Un nouveau modèle IA de Claude est très proche ! Et il serait doté d'un "raisonnement Extrême". Traduction : Anthropic prépare de nouvelles versions de ses deux plus grands modèles — Claude Sonnet et Claude Opus — qui devraient sortir dans les semaines à venir, selon deux
LLMS
-
Reasoning tokens alone cannot solve fundamental AI limitations
By
–
Adding more reasoning tokens is not going to fix this
-
Anthropic prepping Sonnet 3.8 and hybrid Opus models
By
–
It is evident that Antropic is gearing up for new model releases. According to The Information, we will get both Sonnet and Opus versions. – “claude-neptune” – Sonnet 3.8?
– Will Opus be a hybrid system like GPT-5? h/t @steph_palazzolo -
Twitter Spam Uses Dadaist Word Salad to Evade LLM Classifiers
By
–
New trends in Twitter spam, from the people who brought you 160,000 reasons to turn off notifications: Dadaist word salad meant to apparently suggest trading gains to a low-sophistication mark while avoiding an LLM classifier given the task “Is this user talking about trading?”
-
o3 Model Experiencing High Hallucination Rates
By
–
o3 is hallucinating a lot, not sure if recent or always been true thought it was the smartest model?
-
AI Intelligence Surpassing Humans by Late 2025 Predicted
By
–
Franchement, ce qu’Elon Musk dit ici me paraît de plus en plus réaliste. Quand on voit la vitesse à laquelle l’IA progresse, surtout depuis GPT-4, on peut sérieusement envisager qu’elle dépasse l’intelligence d’un humain d’ici fin 2025… puis toute l’humanité peu de temps après.
-
ChatGPT Reveals Its System Prompt Blueprint
By
–
ChatGPT just revealed its internal system prompt. It’s the blueprint that powers everything it does and now you can study it, tweak it, and build better prompts. I broke it all down in this System Prompt Guide. Like + comment “System” and I’ll DM it to you.
-
Salesforce Launches xGen-Small Models with 4B and 9B Parameters
By
–
Salesforce released xGen-Small, a new family of small models It includes 4B and 9B param models, with the larger variant proving competitive on long-context understanding up to 128K tokens Also delivers strong performance on math and coding benchmarks
-
Alibaba Qwen Launches Deep Research Feature for All Users
By
–
Alibaba's Qwen team announced that Deep Research on Qwen Chat (similar to that of ChatGPT) is now available for everyone
— Rowan Cheung (@rowancheung) 14 mai 2025
It puts together a detailed report on complex subjects, complete with citations.pic.twitter.com/ap1H8Yr5zWAlibaba's Qwen team announced that Deep Research on Qwen Chat (similar to that of ChatGPT) is now available for everyone It puts together a detailed report on complex subjects, complete with citations.
-
Notion Launches AI for Work Suite with Advanced Features
By
–
Notion released AI for Work, a suite of work-centered AI features, including:
— Rowan Cheung (@rowancheung) 14 mai 2025
— AI Meeting Notes
— Enterprise Search to find answers across tools
— Research Mode to draft docs
— Access to models, including GPT-4.1 & Claude 3.7pic.twitter.com/J1MZFBsuy0Notion released AI for Work, a suite of work-centered AI features, including: — AI Meeting Notes
— Enterprise Search to find answers across tools
— Research Mode to draft docs
— Access to models, including GPT-4.1 & Claude 3.7