Blog News AI #news

Meta has announced that it will use public content from European users to train its AI. The goal is to improve Meta AI's accuracy in EU linguistic and cultural contexts, but the move raises ethical and privacy questions. In this article, we look at the benefits, risks, and future implications.

ByteDance presents Seed-Thinking-v1.5, a new language model capable of “reasoning” with a step-by-step method inspired by human thought. In this article, we delve into its features, benchmarks, and why it could represent a serious alternative to giants like OpenAI and Google.

Discover Llama-3.1-Nemotron-Ultra-253B-v1, the new AI model released by NVIDIA: 253 billion parameters, optimized efficiency, advanced reasoning capabilities and ready to be used in enterprises. Everything you need to know.

Meta has brought artificial intelligence directly into our WhatsApp chats. It's called Meta AI and it's an assistant that can answer questions, suggest ideas and help in real time. But how does it really work? And most importantly... can it be removed?

Google announces a program to fund short films that explore the relationship between humans and artificial intelligence. An initiative designed to merge storytelling, emotion and technology, with the support of Range Media Studios.

YouTube introduces Music Assistant AI, a tool that lets you generate custom music tracks for long-form videos directly from text commands. Royalty-free, custom, and ready-to-use music powered by artificial intelligence.

OpenAI is preparing to release GPT-4.1, the new evolution of the multimodal model already known for its power. But not only that: “mini” versions and an even more advanced reasoning model are also coming. Here are all the expected new features.

ChatGPT boosts its memory by combining what it has learned from your explicit requests with your chat history. But what does this really mean? What can it remember? And most importantly: can you control it? Here's everything explained clearly.

ChatGPT 5 is coming and it promises a total revolution: it will be free, smarter, multimodal and without the confusion between a thousand models. But is it really like this? Let's discover together all the new features announced, the details on the features and what can really change in our daily use of AI.

Meta launches LLaMA 4 with three models dedicated to different AI applications. It is the beginning of a new generation natively multimodal.

Meta launches LLaMA 4 but blocks European users: promises of openness and proprietary controls raise doubts about the project's actual open nature.