Workshop notes
What gets thought
What gets thought
between two builds.
Short essays, often polemical, always owned. On AI that isn't intelligent, sovereignty we think we have, robots that fit in your hand, and search engines that now read like machines.
Essays are published first in French. English translations follow. Meanwhile, you can read the originals at galaadmf.fr/blog.
Semantics · 9 min · FR
Why I stopped saying "artificial intelligence"
An LLM doesn't think. It predicts tokens.
Read (FR) → Mythology · 8 min · FRThe shield and the grail: why Galaad
The quest more than the arrival.
Read (FR) → Sovereignty · 10 min · FRYour AI should not live at someone else's house
Personal PaaS, LiteLLM, Cowork Souverain.
Read (FR) → Search · 9 min · FRGEO/AEO: how AIs will read your site in 2026
End of PageRank, rise of AI engines.
Read (FR) → Robotics · 10 min · FRThe companion robot that fits in your hand
Silver Economy, Article 51, small over spectacular.
Read (FR) → Workplace · 9 min · FRAlan, or the AI that creates jobs instead of cutting them
600 employees, +50 % growth, zero layoffs. AI on a leash.
Read (FR) → Trends · 10 min · FRAnthropic's study on how people actually use AI
1M conversations analysed. 52 % augmentation, 45 % automation.
Read (FR) → Writing · 8 min · FRHow to spot a text written by an AI
Seven tells. And how to write differently if you're human.
Read (FR) →