that really should be included. https://t.co/ClBx2SEvhP
— vitalik.eth (@VitalikButerin) July 5, 2026
Thirteen days later, Buterin summarized the interim results of the experiment. Neither any researcher nor any advanced AI script managed to find the text. The failure of modern deanonymization algorithms prompted Buterin, 13 days after launching the quest, to release a key hint and point to a systemic error made by the search bots.
What’s hidden in Ethereum documents, and where did AI stumble?
Since the material in question is of importance for the Ethereum ecosystem, the hidden text could be a technical proposal for improving the network, an analysis of cryptographic protocols, a mathematical study, or a scaling concept. According to Buterin’s own estimate, there are between 200 and 2,000 documents of similar scale online, which significantly narrowed the sample for analysis.
However, automation lost to humans because of the banal limitations of its settings. According to the Ethereum co-founder, AI scripts stumbled over their own rigid filters.
In his latest statement, Buterin recommended that researchers expand the search area. He noted that he had seen many search attempts and AI scripts that simply ignored entire categories of documents, even though they should be taken into account.
Neural networks became hostages of standard templates, scanning only official blogs or technical specifications while completely missing other layers of publications.
This experiment clearly confirms why Buterin consistently maintains his AGI skepticism and points to the vulnerability of modern AI models when facing non-standard tasks. The failure of AI in the first round proves that the creation of an all-powerful superintelligence is still far away, while human secrecy remains stronger than algorithms.