FAQ

A small research lab built around one question: can the architectures of learning systems be derived from entropy, energy, and wave dynamics, instead of being found by trial and scaled?

The work is theory first. Derivations are tested in small numerical experiments, and anything that survives gets built and run. The research program states the question precisely, lists the established results it builds on, and fixes the sequence of tests it has to pass.

Enkidu, in the Epic of Gilgamesh, was formed from clay. The motto on the front page, ex luto ad astra, reads "from mud to the stars", and the parallel is deliberate: this program treats intelligence as something that arises from raw physical substrate, from matter, energy, and noise, and that can be understood at that level.

The glyphs 𒂗𒆠𒁺 are the name's Sumerian form. Cuneiform is the oldest writing that can still be read: thought pressed into wet clay and preserved. Durable structure out of mud is the oldest precedent for what this program claims about intelligence.

Today: a small lab. The intended end state is an institution that carries both the science and the engineering, where what gets derived gets built, and where systems, software, and knowledge compound in one place over a long time.

It is not structured as a product startup, and nothing on this site is positioned for a funding cycle. The current state of the work, including what has not been done yet, is described in the research program.

Scaling what works is already being done well, by heavily funded groups. The reason to work on foundations instead is the historical pattern: fields acquire their engineering power when they acquire their theory. Thermodynamics turned engine design from craft into calculation. Electromagnetic theory produced radio. Information theory took communication from rules of thumb to provable limits.

Machine learning today has powerful systems and empirical recipes. Statistical learning theory analyzes parts of what exists after the fact; it did not produce the architectures and does not say what should come next. If architectures can be derived, the systems that follow differ in kind. If they provably cannot, that proof would itself be a foundational result. Either outcome is worth having.

No. The waves in this program are classical. Superposition and interference here are properties of classical fields, the kind that water, sound, and light exhibit, and the surrounding mathematics is statistical mechanics and information theory. No quantum hardware is assumed, required, or claimed. If a result ever depends on quantum effects, it will say so explicitly.

A stated research program with an explicit falsification sequence. Dated lab notes, published as they are written. Formal groundwork and small-scale numerical experiments in progress.

There are no products, no benchmark claims, and no announced breakthroughs. Technical results will be published when they hold, together with the material needed to rerun them. Until then, the honest record of progress is the notes.

The phrase "artificial intelligence" now names a market and a funding climate more than a research question, and most of what it covers is not what this lab works on. The vocabulary here is narrower because the claims are: learning systems, inference procedures, computational substrates. Precise words are part of the method.

Write to research@enkaidu.com and include something you have finished: a derivation, a proof, a simulation, a system that runs. Background in mathematical physics, information theory, or systems engineering helps. Finished work counts for more than pedigree.