On vit un moment assez fou. Voir le développeur de Minecraft partir en croisade contre l’IA qui code, jusqu’à la comparer aux NFT, est assez révélateur. (De mon point de vue, il est resté bloqué sur l’état de Copilot d’il y a deux ans et ne s’est pas vraiment intéressé à l’état
LLMS
-
Fine-grained VLM Reward Model with Stepwise Confidence Metrics
By
–
Finally! A fine-grained VLM-based general reward model with a stepwise confidence metric. Also novel segmentation and fusion of incremental, forward, and backward prediction. https://t.co/m5k2ghkUD7
— Ken Goldberg (@Ken_Goldberg) 30 décembre 2025Finally! A fine-grained VLM-based general reward model with a stepwise confidence metric. Also novel segmentation and fusion of incremental, forward, and backward prediction.
-
OpenAI’s 2025 Ship Cycle: GPT-5 Releases and AI Agents
By
–
OpenAI’s 2025 ship cycle was relentless: GPT-5 (Aug 7), GPT-5.1 + Codex-Max (Nov), GPT-5.2 + GPT-5.2-Codex (Dec), plus new ChatGPT Images (GPT-Image-1.5) with 4x faster, more consistent edits. Agents are the real story. Track the tools and more on Futurepedia.
#OpenAI -
Eric Schmidt Warns About AI’s Transformative Power Within 5 Years
By
–
L’ancien PDG de Google, Eric Schmidt, lâche un avertissement glaçant sur l’avenir de l’IA.
— VISION IA (@vision_ia) 30 décembre 2025
« D’ici 5 ans, l’IA pourrait gérer un contexte infini, un raisonnement en chaîne avec des solutions en 1 000 étapes, et des millions d’agents travaillant ensemble.À terme, elles… pic.twitter.com/Jg7O1UtV7iThe former Google CEO, Eric Schmidt, issues a chilling warning about the future of AI. "Within 5 years, AI could handle an infinite context, chain-of-thought reasoning with solutions in 1,000 steps, and millions of agents working together. In the long run, they will develop
-
Scaling AI Workloads with Ray: From Data to Agents
By
–
🎙️ In this episode of Yet Another AI Podcast, we sat down with @lindavivah from @anyscalecompute on scaling AI workloads with Ray from data processing to training, inference, and agents.
— AI21 Labs (@AI21Labs) 30 décembre 2025
We look at Ray’s roots in RL, why tools like vLLM build on it, and when Ray vs. SaaS makes… pic.twitter.com/9W8QPxztiOIn this episode of Yet Another AI Podcast, we sat down with @lindavivah from @anyscalecompute on scaling AI workloads with Ray from data processing to training, inference, and agents. We look at Ray’s roots in RL, why tools like vLLM build on it, and when Ray vs. SaaS makes
-
Minimax-M2.1 ranks 6th in web dev arena
By
–
Minimax-M2.1 secures 6th spot on web dev arena, becoming a leading open source model in this category.
-
Claude Opus vs GPT-5.2: Capabilities and Behavioral Differences
By
–
Having used both Claude 4.5 Opus and GPT-5.2-xhigh side by side I can say that Opus has a beautiful mind, it's like an artisan who can come up with wonderful ideas & quickly. But if I don't want the model to suddenly flip the table and kick me in the nuts, I switch to GPT-5.2,
-
User asks Grok to write in similar style
By
–
Hey @grok could you write some stuff in similar style?
-
Meta’s Manus AI scores 2.5% on Remote Labor Index
By
–
Manus AI, acquired by Meta, is a current SOTA with a 2.5% achievement on the "Remote Labor Index" benchmark (Introduced by Meta). The Remote Labor Index measures AI automation of remote work. Soon to know how much 2.5% is worth
-
Grok analyzes YouWare interview: China’s AI tools reshape innovation
By
–
Grok says, about this interview: "It's not just a product demo. It's a window into the evolving AI world, where China's emphasis on accessible, optimized tools is reshaping who can innovate. Viewers emerge with practical knowledge of YouWare's capabilities, inspiration from its