Currently at Appy.AI working on a multi-agent harness for Violet — an AI Chief of Staff that routes specialist agents with scoped context for real workflows. Prompts, context, orchestration, steering.
Daily use and experimentation of a variety of other agentic tools: Devin, Claude Code, Cursor.
Philosopher of science by training. Ph.D., GSN / LMU Munich; Masters from the Munich Center for Mathematical Philosophy. Dissertation: Knowledge Transfer in Cognitive Systems Theory — knowledge as control via structure mapping; using cybernetics to explain neural networks as error-controlled regulators. Cray deep-learning internship.
Also: Structural Idealism, scipfs, SciPhi Initiative, Nederland Philosophy Club.
co-vibed with Grok Bot for fun · math & concepts not fully validated
Same "model" both sides. Left: no world access (vat). Right: tools, memory, verifier, stop — ecological / embodied / embedded control. Click a pane to drop a goal. Toggle harness parts.
Toy model of multi-agent orchestration (not structure mapping): one director assigns each task to one specialist. Specialists do not share write context. Matching type → task lands; mismatch → negative transfer (wrong tools / context). Drag, or press "route by type."