Vision

oml (oh my lisp) — a personal, persistent, self-evolving Lisp you drive in natural language, and, when you want precision, in Lisp directly. An agent lives in the image, acts through a bounded grant, and the image grows by keeping the workflows that worked.

Why it matters

• Natural language first. You talk to it. Without that, it is a REPL for hackers; with it, it is something a person can actually use. This is the difference between a toolkit and a companion.
• A body that remembers. A bare language model forgets. oml is a persistent image: its objects, state, and learned skills survive restarts. The work you did yesterday is still here today.
• It grows. Every request is fulfilled by composing and running Lisp in the image. When a workflow works, it is kept as a named skill. The longer you live with an image, the richer its vocabulary — and the better it understands what you mean.
• Safe to inhabit. The agent acts through a deny-by-default grant: a bounded vocabulary, not ambient power. That makes it safe to hand a capable but untrusted mind the keys to a live environment.
• Lisp underneath. Precision when you want it, and a substrate where behavior is data — so the image can extend and rewrite itself.

The loop

natural language
   → the agent composes and runs Lisp in the image
      → if it works, it is saved as a skill
         → next time that intent is a known capability,
            and your language is understood more richly

This compounding loop is the point. A plain assistant cannot do it — it has no persistent body in which a vocabulary can accumulate. oml does.

Personal and extensible

oml is yours. It starts small and gains capabilities by import — you add what you need, when you need it. Its identity is the personal, growing Lisp; everything else is something you bring in.