Oct 20, 2026 · 1h 32m
AI red teams expose Bitcoin’s custody weak points
The Fight to Defend Bitcoin From AI | Rob Hamilton
The episode shows how AI-assisted security testing is changing Bitcoin’s threat model while making user competence and custody design more consequential.
- 1AI-assisted review scanned roughly 650 Bitcoin-related codebases and uncovered thousands of potential vulnerabilities.
- 2The Coldcard incident exposed a vendor-specific failure, not a compromise of Bitcoin or Bitcoin Core.
- 3Safer custody depends on tested procedures, resilient backups, and trade-offs among multisig, recovery, sovereignty, and trust.
Don't miss
Hamilton explains how a Coldcard vulnerability led to an AI-assisted volunteer effort scanning roughly 650 Bitcoin-related codebases and finding thousands of potential issues.
The brief
Rob Hamilton recounts how AI helped investigate a serious Coldcard vulnerability, then catalyzed a volunteer red-team effort that scanned roughly 650 Bitcoin-related codebases.
The findings—thousands of potential issues, including critical and high-severity bugs—support a blunt conclusion: audits reduce risk, but no software can be assumed secure.
The conversation separates Bitcoin’s core from its surrounding infrastructure: Bitcoin and Bitcoin Core were not seriously compromised, while dependent projects exposed a broader attack surface.
Hamilton turns from diagnosis to practice, urging test transactions, verified backups, dry runs, and custody designs that match a user’s competence and tolerance for risk.
The sharpest policy argument arrives late: restricting open-weight AI may hinder defensive research even as attackers, banks, utilities, and hospitals face the same accelerating arms race.
Books & mentions
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.