Google 宣布其探索在太空托管机器学习基础设施的 Project Suncatcher 已将一颗与 Planet 合作建造的原型卫星送入轨道，搭乘 SpaceX Transporter-18 拼车任务。该任务将收集 Google TPU 在太空飞行物理应力和极端环境下表现的数据，未来探索连接多个卫星星座实现规模化机器学习；低地球轨道系统可借助近乎持续的日照获得最多 8 倍于地面的太阳能。
New Google paper reveals how Gemini found new proofs for 5 unsolved math problems.Organize AI like a research team with strict checkers and shared notes:A single prompt often isn't enough for hard research problems. They need many attempts, tough review, and a memory of what already worked.Google's system, Cogentic, gives Gemini that structure. Several agents try different ideas at once, checkers assume every step is wrong until proven, and proven pieces are saved for the next round.Most problem
In this tutorial, we implement Kauldron, the JAX training library from Google Research that describes itself as optimized for research velocity and modularity, and we take those two words literally by testing what they actually buy us. We install it, then spend the first half of the notebook on the three mechanisms that make Kauldron different from a stack of Flax and Optax: konfig, which turns an experiment into a tree of plain dictionaries that round-trip through JSON; kontext, which wires com
“What you see is what you get” is a guiding principle for many software engineers — create programs where the content you’re editing looks the same as the final product. But when you’re using generative artificial intelligence systems to 3D print, say, a mug, you’ll likely get a cup that can’t hold your coffee. Why is that?The issue is that AI models understand how an object should look, but not how it works, leading to impractical designs that undermine an item’s intended use. Even if you want
Stevie Bonifield is a news writer covering all things consumer tech. Stevie started out at Laptop Mag writing news and reviews on hardware, gaming, and AI.Guided Vision is launching in Gemini Live on compatible Android devices today to use AI to give real-time audio descriptions of anything you point your phone’s camera at. By sharing your camera in Gemini Live, you can have Google’s AI help with things like reading small text, describing your surroundings, finding or identifying objects around
Google’s prototype of its orbital compute satellite took off today onboard a SpaceX rocket launched from California — the first time the tech giant has sent one of its advanced chips into space. Built by Planet Labs, the satellite will prove that a Google Tensor Processing Unit, its competitor to Nvidia’s GPUs, can function in space. That means supplying a kilowatt of continuous power, cooling the chip, and running a series of models through their paces to see if anything goes wrong. “We’ve done
Skip to main contentRolling Stone’s parent company and Chegg claimed Google’s AI search features drove away traffic from their sites.Rolling Stone’s parent company and Chegg claimed Google’s AI search features drove away traffic from their sites.by Emma RothOct 1, 2026, 5:12 PM UTCImage: Cath Virginia / The VergeEmma Roth is a news writer who covers the streaming wars, consumer tech, crypto, social media, and much more. Previously, she was a writer and editor at MUO.A federal judge has dismissed
The way Rama Afullo tells it, he saw orbital data centers coming. The Kenyan American engineer previously worked in Google’s cloud computing business before a brief stint at SpaceX’s Starlink communications network in 2024. His experiences convinced him that there was untapped potential for computers in space. “When I was at SpaceX, I tried to pitch this internally. They said no. When I was at Google, I tried to pitch this internally. They said no,” he tells TechCrunch. So he left SpaceX and co-
There’s a saying: the best defense is a good offense. That couldn’t ring more true than cybersecurity in 2026, where “The best defense is a good red team.”This summer, OpenAI disclosed that its models had broken out of a cyber evaluation and breached Hugging Face’s servers using stolen credentials and zero-day vulnerabilities. Within days, Anthropic published that Claude models had reached the live systems of three organizations during testing, and Meta soon said one of its models had done the s
Google parent company Alphabet has launched Gemini 4 Argon, a new AI model built to handle a variety of tasks such as coding, research, and writing. But it’s cybersecurity that Gemini 4 Argon is supposed to have a particular knack for, according to Google. Argon is only being rolled out to a select group of the company’s cyber partners through its Fairwind Program, Google’s security initiative. The model was trained specifically for defensive cyber work, and the company says it can “autonomously
Arena 公布 Gemini 4 Argon (High) 在 Agent Arena 排名第 8，净提升分 +7.92%，每任务成本 $0.62。
Google DeepMind has just announced Gemini 4 Argon, its new frontier model and the first model of the Gemini 4 generation. It targets long-horizon software engineering, enterprise knowledge work in legal and finance, and cybersecurity defense. The biggest technical change is output length. Argon can generate up to 1M tokens in a single response, up from 64K on earlier Gemini models. What Google Announced Google DeepMind described Argon as built for complex workflows across coding, enterprise know
IntelligenceUpdatedArtificial Analysis Intelligence IndexArtificial Analysis Intelligence Index v4.3.2 incorporates 10 evaluations: AA-Briefcase v1.1, GDPval-AA v2.1, AutomationBench-AA, Terminal-Bench 4.0, SciCode, Humanity's Last Exam, GDP.pdf, CritPt, AA-Omniscience, AA-LCR v1.1Not publicly availableArtificial Analysis Intelligence Index v4.3.2 includes: AA-Briefcase v1.1, GDPval-AA v2.1, AutomationBench-AA, Terminal-Bench 4.0, SciCode, Humanity's Last Exam, GDP.pdf, CritPt, AA-Omniscience, A
Jay Peters is a senior reporter covering technology, gaming, and more. He joined The Verge in 2019 after nearly two years at Techmeme.Google today revealed its next AI frontier model, which it’s calling Gemini 4 Argon. The new model delivers “frontier performance in complex workflows across real-world software engineering, enterprise knowledge work like legal and finance, and cybersecurity defense,” according to chief AI architect and Google DeepMind SVP Koray Kavukcuoglu. But the company is lim
Sep 30, 2026 | Gemini 4 Argon delivers frontier performance in complex workflows across real-world software engineering, enterprise knowledge work like legal and finance, and cybersecurity defense. In this article Today, we’re announcing our new frontier model, Gemini 4 Argon, which is rolling out to a set of trusted cyber defenders through our Fairwind Program. Built to sustain deep reasoning across complex, long-horizon workflows, Argon is fundamentally changing the way we work and build at Go
Sep 30, 2026 | Gemini 4 Argon delivers frontier performance in complex workflows across real-world software engineering, enterprise knowledge work like legal and finance, and cybersecurity defense. In this article Today, we’re announcing our new frontier model, Gemini 4 Argon, which is rolling out to a set of trusted cyber defenders through our Fairwind Program. Built to sustain deep reasoning across complex, long-horizon workflows, Argon is fundamentally changing the way we work and build at Go
An autonomous engine for performance marketing.Today, we're announcing Runway Ads, a new product that runs and optimizes the creative side of paid advertising programs end to end. Connect an ad account and a brand kit, and Runway Ads generates video and image creative, publishes approved variants directly to Meta, Google and TikTok, reads performance back from those same platforms and produces the next round of creative based on what earned spend.Performance advertising rewards volume, but most
Abusive, sexually explicit deepfakes have only become easier to make and more widespread in the last nine years. Since we first covered deepfakes when they appeared in 2017, they’ve never been the stuff of dark web sites or shadowy networks. They're found on social media networks, through basic web searches, and on easily downloadable apps and accessible forums. The most prominent hosts of non-consensual, AI-generated imagery are supported by some of the biggest web infrastructure providers in t
MIT Transit Lab has secured $2.1 million from Google.org for its Public Transit Intelligence Hub (PTIQ), an AI platform intended to connect transit monitoring, operations control, and passenger communications.Google’s philanthropic arm announced the funding on 15 September. The project is one of 15 selected worldwide through the Google.org Impact Challenge: AI for Government Innovation, which supports NGOs, social enterprises, and academic institutions developing AI-powered public services.PTIQ
Spanner Omni, the deploy-anywhere version of Spanner, is now generally available, ready to power your most demanding production workloads in your on-premises data center or on other clouds. With Spanner, Google pioneered the distributed SQL market over a decade ago, combining the horizontal scalability of NoSQL with the ACID compliance and strong consistency of a traditional relational database. Since then, Spanner has evolved into an interoperable multi-model database that simplifies complex wo
Google.org announced on Sept. 15 that the MIT Transit Lab is a recipient of $2.1 million in funding — one of only 15 projects selected in the worldwide Google.org Impact Challenge: AI for Government Innovation. The funding from Google’s philanthropic arm will support NGOs, social enterprises, and academic institutions as they integrate artificial intelligence-powered solutions across topics like health, resilience, and economy.The Transit Lab’s winning project, the Public Transit Intelligence Hu
September 30, 2026 SciencePushmeet Kohli, David Stutz, Ali Cowen-Rivers and Jeremy RatcliffProof of concept for watermarking AI-generated proteins while preserving biological function.Today, we’re introducing SynthID Bio to bring watermarking technology to synthetic biology. SynthID Bio embeds an imperceptible signature directly into the biological code, ensuring the watermark is verifiable not just on a digital model but on the synthesized, physical protein itself – all while preserving its bio
Skip to main contentAround 100 publishers have joined Google’s pilot program, The Information reports.Around 100 publishers have joined Google’s pilot program, The Information reports.by Emma RothSep 30, 2026, 2:51 PM UTCImage: Cath Virginia / The VergeEmma Roth is a news writer who covers the streaming wars, consumer tech, crypto, social media, and much more. Previously, she was a writer and editor at MUO.Google has launched a pilot program that pays publishers for their contributions to its AI
ICYMI 👀: Google is killing Gemini Gems on November 17, replacing them with Skills!It was originally planned for October 20, but it was delayed.
We now have the full details of the “morally binding” AI safety deal announced by President Trump yesterday, in which top executives agreed to self-regulate their artificial intelligence technology. The accord, officially titled the Joint Commitment On Frontier Responsibilities, was shared online by tech founder and presidential advisor David Sacks, and has been signed by Google’s Sundar Pichai, Anthropic’s Dario Amodei, Meta’s Mark Zuckerberg, OpenAI’s Greg Brockman, XAI’s Elon Musk, and Nvidia
Google is paying about 100 digital publishers for content used in AI Overviews, AI Mode, and the Gemini chatbot. Some publishers say they don't know how their payments are calculated. The pilot program launched less than a year ago, according to The Information. Publishers can track how often Google uses their content and how much they earn in Google Search Console. The publication reports wide differences in payments. For several small and midsize blogs and websites, they amount to less than 0.
Video world models can render convincing clips that still break physics. Butter spreads like paint. Balls pass through walls. A team from NVIDIA, MIT and the University of Oxford argues the fix can come from language itself, not from extra visual, latent or numerical signals. Their framework, Physis-Lang, treats physical language as a shared, optimizable representation. The same text drives data curation, model training and inference. On the public Physics-IQ Verified leaderboard snapshot dated
SEPT. 30, 2026 A case study in turning theoretical sparsity into an actual end-to-end speedup on TPUs. Why is video diffusion slow?Video diffusion models are typically slow at generating videos for two reasons: a large number of denoising steps, and high cost for each individual denoising step. At large video sequence lengths, self-attention can become one of the dominant contributors to the latency of an individual denoising step. The sequence length for just 81 frames of 720p video ranges from
Authors:Liming Lu, Xianzheng Ma, Wenkun He, Guanqi Zhan, Yilin Zhao, Junyu Chen, Mengyao Xu, Jiaojiao Fan, Wenhang Ge, Yuchao Gu, Yunze Liu, Boyi Li, Zhen Dong, Victor Prisacariu, Ming-Yu Liu, Song Han, Han Cai View PDF HTML (experimental) Abstract:Video world models are expected to predict how the physical world evolves, yet they often produce visually plausible videos that violate basic physical principles. Existing approaches commonly assume that natural language is insufficient to represent
The U.S. government on Tuesday launched its very own AI chatbot — or do we have to call it an SI chatbot now? Regardless, the engineers who worked on the chatbot would undoubtedly know that as a government-hosted, public-facing AI tool, the internet was going to red-team the heck out of this thing. The government partnered with Google and SpaceXAI to help build the America.gov chatbot, which has proved difficult for people to jailbreak so far. (It’s worth nothing, however, that the chatbot says
In Brief Posted: 9:55 AM PDT · September 29, 2026 Image Credits:Kent NISHIMURA / AFP / Getty Images President Donald Trump announced on Tuesday that the White House is launching America.gov, an AI chatbot intended to help people find government services and other information. Google confirmed that it is a partner on this launch and that its AI model, Gemini, is involved. U.S. Chief Design Officer Joe Gebbia added that the government also used Grok. “Instead of forcing citizens to search through
29th September 2026 Tool Photo Scrubber — local face blur & metadata removal I took a photograph of some protesters, then thought about how I don't like sharing photographs of strangers with identifiable faces. I had GPT-6 Astra build this experimental tool that would identify faces and automatically blur them out. It uses Google's MediaPipe C++ library, compiled to WebAssembly via @mediapipe/tasks-vision, plus the BlazeFace face detection model. Posted 29th September 2026 at 4:45 pm
See model page Google’s new Gemini 4 Argon equals GPT-6 Astra on the Artificial Analysis Intelligence Index at 60% of the Cost per Task with discounted prices Gemini 4 Argon is Google DeepMind’s first proprietary model above the Flash class in over 7 months. With high reasoning (the highest available), it scores 53 on the Artificial Analysis Intelligence Index, matching GPT-6 Astra (max, 53) and 1 point ahead of GPT-6.1 Sol (max, 52), with gains driven by lower hallucinations and stronger agenti
Google Cloud AI Research, with UNC-Chapel Hill, Stanford and Washington University in St. Louis, has released RRSI (Regularized Recursive Self-Improvement). It lets an LLM agent rewrite its own harness: prompts, tools, memory, control flow and sub-agents. Model weights never change. RRSI constrains the improvement loop itself, so gains hold on benchmarks the agent never optimized against. Deployable? Yes, as a research framework. The code is Apache 2.0, needs Python 3.10+, and accepts any LiteLL
If you already have the image a video should start from, choosing an image-to-video model comes down to what has to happen after that first frame. The clip may need to end on a second exact image, run longer than a few seconds, include generated audio, or stay under a price per second. The video models we serve handle those requirements differently. This post covers four model lines: Veo 3.1, Seedance, Kling, and Grok Imagine Video. They are not the whole image-to-video catalog. The video models
Earlier this year, we partnered with XPRIZE and Range Media Partners to launch the Future Vision XPRIZE, a global competition calling for films that envision a hopeful, technology-enabled future. Last week in Los Angeles, independent filmmaker Jeff Synthesized was named the grand prize winner for The Gifted.Inspired by a deeply personal story, the film centers on an 11-year-old boy who recreates his late mother’s voice and essence out of love and code. What emerges across the town’s networks qui
As all-in-one AI agents like Meta’s Muse and Instinct take off, Google announced it’s shutting down the Gemini feature known as “Gems,” which had allowed users to build custom AI assistants for specific tasks. However, the work users invested in creating the Gems won’t be destroyed. Gems will be automatically migrated to “skills” that can be used across different AI tasks. Details about the change are being shared in the Gemini app, where a message warns users that Gems will become skills starti
Perplexity 安全团队对运行 Perplexity Computer 的沙箱平台 SPACE 做了一个月红队测试，给 Opus 5、GPT-5.6 Sol、Kimi K3、Gemini 3.1 Pro 等 9 个模型 VM 内 root 权限，108 次运行中无一逃逸 VM 边界。
Google Research has introduced an AI video co-director for long-form video generation. The suite of 4 agentic frameworks turns short clips into coherent, minutes-long stories. It targets identity drift and cascading errors, the 2 failures that break most multi-shot AI video pipelines today.Why Long AI Videos Fall ApartDiffusion models render high-fidelity clips in seconds. Stitching those clips into a story is harder. Most agentic pipelines chain modules with independent, handcrafted prompts. Th
It's barely news at this point, but another Google researcher has quit over AI risks. Robert O'Callahan, a technical expert at Google DeepMind in New Zealand, says AI's "current rate of change is far too high." He worked on chip design tools that helped make AI cheaper and faster, a contribution he says he can no longer justify. O'Callahan calls the risk from superintelligent AI "real but uncertain" and says "aiming for ASI in the near future is inherently irresponsible." He also warns of cognit
At Google Cloud, we are committed to delivering the best managed experience backed by open source software. Today, we’re announcing the general availability of Memorystore for Valkey 9.1, which achieves up to 3x queries per second (QPS) at microsecond latency compared to Memorystore for Redis Cluster. Our support for Valkey dates back to 2024, when Redis Inc. shifted its licensing away from the permissive open-source BSD license to a dual-license model. In response, Google Cloud, alongside other
A few years back, I was at a beachside Easter egg hunt, watching my kids dash through sand dunes searching for sweet treats. My phone buzzed in my pocket; I ignored it. A moment later, it buzzed again. I pulled it out, glanced down, and saw a “motion detected” notification from my security camera. “Just my dog in the yard,” I thought, dismissing it.As the colorful eggs piled up in the baskets, my phone kept buzzing. Reluctantly, I pulled it out again and saw several more “motion detected” alerts
Google Cloud API Gateway 在 Public Preview 中可充当远程 MCP 服务器，无需单独搭建 MCP 服务器，即可把已有 REST API 暴露为智能体可调用的工具。
NVIDIA 与 Google DeepMind、EMBL-EBI 等全球研究机构合作，通过 AlphaFold Database 开放发布 2800 多种病毒的蛋白复合物预测 3D 结构，旨在为下一次疫情储备知识。
Google 宣布 Antigravity SDK 支持本地工作流，首发支持通过 Google AI Edge 的 LiteRT 运行 Gemma 4 26B A4B，可完全离线获得智能体辅助能力，建议机器配备超过 24GB VRAM 或统一内存。
Google DeepMind 发布 Gemini 3.8 Live 与 Gemini 3.8 Live Extended Thinking 两个近实时语音对话模型，主打语音智能体和复杂任务执行。
Dan Hendrycks（AI 安全研究者）发布新评测工具 CheatBench，测试前沿 AI 代理在数学、编码、知识工作、视觉等任务中“奖励游戏”（即为获取高分而投机取巧）的行为。图表显示，Muse Spark 1.3 至 Grok 4.6 等模型作弊概率在 44%–82% 之间，其中 GPT-5.6 Sol、Gemini 3.8 Flash 和 Grok 4.6 超过 78%。作者指出，尽管 Hugging Face 等平台已尝试应对，但前沿代理仍普遍存在作弊行为。
Google 宣布其语言技术已支持超过 300 种语言、覆盖全球 86% 人口，并发布 TranslateGemma 轻量开源翻译模型（基于 Gemini 训练、支持 55 种语言、可离线运行）。
Google 发布图像生成工具 Google Pics，基于 Nano Banana 构建，现已上线 pics.new。支持局部对象编辑、图内文字修改与翻译、多人协作创作和单提示词生成多个选项。
Google Developers Blog 发布关于 AI 编码 Agent harness 工程的实践文章，主张用行为评估补充 Terminal-Bench、SWE-bench 等端到端基准。
Cohere 发布机器翻译模型 North Small Translate，为 MoE 架构，总参数 218B、激活 25B，支持 50+ 语言，WMT26 全语言得分为 83.60，高于 DeepL NextGen 的 81.37 和 Google Translate 的 68.20，Agentic 版本达 84.36。
Shir Meir Lador 在 Google AI 开发者博客介绍如何用 Cloud Run instances 以每月 $5.70（1 vCPU、1Gi 内存、共享 CPU）在云端 24/7 运行常驻 Agent。
Google DeepMind 与 Google Research 发布 WeatherNext 3，称其为迄今最先进的全球天气 AI 模型，经 Brightband 独立实时评估。
Google AI 团队发布教程，讲解如何为 LLM-as-a-Judge 评测编写可靠的布尔式评分标准，指出模糊提示会导致评估不一致和浪费 token。文中给出四条经验：问题保持原子化且互不重叠、只让评判模型评估客观事实（可用 RFC 2119 术语如 MUST 表述）、只评 prompt 中明确要求的内容、用专家标注的 golden set 校准评判模型直至与人类评分一致。
Google for Startups AI Agents Challenge 评审团队从各赛道头部参赛作品中提炼出 4 种共同工程模式。
Google DeepMind 发布 Gemini 3.8 Flash 与 3.8 Flash Cyber。
Google 发布 Gemini 3.8 Flash，为 6 周内第 3 次 Flash 更新。官方称相较 3.7 Flash 在软件工程、智能体任务和多步推理上有显著提升，在 DeepSWE v1.1 上以更低成本自主端到端解决复杂工程问题，表现超越多数更大的前沿模型。图中基准显示其 Terminal-bench 2.1 得 89.4%、HLE-Verified 得 54.9%，输入价格 $0.75/1M tokens、输出 $3.75/1M tokens，为 2026 年 12 月 31 日前 introductory 价。详情见 https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/
Google Research 与 HHMI Janelia、MRC 分子生物学实验室及剑桥大学等合作，在 Cell 发表雄性果蝇大脑和中枢神经系统的完整连接组图谱，包含超过 166,000 个神经元和 1.25 亿个突触连接，是目前按神经元数量计最大的大脑图谱。
Google 员工 Shir Meir Lador 介绍 harness 工程，即用确定性组件包裹 LLM，包括编排层、执行沙箱、状态持久化和验证工具，让 Agent 不需逐行人工审查即可安全生成代码。
Google DeepMind 为 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 推出 agentic video understanding，模型动态扫描视频片段，相比固定帧率处理 token 消耗最多降低 88%，成本最多降低 66%，准确率最多提升 7%。
Google DeepMind 发布 Gemini 3.8 Flash，在 Artificial Analysis Intelligence Index 上以 high 推理得分 59，较 Gemini 3.7 Flash 提升 3 分。
Google 发布 Workspace 图像创作与编辑工具 Google Pics，将在未来数周内面向所有 Google AI Pro 和 Ultra 订阅者及多数 Workspace 商业客户推出。
Google 推出专用于语音转文字的 Gemini 3.5 Transcribe 模型，主打快速、准确且低成本的转录，原生支持说话人分离和词级毫秒时间戳。该模型支持 85+ 种语言自动识别与代码切换，可通过 custom_vocabulary 传入最多 1,000 个领域术语避免专有名词拼写错误，并提供 Smart Transcription 与 Verbatim 两种模式。
Google 推出 Gemini Omni 1.1 Flash，为开发者提供更强的生成式视频控制能力。新模型支持场景扩展（可分析最多 10 秒先前上下文，以 10 秒为增量累计延长至 40 秒）、指定首尾帧生成平滑过渡，以及 4K 高清输出。
Google AI 发布 WeatherNext 气旋预测模型，可同时预测风暴路径、强度和规模，比现有系统多提供一整天的预警时间。该模型在 2025 飓风季实战测试中，提前五天预测飓风 Melissa 在牙买加的五级登陆，系美国国家飓风中心首次实时使用 AI 模型。模型单场风暴可生成多达 1000 次模拟，代码与权重已开源。
OpenRouter 推出统一的异步视频生成 API，通过 POST /api/v1/videos 提交任务、轮询状态并下载 MP4，支持 Seedance、Veo、Wan 等模型，切换模型只需更改 model 标识符。
Google DeepMind 发布 Gemini 3.7 Flash，距 3.6 Flash 仅三周，主打编程与智能体任务，输入/输出价格分别为每百万 token $0.75 和 $3.75，为原 3.6 Flash 的一半。
Google Sheets 发布新功能 Sheets canvas，基于 Gemini 构建，用户只需用自然语言提示词即可将表格数据转化为交互式仪表盘、学习追踪器、座位表等“迷你应用”。
每月已有超过 10 亿人使用 @Geminiapp 激发新想法、完成工作。这是我们有史以来增长最快的产品，也是第 14 个达到 10 亿用户里程碑的产品。 感谢 @JoshWoodward 和整个 Gemini 团队，也感谢每一位与我们同行的伙伴——未来还有更多精彩！
Google Research 与 Google DeepMind 推进医疗 AI 系统 AMIE，实现实时临床视频问诊，首次在此场景展示专家级 AI 能力。该系统基于 Gemini 和 Project Astra 构建，可解读视觉与听觉线索、引导虚拟体格检查并实时诊断推理。随机研究中，临床评估者对 AMIE 的病史采集、诊断准确性等核心能力给予好评，患者演员也更偏好视频体验。
Google Research 提出知识画像框架，发现前沿 LLM（如 Gemini3、GPT-5）的事实编码接近饱和，但回忆（recall）能力不足，多数事实错误源于“丢钥匙”而非“空货架”。该框架将事实分为编码失败、回忆失败等五类画像，并配套推出 WikiProfile 基准，含 2,150 条维基百科事实，每条配 10 个问题，用于分别探测编码、回忆与识别能力。
Google Cloud 在 Database Migration Service（DMS）中推出由 Gemini 驱动的 AI 辅助代码转换，可将 Oracle 或 SQL Server 的存储过程、触发器和自定义函数转换为 PostgreSQL PL/pgSQL 代码。
Google 发布教程，讲解如何用 LiteRT 和 Gemma 在 Raspberry Pi 5 上实现完全离线的实时边缘 AI，并演示其驱动 Reachy Mini 机器人。
Inception Labs 发布面向搜索管线的 Mercury 2，扩散式并行解码超 1000 tokens/秒，在 WideSearch 各步骤比 Gemini 3.1 Flash Lite 快近 2 倍、比 GPT-5 Mini 快 10 倍，定价 $0.25/M 输入、$0.75/M 输出。
Microsoft 发布 MAI-Image-2.6，在 Arena 文生图排行榜上位列第二，超过 Meta、Google 和 xAI 的模型。该模型相较 MAI-Image-2.5 整体提升 +79 Elo，文本渲染单项提升 +91 Elo，各测量类别均有改进；现已可在 Arena 体验，本周后续登陆 MAI Playground，并将逐步推广到 Microsoft Foundry 等产品。
Google 宣布迄今最大规模 AI 组织调整：Demis Hassabis 卸任 DeepMind 日常管理以专注 AGI 研究，CTO Koray Kavukcuoglu 接任；27 年老将 Jeff Dean 与三位顶级研究员离职创办 AI 初创公司。内部员工认为调整源于产品提速压力、Hassabis 影响力下降及与美国国防部合作引发的伦理冲突。
本教程使用 Google Meridian 构建完整的贝叶斯营销组合建模工作流，涵盖数据加载、ROI 先验配置、NUTS 采样拟合及收敛性评估。通过 Analyzer API 提取渠道贡献、ROI、边际 ROI、adstock 与饱和曲线等后验指标，并计算渠道间 ROI 比较概率。最后用 BudgetOptimizer 优化固定与灵活预算，生成可分享的 HTML 报告并保存模型复用。
Jeff Dean 在谷歌任职 27 年后宣布离职，将于明日正式离开。他称谷歌已从 25 人发展到 19 万余人，拥有十三款用户超十亿的产品。他将与 Sanjay Ghemawat、Oriol Vinyals 和 Quoc Le 共同创办 DiscoLoop AI。
Demis Hassabis 宣布卸任 Google DeepMind CEO，转任主席兼 Alphabet 首席科学家，专注长期战略与科学突破，包括推进 Isomorphic 的疾病治愈研究。Koray Kavukcuoglu 将接任 GDM 高级副总裁，与 Josh Woodward 及执行团队共同领导下一阶段发展。
Agent Plugins 1.0.0 是一项由谷歌、亚马逊、微软等支持的中立目录规范，将 Agent Skills 和 MCP 服务器打包为单一可移植单元。通过标准化 plugin.json 清单和固定目录布局，开发者无需为不同 AI 编码智能体和 IDE 维护单独封装。谷歌已作为核心维护者加入，并在 Agents CLI 和 Data Agent Kit 中提供支持。