Hi, we’ve been building NØNOS, an experimental OS aimed at “clean sessions” where the system is designed to minimize persistence and keep trust explicit. The public repo right now is the kernel/boot path + early desktop/QEMU flow.

What it is: a Rust microkernel with capability-based access control and a “runs entirely in RAM / zero persistent state” model (by design).

How to try it (QEMU):

Follow the repo README to build/run, or use the docs “Running in QEMU”. The kernel repo includes expected boot output and troubleshooting notes.

What I’d love feedback on:

Whether the “zero-state” model is compelling vs. existing approaches (live OSes, VMs, immutable systems).

Threat model gaps / places where the design assumptions are shaky.

What would make this actually easy to test safely (prebuilt images, reproducible builds, hardware support priorities).

If you want deeper design context, see the whitepaper.