AI Pioneer
Manuel Fuß.
I want to understand how AI arrives at an answer. What does it pick up from my question? What does it miss? And why can changing a single sentence lead to a different result?
I like exploring those questions through experiments I can run myself. I give an agent a task, look at the choices it makes and try again with different context. The results that surprise me are the ones I want to look at more closely.
What I’m working on
I’m a Senior AI Native Engineer at re:cinq.
I build the workflows teams use to ship software with coding agents. I work alongside client engineers in their own codebase. Together we describe what a change needs to do, give the agents the context they need and review what’s ready to ship. The team should be able to keep working this way after we leave.
What interests me about AI
In my talks I often come back to a question I encounter in my own experiments: how do we notice when an AI system does something we didn’t intend? I’m interested in who can spot it and what it would actually take for that person to step in.
A game I made
AI Doomsday
In AI Doomsday, you decide how much control to hand over to AI. I wanted to turn that question into something you could play.