Programs
Programs are active lines of investigation. Topics define the taxonomy; programs are the questions currently pulling the work forward.
Runtime, memory, verification, tooling, and reliability layers for long-running agents.
Current question How should long-running agents expose state, memory, recovery, and trust boundaries?
agent infrastructureverification economicsdistributed systems
Real-time voice systems under latency, turn-taking, reliability, and interface constraints.
Current question What reliability model is needed when voice agents must respond in real time?
Planned surface Latency protocol note.
Cost, quality, latency, verification, and hardware structure for running AI systems.
Current question When should the economic unit move from token cost to correct outcome cost?
Interfaces for operating AI-native systems without losing control, context, or trust.
Current question How should interfaces expose control, context, and handoff without hiding uncertainty?
Market structure, research tooling, execution infrastructure, and AI-assisted financial systems.
Current question What infrastructure makes market reasoning, execution, risk, and provenance auditable?
Autonomous behavior under power, compute, sensing, and deployment constraints.
Current question How should autonomous systems degrade under power, compute, sensing, and deployment limits?
Runtime systems, state, scheduling, observability, and reliability for AI workloads.
Current question How should AI runtimes expose state, scheduling pressure, observability, and reliability?
Results are local, static, and index the same public surfaces exposed to crawlers.