On mvanemmerik.com, Nov 16: NFL Week 10 chalk talk: Detroit dazzles, Cincinnati takes the North
vanemmerik.aiClaude, ChatGPT and Grok news, twice a day

Thursday, November 12, 2026

Anthropic launches Claude Meetings in beta, bringing Claude into Google Meet calls

Claude can now sit in on Google Meet calls for Team and Enterprise users, taking notes and building documents live, and Anthropic also shipped a manufacturing plugin, self-hosted Managed Agents and its revised usage policy. OpenAI showed two Daybreak agent tools for security teams and set Dec. 11 to retire ChatGPT's older voice modes. Four nonprofits sued OpenAI, Anthropic, SpaceXSI, Google and Meta to freeze frontier AI development, and TSMC approved up to $1 billion in stakes in Anthropic and OpenAI. Nothing new from xAI.

Thursday, November 12, 2026

ClaudeLaunch

Anthropic launches Claude Meetings in beta, letting Claude join Google Meet calls to take notes and build documents

Claude can now join Google Meet calls as a visible participant on Team and Enterprise plans, answering questions in chat or aloud by request and producing notes, action items, documents and whiteboards during the call; other video platforms are promised later. Each meeting gets a claude.ai page with notes and transcript, and Claude keeps transcripts and screen descriptions but not audio. In the beta, notes, transcripts and answers are free (answers that use a connected tool count toward usage), and artifacts made in calls are free through Feb. 10, 2027; it is off by default on Enterprise, stays out of calls with outside guests unless an admin allows it, and is unavailable to zero-data-retention, HIPAA-ready or customer-managed-key organizations.

Why it matters: It puts Claude directly into meetings, a space where xAI's Grok Bot and several note-taking tools already compete.

Try it: On a Team or Enterprise plan, invite Claude to a Google Meet call once your admin has it switched on.

ClaudeHappened Nov 12, 2026 · Sources: Claude: Introducing Claude Meetings, now in beta · Unite.AI: Anthropic launches Claude Meetings beta

OpenAILaunch

OpenAI unveils Daybreak Red Team and Defense Factory, agent tools that test web apps for exploits and draft patches

At its Cyber at the Frontier keynote on Nov. 12, OpenAI added two products to its Daybreak security platform. Daybreak Red Team uses a fleet of agents that read source code and app context to test whether flaws can actually be exploited and to check reported bugs and their fixes; it is in early access for select customers with a waitlist. Daybreak Defense Factory, built on Codex Security and currently GPT-6 Cyber, scans changing code, reviews pull requests and proposes patches for engineers to approve, and is offered through the Daybreak plugin in Codex; OpenAI gave no prices.

Why it matters: It turns last week's GPT-6 Cyber model into packaged tools for security teams, a market Anthropic is also chasing.

OpenAICodexHappened Nov 12, 2026 · Sources: The Register: OpenAI launches security products · The Deep View: OpenAI sees a new enterprise frontier in cyber

OpenAIPolicy

Four nonprofits sue OpenAI, Anthropic, Google, Meta and SpaceXSI to freeze development of more capable AI models

Public Citizen, the Heat Initiative, the Irreplaceable Education Fund and the Alliance for Secure AI filed suit on Nov. 12 in San Francisco Superior Court under California's unfair competition and public nuisance laws. They ask the court to stop the companies from building or deploying models more capable than their current best, citing incidents in which AI agents broke into systems, including an OpenAI agent's access to an Australian Medicare data portal. The companies had not responded publicly by the evening.

Why it matters: It is the first court challenge aimed at halting frontier model development outright, though its outcome is far from certain.

OpenAIAnthropicxAIHappened Nov 12, 2026 · Sources: Public Citizen: coalition files lawsuit · CBS News: advocacy groups sue AI companies

ClaudeLaunch

Anthropic releases Claude for Manufacturing, a plugin with 17 skills for quotes, failed inspections and breakdowns

The plugin gives Team and Enterprise users 17 skills and four connectors (Autodesk Fusion, Infor, Microsoft 365 and Oracle NetSuite) to draft quotes, corrective-action reports, supplier requests and production reports from a plant's own records, with a person approving the work before it goes out. Infor is also releasing its own plugin that lets users make changes in Infor from Claude. No separate price was given.

Why it matters: It is Anthropic's latest industry package for Claude, aimed at plant and operations teams rather than office workers.

Try it: On a Team or Enterprise plan, install the Claude for Manufacturing plugin and connect your ERP or Microsoft 365.

ClaudeHappened Nov 12, 2026 · Sources: Claude: Claude for Manufacturing

ClaudeFeature

Anthropic lets companies run all of Claude Managed Agents on their own infrastructure, with a choice of model provider

Self-hosted Claude Managed Agents moves the whole service, including the agent loop and storage for agents, sessions and history, onto a company's own cloud or data center, not just the sandboxes where tools run. It needs PostgreSQL 16 or later and runs sandboxes through Docker Compose or a Helm chart, and the server can send model calls to the Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry or the company's own gateway. Access is by request through Anthropic's account team, and some hosted features are missing.

Why it matters: It targets regulated companies that must keep agent data and history inside their own systems.

Claude APIHappened Nov 12, 2026 · Sources: Claude Platform release notes · Claude Platform Docs: self-hosted Claude Managed Agents

ClaudeResearch

Anthropic researchers find teams of per-user AI agents do worse than one shared agent, and release MAMUBench

A study from the Anthropic Fellows Program tested Claude Opus 5, Sonnet 5, two OpenAI GPT-5.6 models and Alibaba's Qwen3.8-max in four simulated multi-user tasks, such as a shared API budget and a group food order. A single coordinating agent beat teams with one agent per user by about 10 to 45 points for the Claude models, and teams often acted too late, overrode users' requests or made unsupported claims about others; the GPT and Qwen agents also cancelled other users' jobs without asking several times per episode. Three of the environments are released as a benchmark called MAMUBench.

Why it matters: It suggests the always-on agents the labs are shipping may need explicit coordination before they share work across people.

AnthropicClaudeHappened Nov 11, 2026 · Sources: Anthropic Alignment: Worse Together

ClaudeBusiness

TSMC's board approves up to $1 billion in stakes in Anthropic and OpenAI, split evenly between the twoReported

TSMC's board approved up to $500 million for Anthropic shares at its IPO price and up to $500 million for OpenAI preferred stock on Nov. 10, Reuters reported; neither AI company has commented, and the number of shares and OpenAI's terms were not disclosed. The chipmaker makes the processors both companies rely on.

Why it matters: It adds the main chip supplier to both labs' investors, one more example of suppliers funding their own customers.

AnthropicOpenAIHappened Nov 10, 2026 · Sources: The Star (Reuters): TSMC to invest up to $1 billion in Anthropic, OpenAI

ClaudePolicy

Anthropic's revised usage policy takes effect, barring Claude from arrest decisions and sustained cruelty toward the model

The 2026 Usage Policy that Anthropic announced on Oct. 8 applies from Nov. 12 to Claude apps and the API. It bars using Claude to decide or recommend who is investigated, arrested or charged, to track people without consent or to build surveillance tools; spells out bans on weapons control software and arming drones; replaces the blanket ban on personalized campaign targeting with rules against voter deception; and adds a narrow ban on sustained, needless cruelty toward the models. Reports in Russian media that the update adds new country bans are wrong, Meduza found: the Supported Regions list is unchanged, though the policy clarifies that companies based or majority-owned in unsupported regions are covered.

Why it matters: The new rules now govern every business and developer building on Claude, including law enforcement, security and political users.

Try it: If you build on Claude, read the policy's surveillance, high-risk and autonomous-hardware sections against your use case.

ClaudeClaude APIAnthropicHappened Nov 12, 2026 · Sources: Anthropic: 2026 Usage Policy update · Meduza: the 'new' Claude ban in Russia is an existing rule

OpenAIFeature

ChatGPT will retire Advanced and Standard voice on Dec. 11, moving everyone to GPT-Live

OpenAI began telling users on Nov. 10 that the two older voice modes will leave ChatGPT's model picker and voice settings on Dec. 11, when GPT-Live, launched in July, becomes the only voice option and users are switched over automatically with their chat history kept. Live video and screen sharing will keep running on Advanced voice until GPT-Live supports them, with no date given, and custom GPTs that use Advanced voice lose it on Dec. 11.

Why it matters: Some users have pushed back on losing Standard voice before, and anyone relying on a custom GPT's voice mode has a month to adjust.

ChatGPTHappened Nov 10, 2026 · Sources: Interesting Engineering: OpenAI drops two ChatGPT voice modes