Record reality
ModelReceipt identifies the deployment stack and the evidence behind it.
OPEN//FRONTIER PROJECTS
Make AI systems legible.
One project records the system that ran. The other tests whether policy rules can survive the incentives they create.
THE PROJECTS
Two tools for harder questions.
Each project turns a vague AI claim into a record or mechanism that people can inspect.
PROJECTOPEN PROJECT
DEPLOYMENT EVIDENCE
ModelReceipt
What actually ran?
Record the model, software, machine, evaluation, and evidence behind an AI deployment.
Explore ModelReceipt →PROJECTMETHOD RELEASE 1.1
POLICY MODEL + RUN RECEIPT
OpenNash
Which assumptions make the rule hold?
Inspect the matrices and transforms, report every equilibrium, test payoff uncertainty, and save a citable run.
Explore OpenNash →WHY THESE BELONG TOGETHER
Inspect the system. Stress-test the rules.
Model incentives
OpenNash publishes policy assumptions, payoff transforms, robustness checks, and citable run receipts.
Test the claim
Review whether the technical claim is supported and whether the policy is stable.