GoKawiilUnikernels regain interest as AI coding tools ease development friction
Justin Cormack, a former MirageOS and Unikernel Systems developer, discussed in a podcast conversation why unikernels—operating systems built as a single application with no separate userland—are seeing renewed interest. He said the original difficulty of writing low-level components like storage drivers has diminished because AI models can now generate that code on demand. The conversation, part of Cormack's newsletter series, also touched on related technologies including Mirage, Orleans, Haskell, Nix and Cursed.
GoKawiil's interpretation of the reporting above, not reported fact.
If AI coding assistants can genuinely absorb the specialized, low-level work once required to build unikernels, it could lower the barrier to a security-focused architecture that strips out unnecessary OS components and attack surface. This suggests a broader pattern where AI tools may revive previously abandoned or niche engineering approaches by removing the manual labor that made them impractical, though the durability of this shift remains to be proven in production use.
- Unikernels embed application logic directly into a minimal OS with no separate userland or forkable processes.
- Justin Cormack says AI models now generate the low-level code (e.g., drivers, storage stacks) that once made unikernels impractical to build.
- The discussion was part of a newsletter series exploring renewed interest in unikernels alongside tools like Mirage, Orleans, Haskell and Nix.
Source: ghuntley.com — Geoffrey Huntley, 2026-10-10
Published there as: “Unikernels were hard. key word: were”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.