Coldcard Flaw’s AI Discovery Cost Was About $2, Dragonfly’s Qureshi Says

The discovery of a vulnerability in the Coldcard hardware wallet highlights how artificial intelligence is changing the economics of crypto security, according to Dragonfly Managing Partner Haseeb Qureshi, who said product safety is increasingly tied to how much companies invest in AI-powered security testing.

Qureshi proposed a new metric, called Cost of Discovery, to estimate how inexpensively a frontier AI model can reproduce a software vulnerability. Applying the concept to the Coldcard flaw, he estimated the discovery cost at roughly $2.

He also addressed claims that Anthropic’s Claude Code identified the vulnerability in about eight minutes, suggesting web search capabilities may have contributed to the result. By comparison, Qureshi said an offline test using the GLM model reproduced the same issue in approximately 20 minutes.

Source: MTS Live