Programs start as questions. Linked projects and papers are where the work currently stands.
GHOST and the Symbolic Human Life Simulator study ecological selection, lineage behavior, symbolic agents, and competence in constrained worlds.
Can competence be earned by living in a digital ecology rather than imposed by an external optimizer?
Mimir, RouterBot, and DSL Router cover recall, routing, fallback, context budgets, and traces for agentic work.
What substrate do useful AI agents need before autonomy can be trusted?
ANIMUS, Aether, Anima, and Atlas Vehicle Assistant cover local-first workspaces, persistent companions, voice control, and embodied assistance.
How should assistants persist, remember, act, and remain legible to the humans around them?
Self-hosted AI workspace with projects, providers, jobs, and updates.
View projectLocal assistant runtime with memory, voice, plugins, scheduler, and dashboards.
View projectDigital companion with drives, mood, and embodiment.
View projectVehicle voice UX, biometric flows, Android integration, and embedded security.
View projectIncludes Aethon, Team Hub, and The System for coordination, state, progression, and human-in-the-loop control.
How do agentic systems coordinate real work while preserving visibility and human control?
Daily operational assistant for briefs, approvals, customers, and task visibility.
View projectWorkspace-native coordination surface and dashboarded operations.
View projectVoice-first progression system with stats, XP, skills, and retention design.
View project