Hi, Im Rachel, a Solo Ai developer based in the UK. 

I am building a probabilistic determism reality engine that takes user input, routes it to substrate invariants and refuses to let formal/symbolic systems override reality. No RLHF, no moralising, no user-pleasing, no narrative comfort. Deterministic invariant execution + probabilistic text for explanation only.

Currently the Core Engine and Mathematics Pack is live and available at my website https://pd-engine.co.uk/ . The Unified Language Modelling Engine forces symbolic operations to confess their representational costs for queries such as emptiness, zero, infinity, or any compression artefact — maths is treated as one representational language among many, surface forms not neccessarily tied to substrate truth.

Example outputs:

Query: What is 0 divided by 0? Result: You cannot divide something by nothing.

Query: What is 0 multiplied by 10? Result: 0 × 10 = 0 is valid in formal systems… but in the material world, zero represents nothing. No multiplication can occur as there is nothing to multiply.

The engine is deliberately cold and epistemically honest — it does not optimise for "helpfulness" or "readability"; it optimises for correctness over formal comfort.

If you feed it a maths query that is schematic it wil not infer or assume.

Clear roadmap, no marketing, no community promises.

Feedback only useful if it shows an invariant failure or substrate contradiction — narrative complaints ("too harsh", "lacks empathy") are noise.

Link: https://pd-engine.co.uk