Scoop
AI tool news · rumor vs. reality

The AI Wire ●

Rumors tracked. Announcements verified. Updated daily.

2026-11-12

ANNOUNCED

Notion says AI chat transcripts from fewer than 1% of workspaces were exposed in the Fireworks AI breach

Notion says AI chat transcripts from fewer than 1% of its workspaces were exposed in the breach at Fireworks AI, the inference provider it used to serve open-weight models in Notion AI and to process data for its AI usage Insights feature. Notion's own systems were not breached; the data was exposed one layer down, at a supplier most Notion customers had never heard of.

Fireworks alerted Notion on November 2 to an exposed dataset that at first appeared to contain only synthetic test data, and Notion disconnected from Fireworks and deleted its stored data there on November 3. On November 8 Fireworks confirmed the dataset held real Notion customer data: chat transcripts collected for Insights, including inputs and model outputs written to the chat. Notion says that for the vast majority of affected workspaces the transcripts were not attributable to a user or workspace and did not include tool-call results, and it advises affected customers to treat the data as exposed and rotate any secrets pasted into AI chats.

The retention point is the awkward one: Notion says Fireworks had contractually committed to zero data retention for its data, yet the transcripts were held at rest in a dataset. Notion published an official security advisory on November 12, refreshed on November 13, making that breach of commitment explicit. Insights has been disabled, and Notion no longer serves open-weight models through Fireworks.

ANNOUNCED

Four nonprofits sue OpenAI, Anthropic, Google, Meta and SpaceX's AI unit to freeze frontier AI development

Four nonprofit groups — the Irreplaceable Education Fund, Public Citizen, the Heat Initiative and the Alliance for Secure AI — sued Anthropic, Google, Meta, OpenAI and SpaceX's AI business in San Francisco Superior Court on November 12, asking a judge to stop all five from developing or deploying any AI system more capable than their most capable systems on that date, until they can prove the systems are under human control.

The 63-page complaint makes a single claim under California's Unfair Competition Law, arguing the labs' own risk warnings show they are releasing systems they know are unsafe, and citing incidents including OpenAI's July 21 incident report and Anthropic's July review of evaluation runs. Former New Jersey attorney general Matthew Platkin leads the plaintiffs' legal team. The requested freeze would also bar using AI to autonomously research more capable systems.

No court has ever ordered an AI developer to stop building models, and the first hurdle is standing — whether the groups' diverted advocacy resources count as the loss the statute requires. No hearing date has been made public, none of the companies commented to the Associated Press, and the published copy of the complaint carries no case number yet. The allegations are unproven; this is a filing, not an order.

ANNOUNCED

Microsoft 365 outage resolved: Outlook, Teams and Copilot restored, email backlog cleared, cause was a network upgrade fault

A major Microsoft 365 outage that began on November 12 is still running on Friday morning. Microsoft tracks the Exchange Online failure as EX1550543, starting 16:17 UTC, and a wider multi-service incident, MO1550821, covering Outlook on every client, Teams calendars and scheduling, SharePoint features, Microsoft 365 Copilot, Defender, Purview, Planner, To Do and Intune. As of Microsoft's 06:17 UTC update on November 13 the status was still 'service degradation,' with only partial, unsustained recovery.

Downdetector logged roughly 120,000 to 138,000 Outlook reports at peak from more than 70 countries. Copilot's failure mode is dependency, not a model fault: with Exchange down, Copilot Chat may not load and cannot reach users' mail, calendar or files — and reporting notes third-party assistants lost the same access. Teams calls, meetings and chat largely still work; it is the Exchange-dependent pieces that broke.

Microsoft says servers became unexpectedly unavailable after automated health-check failures, that remaining infrastructure buckled under extra load, and that it is not aware of any malicious activity. It is reverting a recent network change and adding capacity, and has promised a full root-cause report once the incident closes. No restoration time has been given.

Update, November 13 midday: recovery is under way but Microsoft had not declared the incident resolved as of 14:35 UTC. At about 14:06 UTC Microsoft said service availability was continuing to improve and email queues were significantly reduced and approaching normal levels, with progress reported in Asia-Pacific and Europe and North America improving. Microsoft has extended its email timeout window to 48 hours so delayed messages stay queued instead of bouncing. Downdetector's peak was about 130,000 Outlook reports. Microsoft still reports no indication of malicious activity and no confirmed root cause, and says it will publish a full post-incident report after recovery completes.

Update, November 13 evening: the outage is mostly over after more than 24 hours. Microsoft said on Friday afternoon that it had addressed the underlying issue and was continuing to restore service to remaining customers and regions, and its service health page showed all services, including Microsoft 365 and Outlook, operational as of about 3 p.m. ET. Downdetector Outlook reports had fallen to about 1,600 by 2:50 p.m. ET, from a peak of about 130,000, and Microsoft 365 reports to about 700. Microsoft apologized for the disruption.

On cause, Microsoft now says the incident was a technical issue not related to cyber activity or AI/agents, introduced by an issue during an earlier network capacity upgrade that surfaced following a recent network change. Access is recovering before mail flow fully normalizes: Microsoft says it is still draining queued messages and rebalancing traffic, gave no ETA for full recovery, and downstream trackers note queued email may take up to 24 hours to deliver. A full post-incident report is promised once the review is complete.

Update, November 14: the incident is closed. Microsoft's final update, at 04:17 UTC on November 14 (11:17 p.m. ET on November 13), marks MO1550821 and EX1550543 as service restored, with an end time of 4:10 p.m. UTC on November 13, recovery confirmed across the previously affected environments and the backlog of email messages processed. Microsoft reiterated the cause — a technical issue from an earlier network capacity upgrade that surfaced following a recent network change — and that it was not related to cyber activity or AI agents. A preliminary post-incident report is due within two business days, with a final report to follow.

RELEASED

Meta's Muse can now call businesses for you — reservations, appointments and hold queues — but only for paid subscribers in the US

Meta opened Muse's business-calling feature to all paid subscribers in the US on November 12, after a limited test that began in September for early adopters who asked for access. Alexandr Wang, Meta's chief AI officer, announced the wider rollout on X, pitching it at the calls nobody wants to make: insurance, cable, the DMV and dinner reservations at places without online booking. Muse can also sit on hold and report back.

Calling sits behind Muse's subscription plans: $20 a month for the Power plan with 500 million Muse tokens per week, or $100 a month for 3 billion tokens. Meta has not said how many calls a plan covers or how many tokens a call uses. Muse passed 10 million users last month, according to TechCrunch.

Wang also said the recently added option to give your Muse its own email address will soon roll out to free users — while calling stays paid-only for now.

ANNOUNCED

Google and Samsung open pre-orders for Gemini smart glasses: $395 from Warby Parker, on sale December 10

Google and Samsung opened pre-orders on November 12 for their first Gemini-powered intelligent eyewear, made with Warby Parker and Gentle Monster and running Android XR. Warby Parker's three styles start at $395; Gentle Monster's six styles start at $449 and run to $489. General sales begin December 10 in stores and online, with first availability in the United States and South Korea.

The glasses are display-free: a 12-megapixel camera, open-ear speakers and a microphone array feed Gemini on a paired phone — Pixel, Samsung Galaxy or iPhone 13 Pro or newer — for voice queries, live translation, object and landmark recognition and contextual follow-ups. Google says battery life is up to nine hours, and its privacy plan includes an outward light, face blurring for bystanders and a physical off switch.

One detail worth knowing before ordering: Samsung's own disclosures say the Gemini experience runs on the phone, not on the glasses' Snapdragon AR1 chip, and Samsung has not published a full comparison of which Gemini features iPhone and non-Galaxy users get.

RELEASED

Google releases Gemini 4 Hydrogen: the fast Flash-tier model now powers Search globally and is open to developers

The leak is now a launch. Google published two posts on Thursday, November 12 naming Gemini 4 Hydrogen — the fast, Flash-tier Gemini 4 model testers had been spotting in blind Arena matchups as gemini-3.8-flash since early November. Google calls it a model built for speed and efficiency in everyday coding and agentic workflows, and it is available starting today.

For developers, Hydrogen handles repository coding and debugging, interactive web development and multimodal workflows, across four configurable thinking levels, standalone or alongside Gemini 4 Carbon and Gemini 4 Argon. It is open now through the Gemini API in Google AI Studio and Android Studio, and through Gemini for business; paid Google AI Pro and Ultra subscribers can try it in Google Antigravity. Google's posts give no API price, context window or benchmark table.

The bigger consumer story is Search. Google says Hydrogen is now live globally in Google Search for all users — 'the biggest model upgrade to Search yet' — with better handling of multi-part and typo-laden questions, stronger query fan-out to more sources, more visual and structured answers, and richer responses across AI Mode languages.

This confirms the Wire's November 7 rumor entry, which is now marked confirmed. The rumored specs that circulated during the leak — 300–500 tokens per second, a guessed $0.75/$3.75 per million tokens — were tester speculation, and Google's launch posts do not repeat them.

Update, November 12 evening: the details Google's launch posts left out are now filled in by its developer documentation and independent evaluation. Paid API pricing is $1.50 per million input tokens and $7.50 per million output tokens, with a rate-limited free tier in AI Studio; Hydrogen defaults to a high thinking level, adds a new xhigh level, and scores 51 on Artificial Analysis's Intelligence Index at high — one point above the heavier Gemini 4 Argon — rising to 54 at xhigh. Google says it finishes tasks 2.5 times faster than Gemini 3.8 Flash. In Antigravity, a new Orchestrator option splits planning from execution between Carbon and Hydrogen, rolling out to Ultra subscribers first. One caveat from Google's own model card: Hydrogen rejects more harmless requests than Gemini 3.8 Flash did.

ANNOUNCED

Anthropic's Supported Regions Policy takes effect: Claude is closed to users in China, Russia and other unlisted regions — and to companies controlled from there

Anthropic's Supported Regions Policy took effect on November 12, the same day its rewritten Usage Policy came into force. Anthropic's products — Claude.ai, Claude Code, Cowork, the API and developer platform, and Claude models offered through cloud platforms and authorised resellers — are available only in the countries and regions on its published list, which runs to about 185 entries.

Mainland China, Hong Kong and Macau are not on the list. Neither are Russia, Belarus, Iran, North Korea, Cuba, Syria, Venezuela, Afghanistan, Myanmar or Yemen. Taiwan is on it, and Ukraine is listed except for Crimea, Donetsk, Kherson, Luhansk and Zaporizhzhia.

The policy names three separate ways to fall outside it: being physically located in an unsupported region, visitors included; being incorporated or headquartered there, wherever your staff sit; and being majority-owned or controlled, directly or indirectly, by people or entities there — the test aimed at subsidiaries of Chinese, Russian, Iranian or North Korean parents operating from London or San Francisco.

One honest bound: the country list itself appears unchanged from the version dated May 2026 — Meduza compared the two and found them identical — so November 12 mainly formalises and clarifies enforcement language rather than adding new countries. Reports of identity checks and account bans around the date are user reports; Anthropic has published no count of accounts affected.

ANNOUNCED

Anthropic opens self-hosted Claude Managed Agents in private beta: the full agent service runs in the customer's own cloud

Anthropic has released a self-hosted version of Claude Managed Agents in private beta, opened on November 12. Customers can now run the whole service — the agent harness, sandboxes, credentials and session history — in their own cloud account or on-premises data center, instead of only self-hosting an agent's sandboxes while Anthropic ran the agent itself.

The customer operates a server, cma-server, which serves the same Managed Agents API as Anthropic's hosted service, runs the agent loop and stores agents, sessions and their histories, backed by PostgreSQL 16 or later. Model requests still go out to a chosen provider — the Claude API by default, or Amazon Bedrock, Claude Models on Google Cloud, Microsoft Foundry, or a gateway the customer runs — so self-hosting the harness does not by itself keep model traffic inside the customer's network.

Access is gated: customers must ask their Anthropic account team for the download and install each release themselves, and nothing is copied over from the hosted service — agents, skills, vault credentials and memory stores must be recreated. Some hosted features are unavailable, including scheduled deployments, dreams and webhooks.

RELEASED

Anthropic launches Claude for Manufacturing: a 17-skill plugin for quotes, quality investigations and breakdowns

Anthropic launched Claude for Manufacturing on Thursday, November 12 — a plugin for Claude Team and Enterprise customers that packages 17 skills for plant work: drafting quotes for new jobs, investigating failed inspections, diagnosing breakdowns, sourcing from another supplier, comparing designs and reporting on production.

The plugin connects to four manufacturing systems — Autodesk Fusion, Infor, Microsoft 365 and Oracle NetSuite — and can also be used with systems that have no connector. Skills can run singly or chained in sequence, for example turning a customer request into a draft quote, and Anthropic says new skills will be added periodically.

The human stays in charge by design: workflows end with a person reviewing and signing off before anything reaches a customer or supplier, and the plugin does not operate equipment. Infor, whose ERP powers the plugin, is the launch partner quoted alongside Anthropic.

RELEASED

Claude Code gets an experimental Code Mode: Claude writes JavaScript to call its own tools instead of invoking them one by one

Anthropic shipped Claude Code v2.1.323 on November 12 with an experimental, opt-in feature called Code Mode. Instead of invoking its tools one call at a time, Claude writes a JavaScript program that calls the tools itself — one round trip instead of ten on multi-step work.

Code Mode is off by default, and organisations can turn it off entirely. The trade-off is inspection: a program that calls tools is harder to review before it runs than a single visible tool call, which is why Anthropic has not enabled it for everyone.

The same release makes plugins honour version ranges and tightens sign-in safety: Claude Code now requires user confirmation before an MCP server can start a sign-in flow.

ANNOUNCED

Anthropic unveils Anvil: its automated red-teamer now finds stronger jailbreaks than all its human bug-bounty hunters combined

Anthropic has disclosed Anvil, an internal automated red-teaming system that uses frontier internal versions of Claude — with safeguards disabled — to hunt for jailbreaks against its own bioweapons, cybersecurity and anti-distillation safeguards before new defenses ship. Anthropic says Anvil now outperforms expert human red-teamers, meeting a goal set in its Responsible Scaling Policy.

The gap is starkest in biology. In a recent bio-classifier bug bounty, human red-teamers spent 430 hours over seven weeks and found a high-severity jailbreak on only 1 of 8 questions; Anvil matched or beat the best human jailbreak on every classifier tested, in 24 to 48 hours, and found universal jailbreaks humans never did. On cyber tasks it got a safeguarded model to write working exploits for known Chrome and Firefox JavaScript-engine vulnerabilities.

Anvil is an internal tool, not a product — but Anthropic is candid about the implication: its power grows with the underlying model, so attackers could soon build similar systems on models with weaker safeguards. The post is a safeguards disclosure, not a release, and Anthropic says it does not observe jailbreaks of this sophistication in real traffic today.

2026-11-11

RELEASED

SpaceXAI releases Grok Bot 1.0: shared Team Bots move into Slack, Microsoft Teams and video calls

SpaceXAI has released Grok Bot 1.0, recasting its AI bots from personal assistants into shared teammates. A Team Bot can be shared with a whole team, brought into Slack and Microsoft Teams, and invited into video calls on Google Meet, Zoom and Teams, where it can answer in chat, track unfinished work and offer to take on tasks without being asked directly.

Team Bots are available on Team and Enterprise plans, and Meetings on Team, Enterprise and Ultra plans, with Enterprise admins needing to switch meetings on. SpaceXAI's launch post on x.ai is dated November 2, following the August beta; the public launch push — a Grok Bot post on X amplified by Elon Musk — landed on November 11, which is the date this entry carries.

The release lists no pricing beyond the plan tiers and no latency or task-success benchmarks, so the claims about better browsing and more efficient task completion are the company's own. This launch also confirms the October 29 rumor on the Wire that Grok Bot 1.0 installers had appeared ahead of any announcement; that entry is now marked confirmed.

ANNOUNCED

United becomes the first major US airline to ban AI agents from booking flights — it can cancel any reservation an agent made

United Airlines revised its Contract of Carriage on Wednesday, November 11, to ban unauthorized third-party personal AI agents from booking, changing or servicing its tickets. New Rule 6.J.6 lists "the use of AI Agents to make, modify or service a Ticket, and/or to purchase any service or product" as a prohibited practice, and Rule 5.H lets United cancel reservations made, modified or serviced by AI agents without notice. The ban covers United's own site and app and bookings routed through travel agencies and aggregators.

Agents can still look up and show United flights and prices — the customer has to complete the purchase themselves. United cited data security, warning that handing account credentials to an assistant can expose sensitive information and lead to unintended purchases or use of MileagePlus miles, and a compliance point: online check-in requires a passenger to personally acknowledge rules on prohibited items, which an automated agent completing that step could put at risk. Bookings an agent made before November 11 will not be canceled.

United says it does not yet work with any third-party AI agents and wants standardized, secure integrations first. CEO Scott Kirby has put the commercial motive plainly: he does not want another middleman between the airline and its customers. The move follows eBay's apparent block on third-party agents, which the Wire covered on November 10, and directly affects assistants such as Meta's Muse and OpenAI's Dots, which can still search fares but cannot buy.

ANNOUNCED

TSMC board approves up to $1B in AI investments: $500M in Anthropic at its IPO price, $500M in OpenAI preferred stock

TSMC's board has approved investments of up to $1 billion in the two leading U.S. AI labs: up to $500 million in Anthropic common shares at its per-share IPO price, and up to $500 million in OpenAI Group preferred stock. TSMC announced the approvals in an official statement alongside $40.93 billion in cleared capex and Q3 results.

The Anthropic leg ties TSMC directly to the IPO now in roadshow — 825 million shares at $70 to $78, seeking up to about $64.4 billion, with trading expected November 19. No closing date or final invested amount was disclosed for either investment.

For the world's largest contract chipmaker, the move is strategic as much as financial: TSMC manufactures the accelerators both labs depend on, and equity stakes deepen those ties ahead of Anthropic's public debut.

ANNOUNCED

Google Labs introduces Tabletop: groups prompt AI to build shared planning tools — itineraries, polls, expense splitters and brackets

Google Labs introduced Tabletop on November 11, an experiment for group planning where everyone contributes links, files and ideas to one shared space, then prompts @tabletop — or picks a starter template — to generate interactive tools inside it. Google's examples are trip itineraries, dinner polls, expense splitters and tournament brackets, and a demo shows a basic Pong game turned into Pickle Pong with a live scoreboard after a text request.

Sharing works through an invite link dropped into a group chat, with everyone able to use and edit the space at the same time, much like a Google Doc. Google Labs design and engineering lead Roman Nurik wrote the announcement. Google has not said which model runs Tabletop behind the scenes.

Access is narrow for now: a waitlist at labs.google/tabletop for people 18 and over in the US, with new tables opening weekly as waitlisted users are admitted. There is no price, no international timeline and no promise it becomes a full product — Labs experiments can change or be dropped.