code:emotion
Designer, systems thinker, and Human-in-the-Middle practitioner in Berlin.
This is my personal hub and public workbench. Code:Emotion is the established design studio: thirty years of craft, interfaces, and proof. Human in the Middle is the AI practice growing from that base: teaching as the front door, training and consulting as the engine, and a future social wing for NGOs. Here I show the work while it is still becoming real.
Branches
Sibling practices with different jobs.
Design
The established studio and design anchor: human-controlled AI products, governance interfaces, identity, editorial systems, and thirty years of craft.
AI practice
Teaching, training, and AI-governance advice for organizations deciding what to automate, what to keep human, and how to remain accountable.
Bridge
Code:Emotion is HITM's design and implementation advantage.
Code:Emotion gives HITM something most AI-governance practices lack: thirty years of interface design and implementation experience. DONE, GLASSBOX, and Panic Defense already show what judgment-preserving interfaces can look like; working prototypes will follow.
The hub holds the complete story while projects are still moving: the design questions, the AI-governance questions, screenshots, decisions, limitations, and next milestones. When a project becomes genuinely demo-ready or finished, it graduates to the branch best suited to present it.
Projects
A public workbench for projects that are still moving. Each entry states what works, what does not, and what comes next.
A trilingual workflow automation for tour CO2 tracking, built in two days and preserved as a frozen proof snapshot.
A guided environment for mapping human ownership, approval gates, risks, and the boundary between human and machine work.
A playful teaching device for making agentic workflows, delegation, and human checkpoints visible.
Writing
Essays, notes, case studies, and practical observations on design, systems, and everyday life with AI.
DONE. is a concept mobile app built around one deliberately antiproductivity constraint: choose a single meaningful task, commit to it, finish it, and then stop.
A controlled study shows different commercial chatbots materially diverge in how they respond to users exhibiting delusional symptoms.
Sony's Ace demonstrates AI moving beyond virtual tasks into high-speed, embodied human domains.
Better systems. Better judgment. More human results.
