par Rémy Bigot | Juil 17, 2026 | AI
A German court ruled that Google is legally responsible for defamatory statements its AI Overview generates, ending the assumption that AI-generated summaries carry the same legal immunity as a search results page. The Regional Court of Munich found that...
par Rémy Bigot | Juil 17, 2026 | AI
China will have a model as capable as Claude Mythos before the end of 2026. A Zhipu AI co-founder said so publicly, and the American think tank IAPS puts the outside date at February 2027. When it arrives, Beijing will not have Washington’s meltdown. The US...
par Rémy Bigot | Juil 16, 2026 | AI
AI agent evals are the structured tests that define what « correct » looks like for an automated task, and without them, an AI agent workforce degrades into expensive, self-repeating loops instead of finished work. Gartner projects that 40%+ of agentic AI projects...
par Rémy Bigot | Juil 16, 2026 | AI
Most companies stall their AI agents at the pilot stage because nobody owns the context, the risk, or the bill, not because the model isn’t good enough. Gartner found 40% of agentic AI projects will be canceled by 2027 due to unclear ROI and weak risk controls....
par Rémy Bigot | Juil 16, 2026 | AI
Mistral et OVHcloud ne jouent plus le même match. Mistral a choisi la vitesse : 1,7 milliard d’euros levés avec ASML, ses meilleurs modèles licenciés à Microsoft Azure, un objectif clair, taper OpenAI et Anthropic à l’échelle mondiale. Tant pis pour...
par Rémy Bigot | Juil 15, 2026 | AI, Asymmetriq
China is not losing the AI race because it lacks chips. It is winning a decisive part of it because it produces the people who build the models. Nearly 40% of researchers presenting at NeurIPS 2024, the field’s top conference, were educated in China, up from 29%...