About the lab
Building culturally relevant AI for humanity.
BlackMind is an independent AI research lab at the Georgia Cyber Warfare Range. We build AI for humanity as a whole — grounded in culture, and held to a standard most of the field reserves for hardware: results you can verify, and claims a proof can discharge.
The work
The lab works along two fronts. Its products bring computational tools to the study of culture and discourse — reading it closely, at a scale a person cannot. Its research pushes intelligence down to its true substrate: training models on raw compiled binary rather than natural language, building differentiable computers whose semantics are recovered from execution and then proven equivalent to a reference by machine, and compiling trained models into verified logic gates that can be taped out in silicon. Across all of it, a claim is only a claim once a proof discharges it.
The founder
BlackMind was founded by Jovonni L. Pharr, an independent researcher at the Georgia Cyber Warfare Range. His work spans language models, formal verification, neural computers, and hardware, and is published on arXiv with widely cited work. The lab is small and instrument-driven: it would rather sharpen its own measurement than trust a number it cannot re-derive.
How we publish
BlackMind publishes for reproducibility. Every number in a result is backed by a committed verification script that re-derives it from a machine-checked receipt — no figure stands on trust alone. This site is the canonical home for the work, maintained alongside its arXiv record so that the lab’s results and proofs stay in one place.
Start with the research, or explore the lab’s public instrument, the Discourse Engine.