Error 404, the story of DeSci
Grant success rates at the NIH are at historic lows. Decentralised science says the institution itself is the problem. This episode follows the money, the tokens, and the research that actually came out of it.
One hundred and thirty four conversations and deep dives on artificial intelligence and health technology, in English and in French, since 2017.
Soixante ans d'histoire, des premières équations de Hansch et Fujita en 1964 jusqu'au prix Nobel de chimie 2024, en passant par le docking moléculaire et AlphaFold.
Grant success rates at the NIH are at historic lows. Decentralised science says the institution itself is the problem. This episode follows the money, the tokens, and the research that actually came out of it.
From rule based computer aided detection in the nineties to the models that lose accuracy the moment they leave the hospital they were trained on. Ten years after radiology was declared finished.
Separating clinical utility from marketing, with a cardiac electrophysiologist who uses these tools in the lab rather than in a keynote.
Comment un laboratoire acculture des centaines de collaborateurs à l'IA générative, sans slide de promesse et avec des cas d'usage qui tiennent après le pilote.
Recorded on the floor at HIMSS Europe in Copenhagen, minutes after the closing session, on what the EHDS actually asks of health organisations.
Why crystals grow differently in orbit, and whether that is a serious pharmaceutical argument or an expensive one.
Someone doing real work on health technology, yourself included. Tell us what you are building and why it matters now. We read everything and we answer.