PHP 两个十年漏洞由 GPT-5.6 Sol 与 Gemini 4 Argon 协助发现

GP @Graham_dePenros原推发布于 2026.10.22

在 X 上打开原推

🚨 🌐 One of the biggest internet security threats in years has exposed just how much AI is changing cybersecurity. Two critical flaws were discovered in PHP, the programming language behind roughly 70% of websites. The bugs had existed for over a decade and could allow hackers to take control of affected servers and access sensitive data. After details leaked ahead of the planned announcement, hacking demonstrations spread online, and Cloudflare reported early attacks. Dutch government websites went offline as a precaution, while security agencies urged organisations to update immediately. AI played a major role. OpenAI’s GPT-5.6 Sol and Google’s Gemini 4 Argon helped researchers uncover the flaws and develop working exploits far faster than traditional methods. PHP has released emergency fixes, but websites still need to install them. The bigger concern is what comes next. As AI improves, finding and exploiting complex security flaws could become faster and easier, giving both security researchers and hackers capabilities that once required months of specialist work. What are your thoughts on this? 🤔💬