06 / About
About SHINGETSU
A Canadian computing research organisation investigating foundational problems in artificial intelligence and computational systems.
Organisation
SHINGETSU (新月) is a Canadian computing research organisation established to investigate foundational problems in artificial intelligence and computational systems. The name derives from the Japanese term for the new moon — the period of renewal and beginning that precedes illumination.
We are an independent research body. We are not affiliated with any university, commercial laboratory, or government agency. This independence is deliberate: it allows us to pursue questions on longer time horizons than commercial pressures typically permit, and to publish findings without regard to competitive interest.
Our work is measured in the quality of the questions we pose and the rigour of the methods we apply. We do not benchmark success by funding secured or headlines generated. We benchmark it by whether our research is reproducible, whether our findings are durable, and whether the problems we address are genuinely important.
Mission
"To investigate foundational problems in artificial intelligence and computational systems, and to advance the state of knowledge through rigorous, reproducible, and independently conducted research."
This mission is pursued through two research programs. The Efficient Intelligence Program investigates the concentration of capability within constrained computational budgets. The Computational Continuity Program investigates the long-term preservation and verifiability of computational knowledge.
Research Principles
Reproducibility
We publish protocols and baselines sufficient to allow independent replication. Results that cannot be reproduced are not results.
Negative Results
A hypothesis that fails is scientifically valuable. We publish negative results on the same terms as confirmatory findings.
Rigour Before Speed
We do not publish prematurely. We accept that thorough work takes time, and that the field is not served by rushed findings.
Independence
We have no commercial interests and no competitive pressures that would incentivise selective reporting or accelerated publication.
Formal Measurement
We are sceptical of informal benchmarks. We develop the measurement frameworks our research requires rather than defaulting to those that already exist.
Durable Questions
We work on problems whose importance is not contingent on the current state of technology. We ask questions that will still matter in twenty years.
Research Governance
Every formally approved SHINGETSU research project receives a permanent identifier and is entered into the central research registry. This identifier is issued once and never changed, deleted, or reused. The registry is a durable record of everything we have formally undertaken — including projects that were cancelled or did not achieve their objectives.
This commitment to permanence reflects our view that the scientific record should be complete. Cancelled projects represent real inquiry and real findings, even when those findings are negative. They belong in the record.