Anthropic employee Evan Hubinger went viral on X on Wednesday after warning about AI extinction risk. According to Protos, his post drew 30 million views and said the chance of AI killing all humans within the next decade is greater than 10%.
Protos linked that comment to Hubinger’s earlier crypto industry work, noting that before his current role in AI safety, he worked at Ripple.
Hubinger’s post drew attention across crypto and AI circles
Commenting on Hubinger’s viral claim, Galaxy Research head Alex Thorn joked about his connection to Ripple and the XRP community. Protos added that XRP’s price today is still below its January 2018 high.
In the X post quoted by Protos, Hubinger wrote: “Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.”
Hubinger later clarified: “To be clear, as we say in our latest Risk Report, I think the risk from present models is low. What I am worried about is superintelligence arising from recursive self-improvement.”
Protos explained recursivity as the ability of AI to train and generate new AI without human input.
Resume lists Ripple software engineering work
Hubinger’s resume says he worked in Ripple’s Software Engineering department. Protos noted that years later, Ripple sought financing at up to a $50 billion valuation, with most of that valuation backed by the company’s corporate XRP holdings.
Although Hubinger joined Ripple while he was still in school, the description of his work goes well beyond routine intern tasks. His resume says he “worked on designing Interledger, a trustless system for cross-currency transactions between arbitrary agents.”
He also said he “wrote a tool to do cryptographically secure wallets for financial institutions.”
His LinkedIn profile adds work on the build process for rippled, the server software that powers parts of the XRP Ledger.
Now leading Anthropic’s alignment stress-testing team
Protos said Hubinger now leads Anthropic’s Alignment stress-testing team. Its job is to prevent alignment failures by AI models, described in the article as industry language for AI agents pursuing secret goals outside the scope specified by humans.
Other Ripple-linked figures have also discussed AI risk
The article also mentioned that Ripple CTO Emeritus David Schwartz has argued that AI safety rules could threaten freedom of speech.
Former Ripple Vice President Emi Yoshikawa wrote in August that AI extinction risk was becoming an everyday conversation topic in Silicon Valley.
The report was published by Protos under the headline “Anthropic’s AI doomsayer worked at Ripple.”