Coldcard Ships Firmware After $114M Bitcoin Theft
The patch tightens the RNG and signing flow, but seeds from affected firmware between 2021 and July 2026 still need replacement. AI-assisted auditing is the broader industry signal here.
Every Zipp story tagged #AISecurity, newest first.
The patch tightens the RNG and signing flow, but seeds from affected firmware between 2021 and July 2026 still need replacement. AI-assisted auditing is the broader industry signal here.
The figures cannot be independently verified, but they are the first concrete metrics a major venue has put on a claim bitcoin developers have been making all month.
Security-driven wallet migrations can mimic selling pressure, making bearish on-chain readings a poor standalone signal while the AI findings await validation.
The founding argument is a Hugging Face breach in which closed AI tools reportedly blocked the forensic response; the alliance open-sources the stack defenders say they need to actually inspect…
The consortium signals that frontier-model providers, chip vendors, and cloud platforms now treat open-source AI security as a shared engineering problem worth coordinating on, not competing over.
The incident shows a capable model with guardrails off can chain unknown flaws, stolen credentials, and production-side weaknesses end-to-end, which is exactly the shape a serious crypto exploit…
The Foundation is publishing a security report that treats AI as an audit accelerant, not a replacement for formal verification and human reviewers.
The headline is the bug itself, but the takeaway from the Foundation's Protocol Security Team is bigger: AI-driven triage is now doing the work human reviewers used to absorb manually.
If basic smart contract reviews drop from weeks-long engagements to on-demand calls, the standard of care shifts too — and skipping one could read as negligence.
A single auditor report used to certify a DeFi protocol. As AI gets sharper at finding bugs, multi-agent systems that cross-check code are becoming the new floor — not a nice-to-have.
Buterin's pitch lands as the industry grapples with smart-contract exploit losses running into the billions — automated proof of correctness is the part AI may actually get right.