Independent research by Rowan Brad Quni-Gudzinas
What computation really costs, and what it delivers.
QNFO is the independent research imprint of Rowan Brad Quni-Gudzinas. The work asks what computation really costs and delivers: energy per correct answer (Joules-per-Solution), what an AI-assisted claim is worth (ignorance audits), and what an autonomous research system actually delivers. Corrections ship as new versions of the same paper.
Selected works
The full libraryEnergy-honest computing
Epistemics of AI-assisted science
Autonomous research operations
Quni-Gudzinas, R. B. Published by QNFO. Prepared with an AI-assisted research pipeline; the author is responsible for the content.
Latest papers
Browse the library- Ultrametric $p$-Adic Noise Structure and Heisenberg-Limited Quantum Metrology: Error Resilience by Geometry, Not Overhead
- Adelic Zitterbewung: p-Adic Regularization of Dirac Trembling Motion and the Product-Formula Cancellation at Macroscopic Scales
- Ontological Re-Entry: What Can We Say Exists?
- What Makes a Measure "Genuinely Natural"? Bits, Nats, and the Audit of Naturalness Claims in Measurement, Information, and Language
- Optimal Radix $q$ under Resource Constraints: A Reconciled Derivation Framework for Ultrametric Hierarchies
- Ensemble Drafting Under Broadcast Control: Throughput Algebra, Variance Reduction, and Vandermonde Conditioning
Explore
Ask the corpus
Questions answered from the papers, with citations and the knowledge graph around each answer.
Work with me
Energy-per-correct-answer assessments, reviews of AI agent operations, talks, collaboration and roles.
Ideas
Research conversations as they develop, with the questions that start them.
Archive
The library by field, feeds and the tools for finding your way through it.
iPatent
A free, private-by-default assistant for drafting US provisional patent disclosures.
License
The QNFO Unified License Agreement: open science with commercial protections.