Compute Governance
ExploringUsing compute as the regulatory handle on frontier AI — chip controls, FLOP thresholds, KYC for cloud.
Territory marked · views not yet written
Governance
The policy
Real-Time Monitoring
DevelopingStanding infrastructure for monitoring AI behaviour and incidents in deployment — the Loss of Control Observatory and adjacent ideas.
Territory marked · views not yet written
Liability and Insurance
ExploringUsing tort, strict liability, and insurance markets to internalise the cost of AI harms.
Territory marked · views not yet written
International Governance
ExploringTreaties, agreements, and cross-border institutions for frontier AI — what's tractable and what's necessary.
Territory marked · views not yet written
Information Security and Model Weights
ExploringWhether model weights can be defended at all, and what changes if they can't.
Territory marked · views not yet written
Power Concentration and Democratic Governance
ExploringThe risk that frontier-AI capability becomes concentrated in a few actors, and what democratic legitimacy looks like for transformative technology.
Territory marked · views not yet written