MilOps
Prototypes, neural mischief and obsessions no one asked for.

Simon Miller
milops.org is my personal lab space for independent projects in AI, systems design, and the occasional deep dive into philosophical nonsense with machine-readable consequences.
I build things here when I have time, energy, and questions that won't go away. It's part portfolio, part playground, and part archive of digital experiments that may or may not go anywhere. But that's the point. Some ideas just need to be built—whether or not they're useful, marketable, or even explainable.
My free-time focus tends to orbit around:
- Researching the moral structures and implicit values in large-scale AI models
- Human-machine alignment
- Tools for understanding, interacting with, or gently poking at intelligent systems
- Open-source tools that no one asked for but someone might eventually need
What no one asked for
Everything here is made independently, without funding, corporate goals, or a secret cabal of advisors. Just a stubborn desire to understand and a mildly unhealthy relationship with language models.
If something here sparks an idea, feel free to reach out. If not—that's okay too. Like the tagline says:
"Prototypes, neural mischief, and obsessions no one asked for."
Recommendation
I might have stumbled across something incredible. Possibly some micro steps towards AGI: Proto-agency in ChatGPT-4o.
Or you could take a look at some of the projects I've made in my spare time.