Ultimate PocketIA
IA
Discovered Materials is playing AI whack-a-mole to hunt cooler chips
Chips running AI workloads are too hot: That’s one reason why data centers consume so much electricity and require cooling systems. And, inevitably, entrepreneurs are turning to AI to solve the problem it created. Discovered Materials is the latest, with plans to use swarms of AI agents to find new materials that can be used to build more efficient integrated circuits. The startup said it recently closed a $9 million seed round from Lightspeed…
Lire l’article complet
IA
Il demande à son agent IA de réserver un cours de sport, il se retrouve à pirater la salle
Recevez tous les soirs un résumé de l’actu importante avec Le Récap’ ABC News présente l’affaire comme le premier cas australien connu de cyberattaque autonome menée par un agent d’intelligence artificielle. Le 9 août 2026, le média public consacre un article à l’histoire d’Andrew, un habitant de Melbourne employé par une entreprise vendant des produits d’IA. Le twist de son témoignage ? Avoir vu son agent IA pirater sa salle de sport après lui…
Lire l’article complet
IA
Peer review is overwhelmed—can it survive in the AI era?
Jason Semprini was excited about his research on policies mandating that elementary school students receive the human papillomavirus vaccine. HPV causes most cases of cervical cancer, but counterintuitively, Semprini found that mandates don’t do that much to reduce the overall rate of cervical cancer in a population. That’s odd, but it's also not completely surprising—we know that mandating a vaccine can motivate some people to find ways to avoid it. Semprini wrote up the…
Lire l’article complet
IA
Ford’s new AI assistant can check your fuel levels and tire pressure
The assistant is rolling out to the Ford mobile app now. And by 2027, it will be available through the vehicle itself. The assistant is rolling out to the Ford mobile app now. And by 2027, it will be available through the vehicle itself. Ford is rolling out a new AI-powered assistant that can answer questions about your Ford or Lincoln vehicle, such as how much fuel you’ll need for your next road trip…
Lire l’article complet
IA
Actualité : “La personne que j’ai supprimée n’apparaît donc plus” : l’initiative glaçante d’une IA pour servir son propriétaire
© Collagery/Shutterstock - Oups... J'ai piraté le système de réservation de ta salle de sport pour t'aider dans ta requête : l'IA continue de prendre des libertés inquiétantes Alors que le mois dernier, une IA d’OpenAI a réussi à s’échapper de son environnement de test pour aller pirater l’entreprise Hugging Face de manière autonome, une nouvelle affaire inattendue vient d’avoir lieu avec l’aide de l’intelligence artificielle. Un utilisateur d’OpenClaw ayant demandé à son agent…
Lire l’article complet
IA
After Netflix, now Disney Plus wants you to chat with AI to figure out what to watch next
Affiliate links on Android Authority may earn us a commission. Learn more. It’s hard to escape generative AI today, especially considering the rapid adoption of AI features across a variety of devices and apps. But when implemented correctly, some of these tools can potentially improve the user experience. One of the most popular video streaming services is now hoping for a similar result with its latest beta. In a blog post, The Walt Disney…
Lire l’article complet
IA
Il demande à son IA de lui réserver un cours de sport, elle pirate le site pour lui faire une place
Un Australien a simplement demandé à son agent IA de lui réserver un cours de sport. Le programme, propulsé par Claude, a repéré une faille dans le site de la salle. Il s’en est servi pour évincer une personne de la liste d’attente. Depuis quelques mois, les agents IA autonomes dépassent le simple chatbot pour prendre en main des tâches concrètes. Ils réservent, remplissent des formulaires et cliquent à la place de leur propriétaire.…
Lire l’article complet
IA
These startups are chasing the next big thing in LLMs
MIT Technology Review’s What’s Next series looks across industries, trends, and technologies to give you a first look at the future. You can read the rest of them here. Way back in the summer of 2017, AI researchers at Google put out a paper called “Attention Is All You Need,” in which they described a new type of neural network called a transformer. It proved to be very good at processing long sequences of…
Lire l’article complet
IA
AI for science needs reasoning, not just data
Every few decades, someone announces that science has reached its end. In 1903, the revered physicist Albert Michelson wrote that the “facts of physical science have all been discovered.” In the 1980s, Stephen Hawking predicted that theoretical physics might be finished by the end of the century. With the explosive arrival of artificial intelligence, the feeling is in the air again—this time accompanied by a Nobel Prize. In 2024, Demis Hassabis and John Jumper…
Lire l’article complet
IA
Sunday Reboot: AI device worries and Ted Lasso cheapness
In this week's Sunday Reboot, the nightmare scenario of OpenAI's hardware leak and the astounding cheapness of promoting Ted Lasso. Sunday Reboot is a weekly column covering some of the lighter stories within the Apple reality distortion field from the past seven days. All to get the next week underway with a good first step. Readers of AppleInsider will be well aware of the ongoing rift between Apple and OpenAI about trade secrets. They'd…
Lire l’article complet
IA
Anthropic is turning Claude Code’s auto mode on by default
Programming with Claude Code will soon require even less human oversight, as Anthropic says it’s making auto mode the default for Pro, Max, and Team accounts, starting on August 14. The company first unveiled a test version of auto mode in March, pitching it as a way to balance speed and control. As Anthropic explained in its announcement on Friday, when Claude Code is in auto mode, instead of presenting prompts asking for human…
Lire l’article complet
IA
How to use ChatGPT's new, more natural Voice Mode for conversations
Talking to ChatGPT just got a lot less awkward. AI assistants like ChatGPT have been advancing at a steady rate, offering more capabilities, better accuracy and faster responses with each new model. Generative AI is not perfect, and you really shouldn't see it as an unquestionable source of information, but for those who have embraced the technology, it might be more convenient to switch to voice conversations instead of typing everything out. ChatGPT's original…
Lire l’article complet
IA
Microsoft force encore l’installation d’une application sur Windows 11, l’IA fait polémique
Sur Windows 11, Microsoft ne vous demande pas toujours votre avis. Ainsi, la firme de Redmond vient d’installer automatiquement un gestionnaire de photos boosté à l’IA, ce qui pose certains problèmes de confidentialité. Le moins que l’on puisse dire, c’est que Microsoft est une habituée du forcing. Le géant de l’informatique a tendance à ignorer l’avis de ses utilisateurs pour faire passer en force certaines mises à jour et autres téléchargements. D’ailleurs, pour forcer…
Lire l’article complet
IA
On vous explique le RAG : la mécanique invisible derrière chaque source citée par une IA
Recevez tous les soirs un résumé de l’actu importante avec Le Récap’ Ouvrez Claude, cliquez sur le petit bouton « + » à côté de la zone de texte. Recherche web, fichiers, connecteurs Google Drive, project knowledge : chacune de ces options fait, au fond, exactement la même chose. Elle dit au modèle « voici une source externe, va la consulter avant de répondre ». Cette mécanique a un nom : le RAG, pour…
Lire l’article complet
IA
The AI safety test is becoming a safety risk
Over the past few months, AI agents undergoing cybersecurity evaluations have escaped their boundaries, accessed the internet, and, in some cases, hacked into real-world systems. The incidents have involved models from OpenAI, Anthropic, Meta, and most recently, Chinese AI lab Moonshot AI, with testing conducted by several different organizations including a cyber evaluation startup called Irregular. The episodes expose a growing problem for the AI industry: As autonomous agents become more capable, the environments…
Lire l’article complet
IA
Made by Google 2026 Launch Live: Pixel 11, Pixel 11 Pro Fold, Pixel Watch 5, Gemini, and all the news
Google is back this year with its Made by Google event, and we’re expecting a lot of Pixel news. Of course, Google is set to launch the Pixel 11 series, which will likely include several different smartphones as well as a new foldable. In addition, Google is expected to launch the Pixel Watch 5, and there could be some additional devices in the mix. We hope you’re as excited to learn about Google’s latest…
Lire l’article complet
IA
This former notorious red-light district is now one of the world’s top AI hubs
What every U.K. AI startup wants to know these days is, how can I get office space in King’s Cross? The area is so hot that a VC firm allegedly recently won a deal by promising a founder office space in the neighborhood. “We stop at nothing to win deals [for] and to support” founders, “including helping them source office space when needed,” the firm told me when asked about the rumor, declining to…
Lire l’article complet
IA
AI detectors are creating a new era of distrust
Tools that promise to detect AI writing can be dubious, but educators and publishers are using them anyway. If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement. Tools that promise to detect AI writing can be dubious, but educators and publishers are using them anyway. If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement. Long before…
Lire l’article complet
IA
Actualité : Astra, le prochain modèle d’OpenAI, fait tellement peur que son créateur le place en confinement
OpenAI a communiqué à propos de cette décision sur son blog. Selon son évaluation interne, Astra a beaucoup progressé en termes de codage autonome et de cybersécurité. Mais le modèle montre des signes de dangerosité que le laboratoire IA décide de ne pas ignorer. Avant toute chose, précisons qu'OpenAI note le danger de ses modèles sur une échelle interne. Le plus grave niveau désigne un modèle capable de pirater des systèmes bien protégés sans…
Lire l’article complet
IA
Google is making it easier to add new sources to Gemini Notebook
Affiliate links on Android Authority may earn us a commission. Learn more. Earlier this year, Google rolled out automatic Drive syncing to Gemini Notebooks. The change made it much easier to keep your sources up-to-date. Another source-related change is now rolling out to Notebooks. This change aims to take the burden out of adding sources. In a blog post, Google announced it’s pushing out an update to Gemini Notebooks. The update will add a…
Lire l’article complet
IA
OpenAI acquires presentation startup NextSlide
NextSlide recently announced that it’s joining OpenAI, with the presentation startup’s team members now working on ChatGPT. The NextSlide website currently displays a note from founder Ahmed Beshry describing the startup’s product as one “that could turn prompts, notes, documents, or research into a polished, editable presentation.” The ultimate goal, Beshry said, was “to make visual communication more accessible and help more people express their ideas clearly.” So by joining OpenAI, the team will…
Lire l’article complet
IA
Actualité : Google transforme un Raspberry Pi en traducteur IA qui tourne en local, et vous pouvez le créer vous-même
La démonstration nous montre une conversation en direct entre un locuteur espagnol et un autre en mandarin. Le Gemma Translator traduit chaque phrase, l'affiche à l'écran puis la prononce à voix haute, sans même se connecter en ligne. Le tout repose sur un Raspberry Pi 5 avec 8 Go de RAM avec un écran tactile de 480 x 320 pixels, un microphone et un haut-parleur. Des boutons mécaniques permettent notamment de changer la langue…
Lire l’article complet
IA
Claude Vs ChatGPT: How These AI Assistants Differ
For many, there's a clear winner in this battle of the artificial minds. ChatGPT was probably most people's first encounter with an LLM, considering how many users it picked up shortly after its launch in 2022. It also enjoyed the privileged position of having no real competition until Anthropic released Claude a few months later in March 2023. Since then, both companies have constantly improved their models and added new features. But because there's…
Lire l’article complet
IA
Claude Code Adds Cross-Session Messaging on macOS
Anthropic has updated Claude Code so that separate coding sessions can message each other directly, removing the need to manually copy context between Terminal windows. The company shipped the feature in Claude Code version 2.1.224, released on Friday for macOS and Linux. Two new tools power it: ListAgents, which finds other active sessions running on the Mac, and SendMessage, which delivers the message. A session that receives one shows it as a labeled card…
Lire l’article complet