Bitcoin AI Security: 6,700 Findings, Unconfirmed Vulnerabilities

An AI-assisted security campaign found 6,700 issues across 425 Bitcoin projects in 55 hours. Of these, 1,029 were labeled high or critical. Here's the thing: the campaign hasn't revealed how many of these findings were confirmed vulnerabilities. This means we don't know the actual security impact. The report highlights the volume of potential issues, not how many were real or fixed. There are no clear definitions for the severity labels, no false-positive rates, and no fix rates. This makes it impossible to determine how many alerts became confirmed vulnerabilities or led to patches. What's interesting is that an AI system can quickly identify a large number of potential issues. However, human experts are still essential for verifying, reproducing, and disclosing these findings. The models do the broad searching, while specialists shape prompts and judge the reports. The bottom line is that while AI can rapidly scan for issues, human expertise remains crucial for determining what truly matters for security.