Search AI Products and News

DeepMind's Mixture-of-Recursions (MoR) enhances model efficiency via dynamic token routing and recursive depth allocation, outperforming Transformers with fewer parameters. Its selective caching reduces memory pressure, proving especially effective above 360M scale, offering optimized AI deployment solutions.....

Google introduces AI-powered dialing for service inquiries, upgrading search with Gemini2.5Pro and 'deep search' for complex queries, enhancing user experience.....

Google has officially launched an AI news summary feature on its Discover platform, sparking concerns within the media industry about traffic loss. The feature is now available on iOS and Android in the United States, replacing traditional news titles and source indicators with three-line AI summaries, primarily covering popular topics such as sports and entertainment. The summaries will indicate the number of sources and warn of potential errors, but the publishing industry is worried that this will worsen the Google Zero phenomenon—users can obtain key information without clicking on the original article, leading to a decrease in website traffic and advertising revenue. Previously, Google had already implemented similar features in search.

Google upgrades its AI note-taking app NotebookLM with a new public note library. The library integrates authoritative content such as The Economist and expert notes, allowing users to ask questions and receive AI summaries. New features include audio overviews and mind maps, with initial resources covering practical topics such as longevity guides and travel tips. The platform has currently accumulated over 140,000 user notes, leveraging AI technology to improve knowledge management efficiency and strengthen the information sharing ecosystem.

Google's AI summaries in search raise publisher traffic concerns. The feature, now live in the U.S., generates news snippets with citations, mainly for trending topics. While Google claims it helps users filter content, publishers report organic traffic drops as non-click searches rose from 56% to 69% since May. Despite new monetization efforts like Offerwall, traffic decline persists. Media reactions remain mixed as outlets like WSJ explore AI a....

Google launched NotebookLM with preloaded research content on 8 topics, featuring AI summaries, Q&A, and podcasts. Over 140K users shared notebooks in 4 weeks.....

OpenAI's acquisition of Windsurf failed, and Google successfully poached its CEO and core team to join DeepMind. Windsurf, formerly known as Codeium, had raised $200 million in funding and had a valuation of $1.25 billion. Google obtained non-exclusive licensing rights to some of its technology but did not acquire the company. Recently, the popularity of code-generation startups has declined, facing competition from large models like Anthropic and Google. Windsurf will be taken over by its former executives and will continue to operate independently.

Google released Gemini, the top embedding model with 68.37 MTEB score, surpassing OpenAI. Based on Transformer, it supports multilingual tasks at $0.15/M tokens, boosting AI applications like search.....

After OpenAI's $3B Windsurf acquisition failed, Google DeepMind secured a $2.4B non-exclusive tech license and poached key talent. Windsurf's founders and top researchers will join Google while the company remains independent, showcasing a new 'reverse acquisition' trend in AI where giants boost capabilities through hires and licensing. Most of Windsurf's 250-member team will stay to develop AI coding tools. The OpenAI deal collapsed due to Micro....

Google DeepMind open sources the GenAI Processors Python library, helping developers build efficient generative AI workflows. The library supports asynchronous processing of multimodal data and optimizes Gemini API application development, significantly reducing latency in real-time applications. Core features include a modular Processor interface, streaming API design, and concurrency optimization, enabling rapid development of real-time applications such as intelligent assistants. Currently only supports Python, but with an open community contribution model, future plans include expanding functionality to cover more scenarios.

Google released HAI-DEF toolkit with MedGemma (medical text/records) and MedSigLIP (image analysis) models. Supports local GPU deployment, privacy protection, and customization via Hugging Face/Vertex AI.....

OpenAI plans to launch an AI browser to challenge Google Chrome. The product is based on Chromium and integrates ChatGPT technology, with 400 million potential users. Its innovation lies in the AI agent feature that can automatically perform web operations, reducing traditional browsing steps. This move may threaten Google's 66% market share and its advertising ecosystem. The industry is currently experiencing a surge in AI browser trends, and OpenAI, leveraging its technological advantages, is trying to seize the opportunity. If successful, it could undermine Google's dominant position in user data and online advertising.

Google releases the open-source tool MCP Toolbox for Databases, simplifying the integration of AI agents with SQL databases. The tool connects to a database with just 10 lines of code and supports secure mechanisms such as connection pool management, authentication, and schema introspection. It is compatible with various Google Cloud databases. As an open-source project, it lowers the development barrier, but currently mainly supports Google ecosystem databases. Future expansion of compatibility may be needed. This tool has the potential to become a standard component for AI development, driving intelligent data processing.