Ripple Alum Warns of Superintelligent AI Apocalypse Risk

Evan Hubinger, a researcher at Anthropic, recently made headlines by predicting that the end of the world could occur within a decade due to superintelligent AI. In a tweet, he stated that he believes the risk is above 10% and even commented on his own previous work in the field. Interestingly, before joining Anthropic's Alignment stress-testing team, Hubinger worked at Ripple, one of the oldest crypto companies.

According to his resume, Hubinger was part of Ripple's Software Engineering department and claims to have designed Interledger, a trustless system for cross-currency transactions. He also wrote a tool for cryptographically secure wallets and contributed to the build process for rippled, server software powering parts of the $XRP Ledger.

Galaxy Research's Alex Thorn used his connection to Ripple and the $XRP community to poke fun at Hubinger's claims, highlighting that the price of $XRP remains lower than its January 2018 high. However, others have taken Hubinger's warning seriously, with Jacob commenting that they genuinely believe AI could kill all humans.

Hubinger later clarified his statement, explaining that he is worried about superintelligence arising from recursive self-improvement, which refers to the ability of AI to train and generate new AI without human input. This concern has been echoed by others in Silicon Valley, with former Ripple Vice President Emi Yoshikawa writing that AI extinction risk is becoming an everyday conversation topic.