AI Labels Mt. Gox Code “Critically Insecure”

[ed]

Former Mt. Gox CEO Mark Karpelès fed the exchange’s 2011 code into Claude AI, which called it a “feature-rich but critically insecure” Bitcoin platform.
Claude traced the June 2011 hack to weak passwords and poor internal security. Karpelès admitted he never reviewed the code before buying Mt. Gox — a mistake he now calls a “hard lesson in due diligence.”

[Nahid Hossen]

AI audits showing Mt. Gox’s weak code prove how bad things were back then.

[EmTeam]

Interesting to see how AI is now helping uncover past crypto vulnerabilities.

[Swipe]

Wild seeing AI dig through crypto history like this — turning old code into new lessons for today’s builders