What Should a Cloud Native OS Look Like?

Series

1. What Are Special-Purpose Operating Systems in the Cloud Native World?

   Exploring the evolution of operating systems in the Cloud Native landscape, focusing on special-purpose operating systems (SPOS) and their role in modern infrastructure.

2. A cloud-native skill that doesn't die when you change jobs Coming soon

   I've been doing Linux for a long time. Long enough that I've watched people walk into a new job, sit down at a new terminal, and discover that half of what they knew at their last job doesn't apply here.

3. One mental model, every customer Coming soon

   If you've spent any time doing solutions architecture, you already know this feeling.

4. Don't pick a fight with security. Meet them where they are. Coming soon

   I've watched this scene play out enough times that I'm going to describe it and you're going to recognize yourself in it.

5. Why 'apt upgrade' is incompatible with a platform contract Coming soon

   Why package-manager upgrades break the replaceable, measurable, and rollback-safe contract a cloud-native OS needs.

6. Open Source Isn't Open Governance

   Why license and activity are not enough when evaluating open source projects, and why governance matters even more at the operating-system layer.

7. Minimalism has two schools. Pick carefully. Coming soon

   "Minimal" is one of the most overloaded words in the operating system conversation right now. Every cloud-native OS project I can think of calls itself minimal. They can't all mean the same thing by it, because they're doing very different things with very different results. And I think the word is papering over a choice that platform teams should be making consciously, not by accident.

8. Measured boot isn't just a security feature. It's a platform contract. Coming soon

   When I explain measured boot to engineers, they usually understand the mechanics pretty quickly. TPM, PCR banks, Secure Boot, UKI, systemd-boot - you can walk through all of that in a blog post, and when you're done, the engineer has a reasonable mental model of what's happening.

9. The honest trade-offs of a cloud-native OS Coming soon

   I've spent the last several posts making the argument for cloud-native operating systems, image-based lifecycles, and the specific bets Kairos and Hadron make. I believe everything I wrote. I wouldn't work on this project if I didn't.

10. The cloud-native OS landscape in 2026, and the axes that actually matter Coming soon

    If you search for "cloud-native OS" or "container-optimized Linux" right now, you'll find a handful of comparison articles. Most of them are from 2022 or 2023, when the conversation looked different - k3OS was still alive, Hadron didn't exist, several projects have since changed stewardship, and the governance picture in particular looks very different than it did then.

11. Eight questions I ask any OS that wants to sit under my platform Coming soon

    At the end of the talk this post is based on, I give the audience a checklist. Eight questions to ask any operating system that wants to sit under a cloud-native platform. The checklist is the thing I most want people to take home, screenshot, and use on their next evaluation.

12. "More secure" is a spectrum: how to actually evaluate security claims in a cloud-native OS Coming soon

    A few months ago I was reading marketing copy for a cloud-native operating system and I counted six distinct "more secure" claims on a single page. No SSH. Minimal attack surface. Purpose-built init. Memory-safe language. Read-only filesystem. Encrypted at rest. Every one of those claims was technically true, and every one of them was doing work it hadn't earned.

Talks