Stable Linux releases often lack crucial security fixes because backporting them to older code is a time-consuming and unrewarding task for human developers. This presents a significant security risk as most users prefer older, tested software versions. To address this, AI agents powered by Codex are being utilized to drastically reduce the time needed for backporting, from days to mere hours per fix. This allows human reviewers to focus on critical software like Ubuntu LTS, Fedora, and the Linux kernel. The AI system handles tasks such as identifying missing fixes, adapting patches, running tests, and preparing submission packets, while engineers review and submit them. This approach has already led to the delivery of fixes for 149 distinct CVEs into stable Linux kernel releases and several fixes for Fedora and Ubuntu packages. The backlog of unfixed CVEs is substantial, with thousands of known vulnerabilities present in stable kernel branches and distribution packages. Many users deliberately avoid updating to newer software versions to maintain stability, creating a dilemma between security and familiarity. The AI-assisted workflow makes tackling this neglected security work practical, enabling a more robust approach to patching the planet. By accelerating the backporting process, the project aims to eliminate the need to choose between stability and security in software.