Publications

Open access to the lab's foundational research. We publish work that shapes the discourse around safe and efficient AI.

2025

The General Theory of Stupidity

A formal model of cognitive vulnerability. Investigating the "smart stupidity" paradox and defense mechanisms against irrationality.

Paper
2025

Power-Survival Stack

Architectural solutions for overcoming the energy crisis in AI scaling.

Preprint
2024

Neural Bytecode: The Efficiency Language

Introducing a dense intermediate representation to decouple logic from linguistics in LLMs.

Preprint