But that’s nothing compared to the latest round of updates. There are an unprecedented 122 security vulnerabilities fixed in iOS 27. Think that’s a lot? The macOS 27 update has 204 fixes listed on Apple’s Security Updates page. Even watchOS 27 adds about 90 fixes.
Thanks to AI coding agents, this is the new normal. You can expect dozens of security vulnerabilities to be fixed every couple of months, at least for a while.
Why is this happening?
You can thank, and blame, AI for this.
AI coding agents have gotten good. Very good. Not only has “vibe coding” become commonplace, but the best AI models can analyze and test very large code blocks and databases. And they find things that people miss.
A few months ago, we wrote about Project Glasswing, a collaboration between Anthropic and a handful of other companies (including Apple) to provide those companies’ security researchers with access to the company’s most advanced AI model, in the hopes of finding vulnerabilities before the models become public and all the bad actors get their hands on them.
The world we live in right now is one of security acceleration. Hackers and malicious groups can use advanced AI tools to find unique vulnerabilities that can, sometimes in combination with each other, compromise systems like never before. The sort of resources and research once only available to nation-states and the largest tech corporations are now broadly available.
