Linda Petrini

I became an AI researcher because I wanted to understand what intelligence actually is — and I stayed because the question kept getting stranger. Interpretability work does that to you: you spend years studying what neural networks are actually doing inside, and you come away less certain about mind than when you started. That uncertainty felt like the right place to build from. Everything I do now — the research, the essays, the coaching — is an attempt to think carefully from inside that gap.

How I got here

I spent four years doing hands-on AI research — interpretability, epigenomics, the slow and absorbing work of trying to understand what’s happening inside black-box systems. It was meaningful work, and it eventually wore me down. The burnout that followed wasn’t a failure — it was information. I went deep into coaching, embodiment, and somatic practice, and what came out the other side was more integrated: someone who could hold the technical and the human in the same hands, without either one cancelling the other out. I now live semi-nomadically in rural Italy, teach acroyoga, and am training in Rolfing.

What I care about now

There is a conversation happening about AI — very loud, very fast, very confident — and almost none of it touches what I find most interesting. Not whether AI will take jobs, but whether it changes what it means to have a thought. Not just whether it’s dangerous, but what we owe ourselves in its presence: what sovereignty means when your environment is shaped by systems you can’t see and didn’t choose. The technical discourse is fast and often disconnected from what people actually feel — the anxiety, the wonder, the sense of something shifting underfoot. I want to close that gap. Not by simplifying the technology, but by taking both sides seriously.

Background

I freelance as a technical writer and researcher, with published work through Palisade Research, Foresight Institute, Anthropic, and the Bezos Earth Fund, and peer-reviewed research at ICLR. My coaching practice draws on seven years at the frontier of machine learning, combined with serious training in somatic and embodiment-based approaches. I’m also developing a Digital Sovereignty course for people who want to understand AI well enough to make real choices about it. I write about AI and life more broadly on my Substack.

If any of this resonates — as a potential collaborator, a reader, or someone who just wants to think it through — I’d be glad to talk.