Might be a dumb question/remark but while this story is super bad, I don’t believe our regulations are based on this one study, right? AFAIK, EFSA and IARC (which only found limited evidence of carcinogenic effect on humans) used more data (with better quality) than just this article: https://link.springer.com/article/10.1007/s00204-017-1962-5
While this article and its conclusions are bad and did a lot of damage in terms of trust, do we know that this retraction means that the conclusions is that it is toxic? I’d be surprised that this is the one study we base all our regulatory decisions on, because I tend to trust EFSA’s conclusion, their work is usually good and up to date.


















Pour l’instant c’est assez propre ce qu’ils ont fait je trouve: la traduction dans la navigateur directement sans passé par un serveur, et les chats LLM dans la barre sur le côté. Tout optionnel et assez compréhensible.
Par contre ils ont aussi fermé plein de trucs à cause du manque de profil (cf pocket) donc si ils se diversifient j’espère qu’ils apprendront de ces échecs :/