AI and me
I’m a data scientist in aerospace. Since December 2025, I’ve been using agentic AI every day and building my own ecosystem of tools, AXM, to entrust it with increasingly complete tasks. This practice brings me up against very concrete questions: how to delegate, verify a result, make room for challenges, or let a system work independently while retaining the means to understand what it is doing.
As this experience grows, I recognise problems that other fields have been grappling with for a long time. How does an organisation distribute responsibility? How does the justice system allow a decision to be challenged? How do living systems preserve information while changing? Building agentic systems gives me a new perspective on these questions; exploring them, in turn, helps me think more clearly about what I’m building.
This blog grew out of those connections. Here I share my reflections through analogies between AI and what exists elsewhere: in science, institutions, organisations and everyday situations. An analogy interests me when it reveals a relationship I hadn’t noticed, suggests a direction to explore or forces me to reframe a problem. Its limits matter too: understanding where the comparison breaks down is part of the journey.
I approach these subjects with the habits of a data scientist and a fascination with systems. I enjoy looking for a common structure behind seemingly distant phenomena, then testing that intuition against what I observe. My experience with AI is the starting point for these essays. Writing them is a way to explore where that experience leads — and to open my ideas to discussion.
Latest posts
- No telescope will save Trisolaris (or your agent) What the three-body problem teaches about the prediction limits of agentic AI, and why the only strategy that depends neither on model generation nor budget is to constrain the dynamics.
- The RNA moment of agentic AI What the origin of life teaches us about self-improvement loops — and why it's the constraint, not the supervisor, that closes the loop.
- What you won't code, you pay for in tokens Agentic AI is replaying the cloud's growing pains: mistaking “the machine can do it” for “the machine should do it.” A measurement, and what it taught me.
- What cassation taught me about evaluating AI Law separates what is right from what is justified. Artificial intelligence is rediscovering that line, in disarray.