What AI do we actually want?

This site is not a manifesto. It is an open notebook. I ask more questions than I bring answers. The main one fits on a single line. What place are we leaving for the human in tech's current trajectory?

I am not against machines. I have been building them for thirty years, and I keep doing so. My wish is more modest. That we take the time to inhabit them.

For a long time I was a sincere enthusiast. An engineer running a company that lives off AI, I carried the good word to the conferences. Then something shifted. A doubt settled in, about the pace above all. About what our teams, our children, our trades have the time to digest when everything moves so fast.

What you will find here is more notebook than pamphlet. A few questions asked out loud, and the will to slow down, the time it takes for the human to stay present.

I What I hold dear

Technology is not neutral. Nor is it a fatality. Generative AI opens many doors. I know it. I build with it every day. What worries me is not the technology itself. It is the pace at which we adopt it. Our trades, our children, our colleagues do not have the time it would take to tame it. Keeping a hold on the tempo remains, no doubt, our last technical freedom.

The human is not reducible to its productivity. There is slow thought, the inner dialogue, the carnal bond to others, the living memory of trades. There is silence, fertile boredom, inhabited doubt. None of these things is measured in gains. All of them fade away when we delegate too much, too fast, to machines that do not know them. Protecting them is not conservatism. It is civilisational hygiene.

What I stand for fits in a single sentence. A tech at human scale. Neither against the human. Nor without tech. Rather a tech that puts itself at the service of what we are, capable of attention, of care, of memory, of judgement. And that accepts the pace at which we can accompany it, without dissolving in it.

The stake is not to slow tech down. It is to keep the human at the centre of the path. notebook entry, 2026

II What I write

I have been writing about these subjects since 2018. Twenty-nine pieces in total. First on LinkedIn for lack of a better place, now here, in HTML composed by hand. Five also exist in English.

If you arrive knowing nothing about me, read these in order. They follow the path of my own thinking, from the first wonderings to the more recent questions:

• November 9, 2025Follow the white rabbit
• November 14, 2025AI: toward the extinction of inner dialogue?
• November 24, 2025Generative AI, the era of artificial emptiness
• January 30, 2026The memory of truth
• March 1, 2026The ultimate spoliation?
• March 10, 2026Chronicle of a more than uncertain past
• March 23, 2026The ideal child of techno-capitalism, age 14
• April 22, 2026For or against AI?

The other twenty-one, including the oldest ones on FrenchTech, hypergrowth, or why I walked away from my six-figure job in 2018, are filed there : all my writings. To follow new ones without depending on any platform : RSS, Atom.

III What I say on stage

I share these reflections out loud. Before executive committees, technical teams, engineering schools, conferences. Four talk tracks in circulation, adaptable to audience and format (60-min keynote, 90-min conference, 3-hour masterclass, two-day training).

• Exec · CSR · HRGoverning AI, training humans. A track on AI governance in companies and responsible employee training, to maximise useful uses without sacrificing sovereignty, judgement or cognitive health.
• CTO · techConfessions of an ex-AI addict: building an impactful tech in the era of generative AI. The honest story of a journey, and how to think the CTO's role as that of a guardian of human and intellectual capital.
• Tech leadersSteering AI without sacrificing your teams' genius. Understanding the cognitive effects of LLMs on engineers, distinguishing the uses that strengthen skills from those that erode them, and drafting a charter for human-machine co-reflection.
• General audienceArtificial intelligence : understanding its mechanisms to remain master of them. An initiation to enjoy AI without dissolving in it. Its promises, its traps, and the simple habits that allow anyone to keep a hand on it.

To book a talk, invite me to chair a round table, or open a keynote, the email is below. I reply within 48 hours.

IV What I say elsewhere, in less order

I co-host, with my friend Thomas Meimoun, the podcast IA pas que la Data. Long conversations with French-speaking AI players : researchers, founders, philosophers, former military. Thirty-three episodes so far, mixing in-depth interviews and monthly chronicles. Anne Beauchart handles editing, marketing, and supported the launch of the podcast.

I also appear regularly on BFM Business with François Sorel, on TV sets when the topic deserves it, and in the specialised press. Everything is listed there : appearances and interventions.

I read a lot. Eighty-three books on my shelves at the moment, sorted into six shelves : AI, philosophy, society, business, tech, documentaries. See my library. If you want to start somewhere : Sadin, Corteel and Carbonell for contemporary critique; Han and Arendt for the philosophy of the present; Harari, Zinn and Graeber for the long fresco.

V Get in touch

The email is pierre.vannier@flint.sh. I appreciate sincere letters, however brief. A reasoned disagreement, an honest question, an invitation to dialogue. I take the time to reply to everything that is not an automated commercial solicitation.

If you prefer another channel : LinkedIn.

My company, the one that pays my bills and that I run day to day, is called Flint. We do AI consulting there with our heads on our shoulders. I do not mix things up. This site is my personal space, not Flint's storefront.