Why did a flaw in PHP hide for over ten years?
Here is a story that sounds made up. A programming language called PHP quietly runs a huge slice of the web: online shops, blogs, and a lot of the site-building kits people use without ever knowing the name. This week it emerged that two critical flaws have been sitting inside it since version 5.4. That is more than a decade of releases. With the right trick, an attacker could take over a server without needing a password.
One of the firms involved estimates PHP sits under roughly 70% of all websites. That is their estimate, not a census, but even half of it would be enormous.
So how does that stay hidden?
Not because nobody looked. This is open code, read by thousands of people over the years. The real reason is much more boring, and much more interesting: it was never worth the money.
The bugs are the nasty kind that live in how the system handles memory when something goes badly wrong. Smelling that something is off is one thing. Turning it into an attack that works across different versions, settings and apps is months of grinding work by very good people. Nobody paid that bill, so nobody found out.
Hold on. Ten years of nothing, and then a few weeks of AI work cracks it. Doesn't that just mean the AI is overhyped, or the code was never that bad?
Fair question, and no. By one security firm's own account, a single researcher with modern models did in about a week what a world-class team would have needed months for. The flaws were real, confirmed and patched on 21 October. What changed was not the code. What changed was the price of doing the hard part.
Think about what that means. Every old, dusty, important piece of software now has a new price tag on its secrets. The ones that were safe only because digging was expensive are not safe any more.