Professor of Computer Vision/Machine Learning at Imagine/LIGM, École nationale des Ponts et Chaussées @ecoledesponts.bsky.social Music & overall happiness 🌳🪻 Born well below 350ppm 😬 mostly silly personal views
📍Paris 🔗 https://davidpicard.github.io/
David Picard
Loading...
found a little buddy and brought her to a nice flower shrub, but why are her wings so mangled? :(
#invertebrate
La barre du +1,5º devrait etee franchie vers 2030.
Aujourd’hui, le Sénat examine une proposition de loi pour lever, dans les territoires d'outre-mer, l'interdiction de recherche, d'exploration et d'exploitation des hydrocarbures. Mais « de manière écologique ».
😤🤯🤬😫
www.lemonde.fr/planete/arti...
Very excited to share our new paper, accepted at ICDAR 2026: "Leveraging Morphology for Historical Script Metrological Analysis"! 📜We present a method to automatically extract high-quality character prototypes and compute stable paleographic measurements:
malamatenia.github.io/morphology4m...
In 2019 it took 6 months for GPT-2 to go from too-dangerous to release. Claude Mythos/Fable ran that cycle in 2 months.
The AI PR singularity is amongst us. In 2029, there will be zero latency between a company saying a model is too dangerous and it being released.
Anthropic doubling down on "safety for thee but not for me" underscores the importance of open source/weight models.
It's not OK for one company to dictate what you are or are not "allowed" to work on. Or even worse, to *poison your software if it thinks you're doing something dangerous*.
