Research
Evidence, not intuition.
Research notes on Bayesian cognition, decision theory, and the formal structure of rational inference — each citing peer-reviewed sources, none peer-reviewed itself.
No. 001
Hallucination Is a Calibration Failure
AI systems do not hallucinate because they lack knowledge. They hallucinate because their training objectives optimize fluency over epistemic honesty — and because no external verification step exists between generation and output.
Framework established — empirical validation pending
7 peer-reviewed sources cited
No. 002
When Heuristics Are Bayesian in Disguise
A formal investigation into whether mental shortcuts and probabilistic inference are adversaries — or the same process operating under resource constraints.
Heuristics as approximate Bayesian inference — evidence supports
9 peer-reviewed sources cited
Suggest a topic
Open call
What should we investigate next?
Submit a research question. We review all submissions and prioritize by demand and analytical tractability.