From Binance to PHP2Shell, AI Is Forcing a Rethink of Web Security
A fresh security incident involving Binance has added to a growing list of major breaches across the crypto sector, including earlier cases tied to Bybit and Bitget. The common thread, as outlined in the source article, is that attackers are no longer focused only on breaking cryptography or stealing private keys directly. They are moving across the broader chain of trust: employee devices, third-party security tools, frontend infrastructure, DevOps environments, and withdrawal authorization flows. The article also links that shift to the rise of AI. It points to the recently discussed PHP2Shell vulnerabilities, two high-severity flaws in PHP’s core engine, as an example of how AI can help uncover deep, long-hidden software weaknesses and shorten the time between disclosure and exploitation. Against that backdrop, Ethereum co-founder Vitalik Buterin argued on Oct. 22 that the long-run end state of cybersecurity may favor defenders, but the transition period will be extremely difficult. His proposed path centers on changing how software is built rather than only racing to find bugs first. The piece highlights two ideas he emphasized: formal verification, which uses mathematical proofs to show code meets security requirements, and minimizing the trusted computing base, or TCB, so fewer components sit inside the critical trust boundary. It also places those ideas inside Buterin’s broader d/acc, or Defensive Accelerationism, framework, which prioritizes technologies that strengthen defense and reduce systemic risk.
