I ve always looked at things structurally and analytically. That s just who I am.
I never had a technical education, but I could still see where tools in the workplace were falling short. Sometimes I could simplify things, automate parts of the work in Excel, and make processes clearer.
Then AI came along, and I started experimenting with it.
I expect some of the claims around APE to raise questions, and that s fair. So rather than overloading the launch page, I wanted to open a space for discussion here.
APE was built through XOX. XOX did not come from theory. It came from repeatedly seeing where AI failed in real work, and realizing I had no coding knowledge to step in and correct it myself. Structure was the only weapon I had. So I analyzed those failures, found the patterns, and turned them into a method.
In XOX, architecture and decisions are locked before execution begins. If you want to ask about the method, the product, or what no code means in practice, ask away.
APE is a deterministic AI-first programming language for unambiguous human-AI collaboration. Built in 6 days using XOX, a method where architecture is locked before execution. Includes a compiler, runtime, decision engine, stdlib, 850 passing tests, and integrations for Claude, OpenAI, and LangChain. 9,000+ downloads on PyPI.