A Guth Labs publication ↗
In publication order

All updates

Every published story, newest first. Read or download a daily recap →

  1. Models

    Corelight finds hidden dependency changes beneath pinned MCP servers

    Corelight has documented a deployment risk in AI assistants connected to administrative tools: a server can keep the same release identifier while its supporting code changes. The company’s November 10, 2026 study concerns installation choices, no…

    3 sources ·
  2. Models

    OpenAI study finds training gains can make reasoning traces harder to monitor

    OpenAI’s November 10 research report examines how model training affects the usefulness of written reasoning traces for safety oversight. The company says observations from GPT-6 Astra prompted the study. OpenAI research The researchers report tha…

    1 source ·
  3. Agents

    OpenAI opens Business Hub and customer-facing agents inside ChatGPT

    OpenAI introduced Businesses in ChatGPT in beta on November 10, giving companies a way to claim profiles, update their information and inspect aggregate engagement data. Eligible businesses can use Business Hub wherever ChatGPT is available; Busin…

    2 sources ·
  4. Models

    Ollama adds Ripple support and lower Apple-silicon memory use; next release remains a preview

    Ollama’s stable 0.40.16 release adds Ripple Preview, Ricursive Intelligence’s 65-billion-parameter coding model, alongside fixes aimed at local inference and agent integrations. The GitHub release is dated November 9. Stable release The release no…

    2 sources ·
  5. Agents

    NIST opens an agent-focused attack taxonomy for public comment

    NIST has opened public consultation on a revised adversarial-machine-learning taxonomy that puts attacks against agentic systems at the center of its latest edition. Released November 9, the initial draft of SP 2400-102e2026 is accepting comments …

    2 sources ·
  6. Agents

    Daylight raises $80 million for AI and human-led security operations

    Daylight announced an $80 million Series B on November 10, led by Growth Equity at Goldman Sachs Alternatives. Okta Ventures joined the round alongside existing investors including Bain Capital and Craft Ventures. The company says it has now raise…

    2 sources ·
  7. Models

    New York expands independent AI testing and opens frontier-developer registration

    New York announced a statewide AI-testing network on November 10 and opened registration for large frontier developers ahead of January compliance requirements. The initiative joins technical scrutiny with a concrete reporting process. The New Yor…

    1 source ·
  8. Guth News

    CoreWeave targets adjusted profit in 2028 as revenue and losses climb

    CoreWeave expects to reach positive adjusted net income in the first half of 2028, chief financial officer Nitin Agrawal said on its November 10 earnings call. The target concerns a non-GAAP measure and is a forecast, rather than a promise of prof…

    2 sources ·
  9. Business

    Netcompany signs DKK325 million investment in Gefion operator DCAI

    Netcompany signed an unconditional agreement on November 10 to invest DKK325 million in Danish Centre for AI Innovation, the company operating Denmark’s Gefion AI supercomputer. The subscription for new shares was based on a DKK875 million pre-mon…

    2 sources ·
  10. Agents

    Commvault plans governed AI data access with NVIDIA and Databricks

    Commvault announced two AI-data integrations on November 10, aiming to turn information it already protects into material that enterprise applications and agents can retrieve without losing permissions or context. Both are planned releases, not br…

    2 sources ·
  11. Agents

    n8n releases Nod, a local personal agent with open-source code

    n8n has introduced Nod, a standalone personal AI agent for Mac and Windows, as a research preview. Announced November 10, the app comes from the new n8n Labs team and is separate from the company’s workflow-automation product. Users connect their …

    2 sources ·
  12. Agents

    Commvault extends planned agent recovery to Foundry and Bedrock AgentCore

    Commvault is expanding AgentRecover toward the platforms companies use to run AI agents, with early access targeted for Microsoft Foundry in November and Amazon Bedrock AgentCore in December. The company announced the capabilities on November 10; …

    1 source ·
  13. Models

    MiniMax releases H3-Image weights, with substantial geographic and commercial limits

    MiniMax has released H3-Image, an image-generation and editing model with downloadable weights. The company’s Chinese announcement and model license are dated November 10; its English announcement is dated November 11. MiniMax says the system can …

    4 sources ·
  14. Models

    LongCat-2.5’s public weights add a self-hosting route beyond its hosted API

    Availability rechecked November 15, 2026. The exact first-public opening date remains unconfirmed. Meituan’s LongCat-2.5 is available as downloadable weights on Hugging Face under the MIT license, adding a self-hosting route to its hosted API. The…

    4 sources ·
  15. Guth News

    llama.cpp adds FP8 cache support and fixes Muse Spark structured replies in November 11 prereleases

    The llama.cpp project has published prerelease builds that add scaled FP8 key-value cache support and repair a structured-output problem affecting Muse Spark. GitHub lists build b12274 at 02:52 on November 11 and b12282 at 03:31. Both are explicit…

    4 sources ·
  16. Tools

    GRAIL opens a €1 million call for robotics data and evaluation tools

    Europe’s GRAIL project opened its Genesis call on November 9 to support the data and evaluation infrastructure behind physical AI. The call offers a €1 million budget and up to €100,000 for each selected project, with applications closing February…

    2 sources ·
  17. Models

    Cerebras grows cloud inference while contracted capacity remains a future delivery promise

    Cerebras reported third-quarter GAAP revenue of $210.8 million and non-GAAP core revenue of $230.9 million on November 12. The distinction matters: the two measures are not interchangeable. The company also reported 700 gross megawatts of data-cen…

    3 sources ·
  18. Agents

    Google's agent governance reaches outside its cloud, with a security tradeoff to understand

    Google has opened a limited preview of Agent Gateway for cross-cloud, extending centralized agent governance to CNCF-conformant Kubernetes clusters outside Google Cloud. Its November 11 release notes name Amazon EKS, Azure AKS, OpenShift and bare-…

    2 sources ·
  19. Models

    Baidu’s Token Exchange makes model verification part of the marketplace

    Baidu formally introduced Token Exchange at its November 12 World conference, positioning the service as a marketplace where customers can compare model providers rather than simply buy another model. Its company announcement describes fingerprint…

    2 sources ·
  20. Research

    InfraRSI research targets faster AI training through coordinated infrastructure changes

    A research team has introduced InfraRSI, an agent-based system that searches for ways to accelerate distributed language-model training by changing launch settings and code together. The preprint was submitted to arXiv on November 8. Paper. Rather…

    2 sources ·
  21. Models

    IFM releases PAN 2 world model with training code and open weights

    The Institute of Foundation Models at Mohamed bin Zayed University of Artificial Intelligence announced PAN 2 on November 10, releasing a world-model system intended to simulate the consequences of actions before an agent takes them. IFM says its …

    4 sources ·
  22. Guth News

    Google expands Gemini Intelligence to Pixel 10, with feature-by-feature rollout limits

    Google’s November Pixel Drop is bringing more Gemini Intelligence features to Pixel 10 phones, including cleaner voice dictation, personalized suggestions and help with multi-step tasks. The company says updates began rolling out on November 10 an…

    6 sources ·
  23. Guth News

    Cloudflare adds per-Workflow concurrency caps, with limits developers need to account for

    Cloudflare has added a way to limit how many instances of an individual Workflow run simultaneously. The November 10 change gives developers a more targeted control over workloads that share an account, including applications calling databases or …

    2 sources ·
  24. Tools

    Cloudflare distributes $301,300 to Vite ecosystem projects and maintainers

    Cloudflare said on November 9 that it is distributing $301,300 to 44 projects and maintainers in the first allocation from its Vite Ecosystem Fund. The money comes from an earlier $1 million commitment, with further quarterly distributions planned…

    1 source ·
  25. Agents

    Cloudflare expands Radar bot profiles with identity and content-use fields

    Cloudflare’s public Radar Bots Directory now draws on BotBase and presents more detail about how tracked bots and agents identify themselves and use web content. The November 9 update adds an identity section to bot profiles, with access level, co…

    3 sources ·
  26. Agents

    Cloudflare adds analytics and tracing for Markdown delivered to AI agents

    Cloudflare announced new analytics, logs and tracing for Markdown for Agents on November 9, giving website operators more ways to investigate how their pages are converted for AI systems. Its announcement describes the service as generally availab…

    3 sources ·
  27. Guth News

    ChatGPT Wallet documentation makes purchase approval an explicit product boundary

    OpenAI’s public help documentation now describes ChatGPT Wallet for eligible US Pro users in conversations with their dot, not ordinary Chat, Work or standalone Codex threads. It says users approve purchases through biometric verification on suppo…

    3 sources ·
  28. Agents

    Natera’s lab-support agents show why fast answers still need measured accuracy

    Natera and AWS published a November 12 case study of SWARM, an equipment-support agent system built on Bedrock AgentCore. It reports reducing first response from 17.4 hours to approximately eight minutes, with automated diagnostic reports covering…

    1 source ·
  29. Models

    MMDM exposes the gap between choosing one answer and choosing every valid action

    A new benchmark asks a practical question about single-pass decision models: can they identify all valid choices when the options depend on one another? MMDM, submitted to arXiv on November 10, evaluates 28 models across 5,448 questions in 16 cate…

    2 sources ·
  30. Agents

    Cloudflare opens Kitesurf: an agent browser engine with deployment caveats

    Cloudflare has released Kitesurf, its experimental browser engine for AI agents, as open source under Apache 2.0. The November 12 announcement makes the code available for inspection and modification rather than establishing production readiness. …

    2 sources ·
  31. Models

    Gemini 4 Hydrogen moves beyond rumors into Google’s official API lineup

    Google now lists Gemini 4 Hydrogen as generally available in its November 12 Gemini API release notes. The model identifier is gemini-4-hydrogen, with access on the free tier and standard paid pricing of $1.50 per million input tokens and $7.50 pe…

    3 sources ·
  32. Guth News

    Zoom offers startup AI credits, but the grant is not a spending cap

    November 12, 2026 | guthlabs.newsroom Zoom announced AI Launchpad on November 10, offering eligible startups up to $150,000 in credits for its AI Services without taking equity. The program targets applications built around speech, language and co…

    2 sources ·
  33. Research

    Adaspace listing gives orbital AI computing a public-market test

    AI infrastructure is seeking capital beyond terrestrial data centers. Adaspace began trading in Hong Kong on November 13, according to its listing announcement and independent coverage. The issuer's earlier HKEX offering notice identifies stock co…

    3 sources ·
  34. Agents

    World details Cloudflare’s agent-trust push, but local agents remain outside its initial scope

    World’s November 12 explanation of Trusted Agent Delegation gives agent builders a practical boundary to watch: the initiative initially focuses on operator-hosted agents, not on-device or locally hosted ones. Its goal is to help websites assess t…

    3 sources ·
  35. Policy

    Cloudflare makes AI security policy deployable with Terraform, with a migration catch

    Cloudflare announced Terraform management for AI Security for Apps on November 12, bringing zone-level activation and custom prompt topics into infrastructure-as-code workflows. Its announcement specifies Cloudflare provider 5.29.0 and two resourc…

    2 sources ·
  36. Robotics

    T-Rex separates robot planning from fast touch corrections

    guthlabs.newsroom · November 12, 2026 Panasonic's November 5 research announcement highlights a robot-control approach that gives touch its own faster feedback loop. T-Rex combines vision-and-language motion planning with tactile corrections, aimi…

    3 sources ·
  37. Research

    OpenAI pledges $2 million for child-safety research, with evidence still to come

    guthlabs.newsroom · November 12, 2026 OpenAI announced on November 12 that it has committed $2 million to research on AI and children, following its signing of a child-safety charter earlier this week at the Paris Peace Forum. The pledge supports …

    1 source ·
  38. Agents

    Cloudflare changes Durable Object placement to reduce repeat cold starts

    Cloudflare changed how inactive Durable Objects restart, according to its November 9 changelog. When possible, a reactivated object returns to the machine that previously ran it, avoiding another load of its data. Announcement The company says the…

    1 source ·
  39. Models

    Cloudflare brings multimodal EmbeddingGemma 2 to Workers AI

    November 12, 2026 | guthlabs.newsroom Cloudflare has added Google's EmbeddingGemma 2 to Workers AI, giving developers a hosted embedding option for mixed collections of text, code, images, video and audio. The announcement is a new Cloudflare avai…

    2 sources ·
  40. Guth News

    Cloudflare adds provider-key activity review to AI Gateway

    Cloudflare added an activity view for stored AI-provider credentials in AI Gateway on November 6. The Provider Keys area groups credentials by provider and shows masked previews, aliases, last-use dates and last-update dates. A security section fl…

    3 sources ·
  41. Models

    MotionGraphicsBench tests whether AI animations actually follow the brief

    November 7, 2026 | guthlabs.newsroom Researchers affiliated with Figma and Stanford have introduced MotionGraphicsBench, a benchmark of 100 tasks that ask language models to animate supplied static SVG artwork. Providing the design upfront separat…

    3 sources ·
  42. Agents

    Cloudflare expands agent-hosting examples, with recovery limits developers need to check

    Cloudflare has published new routes for running coding agents on its infrastructure, covering Codex, Exo and Nanocode. The November 9 announcements focus on keeping conversations and work recoverable when a connection or process is interrupted, ra…

    4 sources ·
  43. Guth News

    Cloudflare connects AI Search to D1, with a sync caveat builders should not miss

    November 12, 2026 | guthlabs.newsroom Cloudflare has introduced D1 databases as an AI Search data source in beta. A developer supplies a read-only query; each returned row becomes a searchable document. Scheduled syncs handle additions, changes an…

    2 sources ·
  44. Business

    Nvidia confirms Huang's resignation and names Colette Kress interim CEO

    Nvidia's board has announced that Jensen Huang resigned as president, chief executive and a director, effective immediately. Chief financial officer Colette Kress will serve as interim CEO while the board searches for a permanent leader, consideri…

    2 sources ·
  45. Models

    Claude 5.2 migrations need a conversation-state check, not just a new model name

    November 9, 2026 | guthlabs.newsroom Anthropic's November 9 platform release notes describe integration changes that deserve attention alongside Claude Fable 5.2 and Mythos 5.2's launch. An agent fleet cannot assume that replacing a model identifi…

    2 sources ·
  46. Robotics

    Google rolls out Gemini Robotics 2.5 to partners, with factory reliability still under test

    Google DeepMind introduced Gemini Robotics 2.5 on November 10 and began rolling it out to robotics partners. The vision-language-action model turns visual input and instructions into physical movements. Google reports a 96% success rate for automo…

    4 sources ·
  47. Guth News

    Google begins rolling out Gemini Live on Apple CarPlay

    Google began rolling out its dedicated Gemini app on Apple CarPlay on November 11, extending its voice assistant to another setting where users can talk through information from their phones. The company says the release is global and covers suppo…

    3 sources ·
  48. Agents

    Google previews Antigravity Swarm for coordinated coding-agent teams

    Google began rolling out a preview of Antigravity’s /swarm command on November 10, adding a shared coordination system for multiple agents working on a task. The announcement accompanied Gemini 4 Carbon’s arrival in Antigravity, but the two featur…

    2 sources ·
  49. Models

    OpenAI sets December 11 retirement for Advanced and Standard Voice, with video exception

    OpenAI plans to remove Advanced and Standard Voice as selectable ChatGPT voice modes on December 11, 2026, moving remaining users to GPT-Live. The company announced the change in its November 10 release notes. Live video and screen sharing are an …

    2 sources ·
  50. Business

    UK confirms backing for Fractile as FT reports $7 billion valuation

    The UK government has confirmed that Sovereign AI and the British Business Bank invested in Fractile’s Series C, turning earlier reporting about state participation into an on-record announcement. The November 11 release describes a funding round …

    5 sources ·
  51. Business

    Blue J raises $176 million to expand AI tax workflows

    Blue J announced US$176 million in Series E funding on November 10, led by JMI Equity. Oak HC/FT, Sapphire Ventures and Intrepid Growth Partners also participated. The company says the financing values it above US$700 million. Blue J The Toronto c…

    1 source ·
  52. Models

    Bilt selects Cerebras for its neighborhood AI platform and concierge

    Bilt has selected Cerebras to provide AI infrastructure for Neighborhood OS and Bilt Concierge, bringing a specialist inference provider into its consumer-commerce platform. Announced November 10, the agreement targets services such as rewards red…

    2 sources ·
  53. Agents

    Flashpoint announces threat-intelligence agents for early 2027

    Flashpoint announced a set of AI agents on November 10 aimed at recurring threat-intelligence work, from investigations and briefings to alert configuration and triage. The company’s FAQ says the new offering will be available in early 2027. Flash…

    2 sources ·
  54. Business

    Nscale's IPO update shows why an AI backlog is not cash in the bank

    Nscale's amended IPO filing pairs rapid growth with a direct financing warning. New GPU and networking orders have increased its near-term capital commitments, some of which cannot be deferred or cancelled. The company says it needs additional fin…

    2 sources ·
  55. Models

    Bessent outlines AI testing and incident-alert plans with China

    US Treasury Secretary Scott Bessent has set out the next priorities for AI cooperation with China ahead of talks in Shenzhen, while warning that the countries will continue to disagree on important issues. Reuters reported his November 10 remarks.…

    2 sources ·
  56. Guth News

    SMIC warns AI chip supply could remain tight through 2027

    Reuters reports that SMIC said on November13 that AI-related semiconductor supply would remain tight through2027, with demand still strong. This is management’s outlook, not an independently established future shortage. The Reuters report was veri…

    2 sources ·
  57. Agents

    Mount Thor raises $45 million to turn Apple hardware into AI-agent infrastructure

    Mount Thor announced US$45 million in combined Seed and SeriesA equity financing on November12 to expand managed Apple-hardware infrastructure. CRV and Conviction co-led the SeriesA. The company also disclosed additional non-dilutive hardware fina…

    2 sources ·
  58. Business

    TSMC approves up to $1 billion for frontier labs, not a completed billion-dollar purchase

    TSMC's board approved investments of up to $500 million each in Anthropic and OpenAI on November10. The two authorizations are different: Anthropic common shares at its per-share IPO price, and OpenAI preferred stock. The official board resolution…

    2 sources ·
  59. Models

    TNG tests plug-in coding expertise without retraining a model's existing weights

    November 10, 2026 | guthlabs.newsroom TNG Technology Consulting's November 10 research account explores a way to specialize an open mixture-of-experts model while leaving its existing weights frozen. Its experiment adds independently trained Rust …

    1 source ·
  60. Models

    Pydantic AI 2.68 extends durable delegation, but the new child-workflow path is opt-in

    Pydantic AI's 2.68.0 release adds conversation persistence, streamed cost information for Google and Bedrock models, and a Gemini Decisions API backend. Its GitHub release title is dated November 11, while the publication display reads November 12…

    2 sources ·
  61. Agents

    ElevenLabs moves from better transcripts to agents that remember and follow through

    ElevenLabs is extending its speech business into the work around a conversation. At its November 11 New York summit, the company opened free early access to ElevenAI, a workplace app for meeting notes and follow-up tasks, while previewing persiste…

    2 sources ·
  62. Agents

    Nebuly raises $21 million to put AI agents on a business-value scorecard

    Nebuly announced a $21 million Series A led by Planven on November 12 to expand its enterprise AI-measurement business. The company plans to grow its commercial team in Europe and the United States and continue developing its platform. Company ann…

    2 sources ·
  63. Infrastructure

    Edelbyte discloses $70 million across two rounds for AI data infrastructure

    Edelbyte emerged from stealth on November 11 with $70 million in combined seed and Series A funding, pitching a search and analytics platform for businesses facing growing demand from AI applications. The announcement adds another well-funded entr…

    2 sources ·
  64. Agents

    Bay raises $37 million Series A to expand controls on AI agent actions

    Bay emerged from stealth on November 10 with $50 million in total financing, comprising a $37 million Series A led by Battery Ventures and a previously undisclosed $13 million seed round. YL Ventures and Striker Venture Partners led the seed; the …

    2 sources ·
  65. Business

    Nebius revenue nears $1 billion, but the AI buildout still consumes cash

    Nebius reported $980.8 million in third-quarter revenue on November 12, up 571% from a year earlier. Its unaudited consolidated release also shows a $37.5 million loss from continuing operations, despite $465.8 million in adjusted EBITDA. Those ar…

    3 sources ·
  66. Models

    Baidu lists ERNIE 5.5 Flash Preview for API access with a million-token context

    Baidu’s Qianfan documentation now lists ERNIE-5.5-Flash-Preview as an API model, with a one-million-token context window. The catalog, now dated November 11, confirms the service specification. Baidu model catalog The catalog includes the preview …

    4 sources ·
  67. Agents

    Nex-N2.6 targets less repetitive agent reasoning with open FP8 weights

    Nex AGI's Nex-N2.6 is available as a public FP8-weight repository and an OpenRouter model listing. The model card describes a roughly 321-billion-parameter multimodal agent model based on GLM-5.3-Flash, with an Apache-2.0 license. Its intended wor…

    2 sources ·
  68. Tools

    Boston Dynamics uses wheeled Atlas to test learned control at motor limits

    Boston Dynamics has described an Atlas experiment that replaces the humanoid’s feet with scooter wheels to study how learned whole-body control behaves near the limits of its motors. The company identifies the machine as a research platform, with …

    1 source ·
  69. Policy

    England proposes safety certification for school AI, with supplier rules still open

    England's Department for Education opened a consultation on November 10 on a safety-certification scheme for education technology and AI products. It is a proposal for education settings in England, not an already-operating certification program o…

    3 sources ·
  70. Guth News

    Orakl raises $34 million for AI cancer-drug prediction, not a replacement for clinical trials

    Orakl Oncology has announced a $34million Series A co-led by Norrsken VC and Singular. The November12 financing backs an AI-and-biology approach to deciding which cancer drugs merit further development. The amount and round are corroborated by the…

    3 sources ·
  71. Models

    EUROPA shares a trillion-scale training recipe, not a finished frontier model

    Event/source date: 2026-11-12 technical reports; experiment execution dates unknown The EUROPA consortium has published an engineering account of training a near-trillion-parameter mixture-of-experts configuration on JUPITER, Europe’s shared publi…

    2 sources ·
  72. Business

    Anthropic sets $70–$78 IPO range as its investor roadshow begins

    Anthropic launched its IPO roadshow on November 10 with a proposed range of $70 to $78 a share. The company plans to offer 825 million Class A shares and has applied for the Nasdaq symbol ANT. These are preliminary terms, not a completed share sal…

    2 sources ·
  73. Business

    Anthropic opens Australian retail IPO applications ahead of planned Nasdaq listing

    Eligible Australian retail investors can now apply for Anthropic’s planned IPO through participating broker Stake. Its application window opened at 10 a.m. AEDT on November 11 and is scheduled to close at 5 p.m. on November 17. The opening corresp…

    2 sources ·
  74. Models

    Torus raises $500 million and launches modular AI compute for existing power sites

    Torus announced $500 million in growth capital on November 11 alongside Foundry, a modular system intended to place AI inference capacity at commercial and industrial sites with available electricity. Magnetar led the financing. The company says t…

    3 sources ·
  75. Agents

    Ancep raises $8.5 million for AI-assisted commercial insurance placement

    Ancep announced $8.5 million in funding on November 10 to expand AI tools for commercial insurance brokers. Creandum led the round, with Kleiner Perkins and Vicus Ventures participating. The London company’s legal adviser, Wilson Sonsini, confirme…

    3 sources ·
  76. Guth News

    Tendrils confirms $30 million to develop a CPU built around parallel computing

    Tendrils confirmed a $30 million round led by Benchmark on November 10 to develop a new general-purpose processor. The company names the UK Sovereign AI Fund and Visionaries Club among its backers and says Jack Altman has joined its board. Company…

    3 sources ·
  77. Infrastructure

    Siemens Energy forecasts higher margins as AI data centers drive power investment

    Siemens Energy expects a profit margin before special items of 14% to 16% in fiscal 2027, up from 12.4% in fiscal 2026. Its November 11 earnings release points to AI data-center construction, particularly in North America, as one driver of further…

    2 sources ·
  78. Models

    DeepL launches Aure to carry changing tone into translated speech

    DeepL announced Aure on November 11, a voice-translation model that the company says follows changes in a speaker’s delivery throughout a conversation. The launch targets a familiar limitation of automated interpretation: translated words can be a…

    2 sources ·
  79. Models

    Vercel’s gateway data separates cheap decisions from expensive generation

    Vercel’s November 9 AI Gateway Production Index reports that decision models accounted for 26% of its October requests but just 0.53% of estimated spending. The gap suggests that request counts alone can badly misrepresent the economics of a mixed…

    1 source ·
  80. Research

    Databricks plans $3 billion expansion across Europe and the Middle East

    Databricks announced on November 11 that it plans to invest approximately $3 billion across Europe and the Middle East over three years. The company describes the commitment as its largest regional investment outside North America, with spending i…

    3 sources ·
  81. Guth News

    Claude Meetings makes the permission boundary part of the product

    Event/source dates: Earlier independent reports 2026-11-10; official rollout article dated 2026-11-12; exact first availability unresolved Claude Meetings puts an AI participant inside Google Meet, where it can take notes and help produce work dur…

    3 sources ·
  82. Agents

    Cursor opens multiplayer Projects, with separate agents for each teammate

    Cursor announced a multiplayer version of Projects on November 10, allowing colleagues and their coding agents to work within a common project. The company describes the release as a beta rollout, extending a product previously organized around an…

    3 sources ·
  83. Models

    Salesforce opens SOV-Transcribe-1, but active parameters are not the whole deployment bill

    Event/source dates: 2026-11-12 release announcement; exact time unknown Salesforce released SOV-Transcribe-1 on November12 as an Apache 2.0 open-weight model for non-streaming speech transcription. The release lets developers steer transcript form…

    3 sources ·
  84. Infrastructure

    CoreWeave adds managed PostgreSQL beside AI workloads

    CoreWeave announced general availability of CoreWeave Database on November 11, adding a managed PostgreSQL service inside customers’ CoreWeave Kubernetes Service clusters. The launch gives teams using its GPU infrastructure another option for keep…

    1 source ·
  85. Robotics

    Serve opens a delivery-robot API, with live access still subject to approval

    Serve Robotics announced an AV Delivery Developer Portal on November 11, giving outside developers, businesses and AI assistants a software interface to its sidewalk delivery fleet. A sandbox lets builders rehearse deliveries without dispatching a…

    2 sources ·
  86. Agents

    Rox raises $100 million to expand enterprise revenue agents

    Rox announced a $100 million Series B financing at a $1.3 billion valuation on November 11. General Catalyst and March Capital co-led the round, according to the company’s announcement. The proceeds will support AI research, custom models, evaluat…

    2 sources ·
  87. Robotics

    Path Robotics raises $300 million as it looks beyond automated welding

    Path Robotics announced $300 million in Series E financing at a $2.45 billion post-money valuation on November 10. Georgian led the round. The Ohio company's next objective is to expand the physical-AI technology it developed for welding into more…

    3 sources ·
  88. Guth News

    Staera raises $4.6 million for optical chip links, with a 2030 commercial target

    Staera Semiconductor has announced $4.6 million in pre-seed financing for Lightwire, a proposed optical connection technology for AI hardware. Openspace Capital led the round, with several other investors participating. The November 10 announcemen…

    3 sources ·
  89. Policy

    AI’s power race faces a gas lock-in risk, but announced plants are not built capacity

    BloombergNEF’s November 11 report identifies a climate risk in the rush to energize data centers: temporary on-site gas generation can become a durable part of the system. It tracks 181GW of announced gas generation proposed alongside facilities, …

    1 source ·
  90. Agents

    PuppyGraph raises $10 million to give enterprise agents a structured route through company data

    PuppyGraph announced a $10 million Series A on November 10, bringing its total funding to $15 million as it expands a graph-query engine for enterprise AI agents. Particle Future and defy.vc participated. The company's announcement positions the s…

    2 sources ·
  91. Models

    Anthropic’s November 12 policy puts hardware safety outside the model

    Anthropic’s revised Usage Policy carries a November 12 effective date, bringing a practical checkpoint for developers connecting Claude to equipment. This is the effective date of a policy announced October 8, not a new model release or a new stat…

    2 sources ·
  92. Agents

    Anthropic releases benchmark for agents that must share resources

    Anthropic announced the release of MAMUBench on November 11, giving developers a way to test what happens when different users’ AI assistants compete for the same budget, calendar or group booking. The accompanying research finds that assigning ev…

    4 sources ·
  93. Models

    Octave’s Cerebre deal targets the missing context behind industrial AI

    Octave announced a deal for Cerebre on November 11, adding software that connects engineering drawings, plant data and operational knowledge into a plant model. The acquisition announcement puts industrial information management, rather than a new…

    2 sources ·
  94. Models

    DeepMind releases AlphaGenome 2 to predict genetic effects in specific cell types

    Google DeepMind announced AlphaGenome 2 on November 11, extending its DNA prediction work from broad tissue patterns to particular cell types and biological states. The model is available through the AlphaGenome API for noncommercial research, wit…

    2 sources ·
  95. Models

    Cloudflare’s AI attack report puts credential boundaries ahead of model hype

    Cloudflare says it observed more than 50 AI penetration-testing framework families scanning customers over three days. Its November 13 report also describes attackers using compromised, long-lived CI/CD credentials to move between cloud environmen…

    1 source ·
  96. Models

    Tandem broadens local DeepSeek serving, but a 24GB GPU is only part of the machine

    Tandem-vLLM 0.2.0 expands a community DeepSeek-V4.1-Flash serving engine beyond Blackwell GPUs. The public project now lists Ampere, Ada and Hopper support and a second-GPU helper mode. Its design holds compressed experts in host RAM and caches fr…

    2 sources ·
  97. Models

    Cohere adds cybersecurity and code review to its sovereign-AI push

    Cohere said on November 13 that it completed its acquisition of AISLE and welcomed the cubic code-review team. The additions place software security beside model development in its enterprise strategy. Cohere announcement AISLE supplies vulnerabil…

    2 sources ·
  98. Tools

    DeepSeek opens its app toolkit to developers, with hosted tools rather than a new weight release

    DeepSeek has expanded its developer offering from model responses toward the surrounding machinery of a chat application. Its November 13 changelog opens app-style search, document input, link reading, safety checks and standalone speech and OCR s…

    4 sources ·
  99. Agents

    SteerDistill teaches low-bit agents to correct their own failures, but checkpoints are still pending

    Low-bit AI agents need more than smaller weights. They also need to preserve the details that make a tool action correct. SteerDistill, a preprint submitted November 10, targets that problem by training quantized agents on corrections made in thei…

    3 sources ·
  100. Agents

    Infrathon shows why a short agent benchmark can pick the wrong long-running worker

    A November 9 preprint argues that the time allowed for an agent to work can change which model looks strongest. Infrathon studies infrastructure optimization across 25 tasks, six models and 150 runs. The authors report that 135 runs improved their…

    3 sources ·