#Anthropic模型被曝发现政府系统漏洞 | If AI’s security testing capabilities continue to keep breaking through, do you think this is good news (stronger security audit tools) or bad news (existing crypto systems could be compromised) for the crypto industry?

Anthropic’s models can autonomously discover large numbers of zero-day vulnerabilities—this is no longer science fiction. AI’s offensive and defensive capabilities in the security arena are improving exponentially.

For the crypto industry, I believe it’s a net positive. Smart contract auditing, formal verification, and fuzzing are areas that have long depended heavily on human effort and experience. AI can significantly lower the barriers and costs. In the future, before a project goes live, running millions of attack simulations with AI will be far more reliable than relying on manual code review.

The bad-news logic is “AI can break existing systems,” but that’s precisely why we must apply AI to the defensive side faster. Quantum computing was once also described as a “cryptography killer,” yet it ended up driving the development of post-quantum cryptography. AI security breakthroughs will force the crypto industry to raise security standards—not cause the entire industry to collapse overnight.

Old degen’s conclusion: Embrace AI security tools, don’t fear them. Those projects still operating on the mindset that “manual auditing is enough” are the truly dangerous ones. The public chains and protocols that can survive in the future will be the ones with “AI-native security.”