Sets intent and retains accountability
Defines the useful result, determines acceptable risk, grants authority, and approves consequential decisions.
The principal
Human direction. Bounded machine authority.
MoonVault is shaped by Karl Ohlemann’s work as a founder-operator and by BlackBoxx.ai’s technical exploration of agents, automation, explicit authority boundaries, human-controlled execution, and evidence-backed workflows.
Karl Ohlemann
The person remains the principal.
Karl sets the thesis, defines the intended outcome, grants authority, and retains responsibility for consequential decisions. MoonVault applies that same operating principle to digital systems: automation can expand human capacity, but it should not obscure who authorized the work, what limits apply, or who remains accountable.
His perspective combines operating experience, governance discipline, digital-asset exploration, and a practical interest in how people preserve control across organizations, technologies, and life transitions.
Relationship map
Three distinct roles, one coherent inquiry.
Develops the trust architecture
Explores identity, delegation, custody, continuity, recovery, and durable evidence as an integrated control system.
Tests the technical patterns
Explores practical agents and automations with explicit authority boundaries, human-controlled execution, and evidence that survives the demo.
BlackBoxx.ai
Where the agent-governance thesis becomes operational.
BlackBoxx.ai is an independent AI Systems Lab. Its work with specialized agents, orchestration, application integrations, and controlled execution creates a practical environment for examining the questions MoonVault is organized around.
How should an agent identify itself? How is its authority derived from a person or organization? What should happen when a task exceeds scope? What evidence should be retained? How does the principal revoke authority and recover control?
MoonVault and BlackBoxx.ai are complementary, not interchangeable: MoonVault develops the long-horizon trust and continuity thesis; BlackBoxx.ai explores practical system patterns that can inform it.
The point of trustworthy automation is not to remove the principal. It is to let the principal delegate without disappearing.