Applicationless Computing

For fifty years, the application has been the unit of software: a discrete thing you install, open, and operate. Agentic systems are already blurring that — but mostly by making applications call each other faster.

This programme asks a different question: what changes when the primitive is not invocation — one entity calling another — but the resolution of intent? When you stop asking "which application do I open?" and the substrate itself decomposes what you want, propagates it, and composes a result?

Most of the interesting problems turn out to be trust problems. An act with real consequences — money moved, content published, a commitment made on your behalf — has to carry evidence of where it came from, what authorised it, and how it can be contested. Today that evidence lives inside application boundaries. Without applications, it has to live somewhere else.

A book-length treatment is in preparation, with academic publications to follow. This page will grow as the work becomes public.