Selfpressionnisme : Et si l'IA nous rendait plus humains ?
The book in brief. Marie Dollé published Selfpressionnisme with EMS in January 2026. My notes do not summarise the book itself. They record what Dollé argued at a conference on AI and humanism at the École Militaire in April 2026, where she spoke from phenomenology and aesthetics about what in human experience escapes optimisation. The ideas below come from that talk.
Key ideas
- Horizon. After Merleau-Ponty, perception is always partial, and the not-yet-seen is what gives depth and orientation to what we do see.
- Sensible knowledge. After Baumgarten, the senses produce their own kind of truth, distinct from logic and not inferior to it.
- Infra-ordinary. With Perec, she asked us to look at what happens when nothing happens.
- Disequilibrium. After Piaget, intelligence develops through struggle, which AI absorbs before it reaches the user.
- Pharmakon. Every technology is both remedy and poison, as her SoftVoice example shows.
Why it matters to my work. Dollé supplied much of the philosophical armature of AI, Magic Skin, and Je-ne-sais-quoi, including the references to Perec and to Derrida. Her answer to my question in the Q&A, that style is the trace of a body in language and something no model can replicate, led me to write that style may be the last frontier.
