Actualités — Page100

Google unveiled Gemini 3 Flash, a faster and cheaper AI model built on the recent Gemini 3 architecture. The company is making Flash the default model in its Gemini app and AI‑enabled search, while still offering the Pro version for more demanding tasks. Gemini 3 Flash delivers notable performance gains on benchmark tests, supports multimodal inputs such as video, sketches, and audio, and is available through Vertex AI, Gemini Enterprise, and an API preview. Early adopters like JetBrains and Figma are already integrating the model, and Google highlights its suitability for bulk, work‑horse workloads.Lire la suite

OpenAI has named former British Chancellor George Osborne as Managing Director and Head of OpenAI for Countries, a role focused on expanding the company’s Stargate data‑center network beyond the United States. Osborne will guide government partnerships, help nations adopt OpenAI tools, and oversee the first round of collaborations with ten countries. His appointment underscores OpenAI’s strategy to deepen ties with governments and accelerate the rollout of sovereign AI infrastructure worldwide.Lire la suite

Google has launched Gemini 3 Flash, a new AI model integrated into its AI Mode that promises search‑engine speed without sacrificing the advanced reasoning capabilities of the Gemini 3 family. The model is available worldwide for free, and Google says it can handle complex queries with greater precision while maintaining the quick response times users expect from traditional search. Alongside Gemini 3 Flash, the company is expanding access to Gemini 3 Pro and the Nano Banana Pro image‑generation tool, though higher usage limits require a Google AI Pro or Ultra subscription.Lire la suite

Amazon is in advanced discussions with OpenAI about a $10 billion investment that would also see the AI startup adopt Amazon's Trainium chips and increase its use of Amazon Web Services. The proposed deal could lift OpenAI's valuation past $500 billion while deepening the company's reliance on Amazon's cloud and hardware infrastructure. At the same time, OpenAI continues to expand relationships with other technology partners, including Microsoft, NVIDIA, Oracle and AMD, prompting investors to scrutinize the circular nature of its financing and compute arrangements.Lire la suite

Google has launched Gemini 3 Flash, a new generative AI model that promises faster performance and higher accuracy than its predecessors. Available now through the Gemini app, search, Gemini API, Vertex AI, AI Studio, and Antigravity, the model delivers notable gains on academic, reasoning, and coding benchmarks while offering lower token costs. Gemini 3 Flash narrows the gap with the larger Gemini 3 Pro, delivering comparable results on many tests at a fraction of the price, positioning it as a versatile option for developers and users seeking efficient, high‑quality AI outputs.Lire la suite

Google is rolling out Gemini 3 Flash, a more efficient version of its flagship Gemini AI model, as the new default in the Gemini app and Google Search’s AI Mode. The upgrade follows the recent Gemini 3 Pro release and promises faster response times, lower operating costs, and more nuanced answers compared to the previous Gemini 2.5 Flash. Tulsee Doshi, senior director of DeepMind product, described the change as a “huge upgrade” for users, highlighting the model’s ability to process images, text, and video together while delivering answers in seconds. Developers will also gain access through a suite of Google AI tools.Lire la suite

Researchers from Purdue University and the Georgia Institute of Technology argue that the growing energy demands of artificial‑intelligence models stem from the longstanding "memory wall"—the gap between fast processors and slower memory. By adopting brain‑inspired architectures that integrate processing and memory, such as spiking neural networks and compute‑in‑memory (CIM) designs, they claim AI can become far more efficient. The study highlights potential benefits for devices ranging from medical equipment to drones, suggesting a shift away from traditional von Neumann designs could dramatically cut power consumption while preserving performance.Lire la suite

OpenAI has introduced GPT Image 1.5, a new image model inside ChatGPT that is faster and more reliable at editing and generating visuals. The upgrade adds a dedicated Images tab with presets, filters, and templates that let users create or modify pictures without typing prompts. Improvements include better handling of text, consistent element tracking across edits, and enhanced style blending. While the tool still produces occasional errors, OpenAI positions it as a professional‑level creative assistant that lowers the technical barrier for high‑quality visual content.Lire la suite

OpenAI has introduced GPT Image 1.5, a new version of its ChatGPT image generation tool that promises faster performance, better instruction following, and more precise editing controls. The model is now available to all ChatGPT users and via API. Its rollout follows an internal “code red” memo that highlighted a competitive push against Google’s Gemini series and Nano Banana Pro, which have recently outperformed OpenAI on benchmark leaderboards. GPT Image 1.5 builds on the earlier GPT Image 1 release, adding granular post‑production features and a redesigned creative‑studio interface within the ChatGPT sidebar.Lire la suite

OpenAI has released GPT-Image Model 1.5, an upgraded AI image generator that is four times faster than its predecessor and offers improved instruction following, editing tools, and legible text generation. The new model is accessible to all ChatGPT users via a dedicated sidebar space and is also available through the API. Its debut comes amid heightened competition with Google’s Gemini 3 and follows OpenAI’s recent partnership with Disney to incorporate iconic characters into future image and video offerings.Lire la suite

Google is rolling out an experimental AI assistant called CC that compiles a personalized “Your Day Ahead” briefing each morning. The service draws on a user’s Gmail, Calendar, Drive, and the web to summarize schedules, upcoming tasks, bills, and relevant updates, while also drafting emails and creating calendar links for quick action. Initially available in early‑access to paid subscribers over 18 in the United States and Canada, Google is opening a waitlist for the feature. Powered by the Gemini model, CC mirrors OpenAI’s Pulse in offering a concise daily snapshot of a user’s digital life.Lire la suite

OpenAI announced that it is rolling back the automated model router for users on its Free and $5‑a‑month Go tiers. The change returns these users to the default GPT‑5.2 Instant model, while still allowing manual selection of advanced reasoning models. The decision follows user feedback indicating a preference for the simpler default chat flow, and comes amid heightened competition from rivals such as Google Gemini and ongoing safety enhancements. Paid subscribers will continue to have access to the model router.Lire la suite

OpenAI introduced ChatGPT-5.2, highlighting a suite of new capabilities that aim to improve the model's adherence to user‑specified constraints, offer richer perspective shifts, provide more personalized recommendations, anticipate potential failures, and help users clarify vague intuitions. The rollout includes claims of better rule‑following, nuanced viewpoint generation, targeted question‑driven recommendations, risk‑aware planning, and structured hunch analysis, positioning the model as a more thoughtful and practical AI assistant.Lire la suite

OpenAI has launched a new version of its ChatGPT Images tool that is four times faster than the previous model and offers sharper instruction following. The update brings stronger capabilities for adding, subtracting, blending and transposing visual elements, as well as improved handling of dense and small text. A dedicated Images section now appears in the ChatGPT sidebar, providing preset filters and prompt ideas. The upgrade arrives amid rising usage of Google’s Gemini chatbot, positioning OpenAI’s visual generation as a key differentiator for its growing user base.Lire la suite

OpenAI’s chief communications officer, Hannah Wong, informed staff that she will leave the company in January. Wong, who joined OpenAI in 2021 and became chief communications officer in August 2024, has been pivotal in shaping the firm’s public narrative, especially during high‑profile challenges. The company will conduct an executive search to replace her, with VP of communications Lindsey Held overseeing the team in the interim. OpenAI’s leadership expressed gratitude for Wong’s contributions and wishes her well as she moves to the next chapter of her career.Lire la suite

The Walt Disney Company and OpenAI have announced a partnership that licenses hundreds of Disney, Pixar, Marvel and Star Wars characters for use in OpenAI’s generative AI tools, including ChatGPT’s image generator and the Sora video platform. Disney will also invest in OpenAI, gaining a seat at the table as the AI firm navigates regulatory and safety issues. While the deal promises new creative possibilities for fans, it also raises questions about brand protection, content moderation, and the impact on traditional animation talent.Lire la suite

Chai Discovery, an AI‑driven biotech startup backed by OpenAI, announced a $130 million Series B financing round that lifts its valuation to $1.3 billion. The round was led by General Catalyst and Oak HC/FT, with participation from Menlo Ventures, OpenAI, Dimension, Thrive Capital, Neo, Yosemite venture fund, Lachy Groom, SV Angel, Glade Brook and Emerson Collective. The new capital brings the company’s total funding to over $225 million. Chai’s platform, now in its second generation (Chai 2), aims to accelerate drug discovery by using foundation models to design molecules and de‑novo antibodies, positioning the firm as a leading player in AI‑enabled biotech.Lire la suite

Creative Commons, the nonprofit known for its open licensing framework, has issued a tentative endorsement of “pay‑to‑crawl” technology that would charge AI bots for accessing website content. While the organization acknowledges the potential benefits for publishers facing reduced traffic from AI‑driven answers, it also warns that such systems could concentrate power and limit access for public‑interest actors. CC outlines a set of principles for responsible implementation, urging openness, throttling options, and safeguards for research and education. The stance reflects a balanced view of protecting content creators while navigating the evolving AI landscape.Lire la suite

Google is rolling out live translation for all earbuds on Android in a beta program that begins in the United States, Mexico and India. The update, part of Google Translate, now preserves tone, emphasis and cadence, and supports more than 70 languages. At the same time, the company is enhancing its Gemini AI to better handle idioms, slang and local expressions. Google’s language‑learning features also receive upgrades, adding pronunciation notes, progress tracking and expanding to 20 new countries, including Germany, Sweden and Taiwan. An iOS version is planned for later, with broader availability expected in 2026.Lire la suite

Merriam-Webster has selected “slop” as its 2025 word of the year, defining it as low‑quality digital content produced in large volumes by artificial intelligence. The term captures growing concerns about AI‑generated material flooding major online platforms such as YouTube, Wikipedia, Spotify and Pinterest. While some sites work to curb the spread, others—including Meta, OpenAI and Disney—are embracing AI‑driven video services. Merriam-Webster also highlighted other notable words from the year, underscoring the cultural impact of AI on language and internet culture.Lire la suite