Public programs
Fields don't stall in general. They stall on a specific constraint. These are the programs we run to loosen those — and the evidence we look at next.
How we choose the work →Diagnose
Identify the bottlenecks
Read field velocity. Name the constraint that is actually in the way.
Intervene
Match the work
Orient, coordinate, resource, build, prove, or enable — whatever loosens that bottleneck.
Learn
Examine the evidence
Watch whether the field moved. Update the hypothesis, continue, or stop.
Draft-source examples. Not approved commitments.
8 published examples. Draft-source programs, labeled as proposed — not a verified active inventory.
Help national institutions turn open models into useful, locally maintained AI capability.
Turn public-goods research conversations into sustained collaborations.
Develop shared research and evaluated AI tools for environmental institutions.
Turn country requirements into usable compute access and technical capability.
Help a senior cross-institutional cohort advance shared missions.
Make a large-scale connectome program technically and financially credible.
Build shared simulation infrastructure after a discovery target is proven.
Supply persistent operational ownership and an AI-native talent pipeline.
Public types here follow the FA2 draft vocabulary. See the methodology for how they relate to the Console toolkit.