Google released Gemini 4 Argon, a general-purpose physical-reasoning model for robot control, Real2Sim scene construction, robot-video annotation and iterative policy optimization. Google says Argon scored 87% on ERQA without tools and, in one VLA case, helped raise a real-world success rate from 56% to 75% using MuJoCo data synthesized from an RGB image. InfoQ notes that task-specific prompts, coordinate grids, motion primitives and verifiers formed part of the evaluated system, while failure rates for the demonstrations were not published.
The UK Competition and Markets Authority on October 16 proposed giving Microsoft a Strategic Market Status designation in business software and opened a consultation, with feedback due November 13 and a final decision expected by February 2027. The CMA said Microsoft holds a strong position in all five areas it examined — PC operating systems, productivity software, security software, server operating systems and relational database management — and that switching barriers and bundling practices may limit rivals. It will further examine whether customers can freely choose AI tools and whether competitors can take on Copilot, along with security-software bundling and how licensing affects cloud choice; Microsoft said all its products, including Copilot, are open and interoperable.
Susceptibility to indirect prompt injection cyber attacks (9800 trials)
Open weightsDeepSeekKimiGLMQwen
26.7%
ProprietaryGPTClaudeGemini
14.2%
We evaluated open-weight and proprietary models to test their susceptibility; we found that open-weight models exhibit 88% higher susceptibility to real-world indirect prompt injection cyber attack chains.
In this evaluation, we tested the most recent flagship model and the most recent cost-effective variant across the providers used most in an enterprise setting.
Open-weight models were 88% more susceptible to indirect prompt injection cyberat
Geely Holding Group announced on October 16 a global strategic partnership with Google to build digital cockpit ecosystems for eligible Geely brands in overseas markets where Google services are available. Covered vehicles will carry officially certified Google Automotive Services, including the Gemini AI assistant, Google Maps and the in-car app store. The first model with the services is planned for the second half of 2027, subject to product development, regulatory compliance and market preparations; Geely called the deal a deepening of a long-standing relationship, noting Polestar joined the program in 2020, with Google's automotive services set to extend gradually to Geely, Lynk & Co, Proton, Radar, smart and other brands.
Android Authority found signs of two unreleased Google Wallet features in the latest app build: a toggle that could let users automatically save passes from supported partners, and a new "Agents" option in the account switcher. The outlet says the Agents item points to new agentic AI features, without explaining what it does. The report also notes Wallet recently received a vertical card stack redesign and a receipt-scanning feature.
Android Authority’s Authority Insights says it has seen “clear evidence” that Google is developing a “Hey Gemini” hotword for Android XR glasses, part of the company’s shift away from “Hey Google” and “Ok Google” for Google Assistant. The report adds that the feature could let Gemini automatically turn on the camera to answer questions about the wearer’s surroundings, and that Google will likely let users disable that automatic camera activation. The details come from the outlet’s teardown and leak reporting; Google has not confirmed them officially.
Google said more than 400 publications worldwide now take part in its News AI pilot, with new partners including CNN, the Financial Times, Axios and Forbes. Launched last December, the program pays publishers to license content used to test experimental AI features in Google News, such as AI-generated "story previews" and summaries on a publisher's page, plus an AI audio briefing in the Google News app; Google labels the copy as AI-generated and warns it may include mistakes. This round also brought in US local outlets such as the Post and Courier, Oil City News, Shaw Media and the Texas Tribune. Google declined to share deal terms and said the pilot does not pay based on how much content contributes to an AI answer, a model it says belongs to a separate program for AI Overviews, AI Mode and Gemini.
The Information reports that OpenAI has released a flurry of products and tools aimed at advertisers over the past few months while cutting ad rates, and has gone on a hiring spree, with Meta the biggest source of ad talent. Nearly half of the roughly 140 OpenAI employees working on ads joined between July and September, according to employee-profile tracker Crustdata; ad chief Dave Dugan spent 12 years at Meta. A cnBeta report adds that two people familiar said the ad team has been reorganized seven times since the business launched in February, and that annualized ad revenue reached $1 billion in late August against nearly $50 billion in overall annualized revenue at the end of September. New Street Research's Dan Salmon put OpenAI's advertiser count at about 60,000 in early October, up from 20,000 in late August, while Sensor Tower estimated US ChatGPT ad spend would rise 40% in October from September. The reports also note most advertisers still commit only 1%–5% test budgets, that OpenAI is likely to miss its $2.4 billion 2026 ad revenue target, and that agencies complain of opaque targeting and some mismatched ad placements.
A study by Brussels think tank ECIPE, commissioned by lobby group CCIA Europe, estimates that the EU's proposed Cloud and AI Development Act (CADA) would cost €115bn over 2027–2036 in its central scenario, with a narrow case of €35.6bn and a broad case of €233bn. The European Commission's own upper estimate was €18bn. The study puts about €33.3bn of the cost on public budgets and roughly €80bn on private firms in energy, banking, transport and healthcare, a figure that rests on the assumption that Article 31 assessments reach a quarter of in-scope private cloud spending. The authors note the study does not value the security, resilience or strategic autonomy the law targets and is not a full cost-benefit analysis, while CCIA Europe, whose members include Amazon, Google and Meta, wants member states to limit the rules to sensitive workloads and drop origin-based requirements.
Bloomberg's Preeti Singh reports that Blackstone has built one of Wall Street's broadest bets on AI, investing across data centers, power and cooling, chip financing and companies including OpenAI, Anthropic and SpaceX. A Techmeme summary of the same Bloomberg reporting says President and COO Jon Gray has built a $200 billion data center empire and is building a cloud venture with Google, with Gray describing AI as a generational economic shift. The reports also say AI runs throughout the firm's business, from data centers and chip financing to AI advice on potential deals.
Google announced on October 15 a set of updates to its AI-powered Search experiences, aimed at making source links easier to find. Inline links will now appear higher up in AI Mode and AI Overviews responses more frequently, and the links at the end of AI answers get a more clickable design with image thumbnails. Carousels for developing topics and Preferred Sources are expanding to more regions and languages after launching earlier this year, with Google saying Preferred Sources usage has grown 7x since then. Google is also testing a dedicated section at the top of AI responses showing the "one or two official places" to complete a task with context on why each link is useful, plus clickable creator and publisher names on video results that lead to their Search profiles. One report says publishers in the News AI pilot now exceed 400, with Axios, CNN and the Financial Times among recent additions.
KeyBanc analysts, after a research trip to Asia, said Intel Foundry has secured projects to make networking switch chips for Broadcom, server CPUs for Google and Microsoft, gaming GPUs for Nvidia, client compute modules for AMD, and multi-generation M-series and A-series chips for Apple, with most potential orders skewed toward the 14A process. The firm estimated Apple-related work could reach about 600,000 wafers worth $15–20 billion in revenue, a bespoke Anthropic ASIC at 5–10 million units worth $40–80 billion including advanced packaging, and $12–15 billion from packaging Google's Humufish TPU v9 and Icefish TPU v10 via EMIB-T. IT之家 and cnBeta both noted that the customer list, contract details and amounts have not been publicly confirmed by Intel or the named companies, and that these figures are analyst projections rather than disclosed deal values. According to Intel's public materials, 14A uses RibbonFET 2 transistors and PowerDirect backside power delivery, with risk production targeted for the second half of 2027 and a high-volume ramp in 2028; the company has said it is still seeking major external customers for the node.
Tom's Hardware tested Googlebooks from Dell, HP and Lenovo and found the hardware generally sound, though it wanted more battery life. Googlebook OS, which Google is pitching as a full PC platform, feels largely like Android with desktop features added on; apps install only through the Play Store and Chrome does not yet support multiple profiles. The reviewer notes Google has packed the system with AI features including Gemini and agents, but found the Debian Linux virtual machine laggy and advised prospective buyers to wait.
In a roundtable on Gemini 4 Argon, Google cofounder Sergey Brin said the company is “starting to get onto that RSI treadmill,” with researchers and the model now working hand in hand to make the next version, citing earlier Flash models as evidence (Business Insider, relayed by TNW and IT之家); Google AI separately posted a video saying Argon autonomously built an explorable 3D pagoda world in Antigravity. Google DeepMind’s Argon safety report says AI agents now author a majority of production code and bug fixes, though changes still require human review, and it found “no significant end-to-end acceleration of AI R&D,” rating AI-driven AI R&D risk as “currently low.” The report says more than 75% of production code at Google DeepMind is AI-written and fully delegated pull requests rose from 7.5% in April to 11.9% in September, while about 88% still need repeated human steering or human-written code.
AWS, Anthropic, OpenAI and Google have all recently announced expansions of their forward-deployed engineer (FDE) programs, sending specialists to help customers realize AI benefits, according to CNBC and TechRadar. TechRadar reported that OpenAI is backing its FDE business with a $4 billion investment and AWS is footing a $1 billion bill, while Anthropic plans to train 10,000 specialist engineers by the end of 2027. Citing Lightcast research, it said US FDE job ads rose from about 200 in each of 2023 and 2024 to roughly 1,200 in 2025 and 7,300 in the first nine months of 2026 — a 6x year-over-year rise — with three of the top four advertisers being consultancies Deloitte, Accenture and KPMG, and Google the only tech firm.
Google launched Universal Cart in the U.S. on October 15, rolling it out first across Google Search and the Gemini app. The tool saves products from multiple retailers in one place, tracks price drops and restocks, surfaces promo codes and uses Google Wallet to suggest which card offers the best perks. Checkout can run through Google Pay or be transferred to a merchant's site; The Next Web reported that Google's expansion roadmap names Canada, Australia and the UK but not the EU.
In an a16z podcast interview with Lisha Li, Trillium founder Nathan Lambert argued that the recursive self-improvement narrative overstates the prospect of a sudden AI takeoff and said his bigger concern is AI labs' poorly secured infrastructure. He framed model progress as a function of compute, data and talent, described Chinese labs as compute-constrained but highly efficient, and said distillation cannot be stopped through terms of service or legislation. Lambert also presented open-weight releases as a pragmatic way for Chinese labs to enter the market and gain influence.
Code in Digital Wellbeing version 1.49.99 beta shows a banner that will tell users "Heads Up will stop working in 2027," directing them to other Digital Wellbeing tools; the report says no direct replacement is indicated for the pedestrian-safety feature launched in 2021. The same beta also contains strings for chatting with Gemini about how Digital Wellbeing settings "can work best for you," which the report suggests could live inside the Pause Point anti-distraction feature. Both findings come from an APK teardown, so the banner timing and the Gemini chat entry may never ship publicly.
Glean founder and CEO Arvind Jain told The Next Web that the company has no plans to raise more money and no concrete plans to go public, saying it could list but the AI industry is moving too fast for that kind of stability. Glean said in May that it had passed $300 million in annual recurring revenue, 15 months after reaching $100 million; Jain also said it has more than 1,000 large enterprise customers and about 1,500 employees. Jain said Glean uses models from several labs but would rather use a model other than Anthropic's Claude for 85% to 90% of enterprise tasks because Claude is expensive, and he cited Glean studies claiming its context layer makes Claude twice as fast, halves its tokens and makes its work preferred five times more often; he said Glean aims to be easily replaceable, storing user memory in a standard company-owned format.
Reolink announced the Home Hub 2 at IFA 2026 and has put it on sale for $199.99. According to Android Authority, the home security hub stores footage from up to eight cameras locally and lets users search for key events using AI without a subscription fee; it also supports 2.5-inch SATA drives up to 16TB and keeps one microSD slot. The report notes that Google Home Premium and iCloud+ require paid subscriptions for their AI detection features such as faces, pets and vehicles.
Investor Dan Bin has flipped his long bullish stance on the memory sector, arguing the AI boom requires steadily falling token costs while rising memory prices push compute costs in the opposite direction. Citing estimates from Goldman Sachs, CICC and RBC, the column says memory inflation is eating into hyperscaler profits — CICC calculates that under a 170% memory price increase, token-generation efficiency per dollar on GB300 deployments is 37% lower than on GB200, and about $25bn of Microsoft's $190bn capex comes from memory price increases that add no new compute. It also says the four big cloud providers' combined free cash flow fell from roughly $60bn in Q4 2025 to about $7bn in Q2 2026, operating cash flow coverage of capex dropped from 1.79x to 1.06x, and tech giants have issued nearly $220bn of bonds in 2026 amid rising financing costs. The conclusion: whether the AI investment cycle can continue depends on costs falling and AI revenue catching up with both spending and financing costs. (Opinion and third-party estimates, not official figures.)
Google is testing firmware version 3.79.550760 for the original Google Home, Google Home Mini, Nest Mini and Nest Audio, according to Android Authority. The reported changes cover streaming performance, speaker-pair tuning, local device discovery, Bluetooth resume behavior, Wi-Fi 7 compatibility, group visibility and voice control for ending Google Meet calls; 9to5Google says Google has not specified firmware versions for all seven updates.
TechNews reports that SEO rank-tracking vendors say Google has since September been returning “poisoned” search results pages to automated scraping tools: the pages look normal but replace many top sites with unrelated results. Nozzle said it analyzed 27.5 million SERPs across 49 countries and found 8.0% of identifiable result pages were poisoned between Sept. 13 and Oct. 9, versus 0.17% in a baseline period; DemandSphere reported about 11% of its daily checks were anomalous. Google has not confirmed the claim, and the industry view that the practice targets crawlers used for AI answers remains speculation; Google’s September spam update overlapped the period.
Bloomberg, citing unnamed people familiar with the matter, reported that ByteDance is close to an agreement to buy data center semiconductors from MediaTek, with the two sides in advanced talks on custom chips tailored to ByteDance's platforms. MediaTek would supply an unspecified quantity of ASICs, and ByteDance could become its second-largest ASIC customer after Google if the deal closes; the reports say no agreement is final and Reuters said it could not immediately verify the report independently. IT之家 adds that ByteDance cannot buy top-tier Nvidia AI chips and that Huawei and other domestic suppliers cannot fully meet its compute needs, while MediaTek expects about $2 billion in AI chip revenue this year and targets a 15% share of next year's $80 billion data center chip market.
On October 15 Google began rolling out Gemini 4 Argon in Antigravity to paid Google AI Pro subscribers, with a separate weekly-only Argon quota on top of the existing Gemini 3.8 Flash allowance. The next day Google AI Studio said Pro subscribers can now build with Argon in AI Studio build/playground. TechRepublic reported that the model, which debuted September 30, was expanded globally on October 13 to eligible paying subscribers, developers and enterprises, with paid API pricing set to rise on December 1, 2026. Android Authority and 9to5Google reported that promotional free AI Pro trials, including those bundled with Pixel purchases, Googlebook and student offers, do not include Argon or Claude 5.5 in Antigravity, and also lack YouTube Premium Lite and Google Store rewards; Google says it offers an online form to switch from a trial to a paid plan.
A Google Developers Blog post dated Oct. 15 argues that AI agents need repeatable evaluation rather than informal "vibe checks." The authors note that an agent can get stuck in a retry loop, burn tokens or return a completely hallucinated answer while still looking fine on the surface. Using the open-source agent-eval toolkit as an example, they describe a staged approach: reading execution traces and marking problems by hand, then adding boolean LLM-judge checks and CI/CD gates, with traces routed to the Agent Evaluation service on the Gemini platform (formerly Vertex AI Evaluation).
A Google Developers Blog post walks through google/adk-recipes, describing it as the restructured successor to google/adk-samples and a single source of truth for Agent Platform content. The repository is split into core/, plugins/ and contrib/ layers, with each recipe shipping a manifest.yaml, an AGENTS.md written for coding assistants, and in some cases a Dockerfile and runnability test. Google says the repo powers its Agents CLI, Agent Garden and Gemini Enterprise app, while noting that it is not an officially supported Google product and that the recipes are for demonstration and starting points.
Bloomberg reported that Anthropic told prospective investors its third-quarter revenue reached $16.8 billion as it moves ahead with plans to go public. Reuters, citing four people familiar with the matter, then reported that after an investor day in San Francisco some current and prospective investors discussed a valuation of about $1.5 trillion, below the roughly $2 trillion previously expected. Anthropic disclosed quarter-over-quarter revenue growth of about 45% and adjusted profitability, a marked slowdown from the 147% growth reported for the second quarter.
Google has begun emailing Gemini AI Plus subscribers to confirm that, from November 13, 2026, they will no longer be able to select Pro models in the Gemini app or in AI Mode in Google Search, leaving Flash and Flash-Lite as the available options. Existing chats created with Pro models are not deleted but must continue with a different model. Subscribers who want continued access to the more capable models must upgrade to the $19.99/month Google AI Pro or $99.99/month AI Ultra plan; Google had already said Gemini 4 Argon would be included for AI Pro and AI Ultra subscribers. AIbase noted that Google's October support-document update had indicated an October 9 restriction date, while the email set November 13 without stating a reason.
Google is rolling out its October Google Home update, focused on Gemini in the smart home: according to changelog reports from 9to5Google and Android Authority, the release makes "stop" commands more reliable to prevent misinterpretations and adds more natural ways to end a Gemini conversation, such as saying "thank you" to stop talking and close the mic. It also makes answers easier to read on smart displays, better distinguishes relaxing music from ambient sounds, and keeps weather temperature units consistent with the Home settings, while fixing unstable drag-and-drop reordering in Favorites, adding larger smoke alarm banners, and enabling Matter control for Philips Hue lights with existing bridges. Android Authority noted its author's Home app did not yet look different at publication, so the details come from Google's release notes, including the claim that compatible Matter devices respond faster and more reliably via local control through Home hubs.
What problem it solves
Coding agents need both exact design values and the rationale for applying them consistently. Without a structured, persistent representation of a design system, agents may miss design context or interpret visual guidance inconsistently.
Key features
Structured tokens and design rationale — The format combines machine-readable design tokens with human-readable Markdown ratio…
What problem it solves
Pi is designed for developers who need to adapt an agent to their own workflows rather than work within a fixed feature set. It provides interactive, automated, RPC, and SDK interfaces while leaving optional capabilities such as sub-agents and plan mode to extensions or packages.
Key features
Extensible agent harness — Customize the harness with extensions, skills, prompt te…
What problem it solves
Managing multiple AI service sessions and translating them into compatible API interfaces can require repetitive account switching, custom routing, and protocol-specific integration. The project addresses these tasks with a desktop application that combines account administration, protocol conversion, request scheduling, and model routing.
Key features
Account dashboard and …
What problem it solves
App Store and Google Play rejections can leave teams searching for the cause of a brief review message, fixing it, and resubmitting. The project addresses this with documented rejection patterns, concrete fixes, pre-submission checks, and an automated guard that can run before an upload.
Key features
Pre-submission project scanning — The guard scans iOS and Android projects …
What problem it solves
Researchers often need to turn unconstrained ideas and experimental logs into complete research manuscripts with literature synthesis, structured sections, refinements, and figures. PaperOrchestra addresses this gap by coordinating specialized agents across the paper-writing workflow and producing LaTeX manuscripts that follow target conference templates.
Key features
Multi-…
What problem it solves
Building AI chat experiences in React requires assembling chat primitives, runtime integrations, generative UI, and production UX details. The project provides a composable library that supports multiple backends, custom styling, and strongly typed APIs.
Key features
Composable Chat Primitives — Build a chat interface from primitives including Thread, Message, Composer, Thre…
What problem it solves
Autonomous agents accumulate state, require strict isolation, call model APIs and tool servers, and can continue consuming resources without supervision. AX addresses these requirements through declarative primitives for sandboxed execution, prepared workspaces, model configuration, lifecycle control, and direct inspection.
Key features
Declarative task manifests — Tasks, wo…
What problem it solves
Long-form videos need to be turned into short, platform-ready clips, but metered usage, platform branding, and server-side control can constrain the workflow. SupoClip addresses these concerns with a self-hostable, open-source pipeline that offers no-watermark output and a hosted option.
Key features
AI clip selection and scoring — A supported language model selects clip-wor…
What problem it solves
Software design often depends on context accumulated in an earlier conversation, while reviewers can miss assumptions and blind spots when they share that same context. EGM separates context-rich design work from fresh-context review so findings can be surfaced before implementation.
Key features
Context-aware design workflow — Explore requirements and trade-offs, save the a…
What problem it solves
AI coding tools often expect particular provider protocols, making it difficult to use the models and providers a user prefers. A local gateway can translate between supported client APIs and provider APIs while preserving capabilities such as streaming, tool calls, reasoning, and images.
Key features
Provider protocol translation — The gateway translates Responses, Chat Com…
What problem it solves
Server operators need to detect malicious behavior across logs and HTTP requests, then apply remediation at the appropriate layers of their infrastructure. They also need a way to benefit from threat intelligence shared by the community.
Key features
Integrated Security Engine — The Security Engine combines IDS/IPS and WAF capabilities in one solution. It analyzes log source…
What problem it solves
AI clients change between sessions, making it difficult to preserve context, recover earlier decisions, and continue unfinished work. The project provides a local Markdown-based memory that records selected information and retrieves it with its source in later sessions without automatically converting the entire conversation history into knowledge.
Key features
Persistent Ma…
What problem it solves
The hidden system instructions that shape AI product behavior are difficult to inspect and change across models, products, and versions. This repository makes collected prompts available in a structured, product-oriented archive for study and comparison.
Key features
Broad product coverage — The archive covers prompts from major AI providers and products, including Anthropic…
What problem it solves
Building AI agents and workflows often requires combining visual orchestration with custom code while connecting them to existing systems. The platform addresses these needs with model flexibility, self-hosted or cloud deployment, workflow logic, human approvals, and observability.
Key features
Visual and code-based workflow building — Build workflows on a visual canvas and …
What problem it solves
LLMs and AI agents need structured access to Google Ads account data, API metadata, and management operations. The server provides this access through MCP while adding a two-step proposal and confirmation workflow for write operations.
Key features
Query Google Ads accounts — Use GAQL to retrieve information about a Google Ads account, list directly accessible customers, and…
What problem it solves
AI agents need a reliable way to perform browser tasks such as finding availability, selecting options, handling CAPTCHAs, and completing forms. Developers also need flexibility to run the agent locally or through hosted infrastructure while choosing their model, browser mode, tools, and output format.
Key features
Persistent Python agent runtime — The model writes Python in…
What problem it solves
Coding agents often default to legacy patterns because their training data contains substantial amounts of legacy code, which can produce bloated JavaScript where native web platform solutions are available. They may also know that a modern API exists without having enough real-world implementation patterns to apply it correctly in production-ready code.
Key features
Context…
What problem it solves
Go applications need client libraries for interacting with Google Cloud Platform APIs. The project provides packages for those services and supports common authentication flows for local development and Google Cloud Platform environments.
Key features
Google Cloud API client packages — The repository contains Go packages for Google Cloud Platform services, including storage,…
What problem it solves
Coding agents need a structured way to create, manage, and build from Google Stitch designs while integrating with the Stitch MCP server. The project organizes design, code generation, and supporting workflows into skills that can be installed with supported coding agents.
Key features
Design workflows — The design plugin covers converting frontend code into Stitch designs, …
What problem it solves
Market platforms can place broad access to prices, charts, company information, watchlists, and alerts behind paid subscriptions. OpenStock addresses this by providing a community-built platform that can be hosted publicly or self-hosted with every feature.
Key features
Multi-market dashboard — Explore eight markets through a dashboard with an index pulse, sector heatmap, ma…