DayWeek
The week

Week 46

8 stories in one edition, 9 to 9 November. Six days to go.

Mo 98 Tu 10 We 11 Th 12 Fr 13 Sa 14 Su 15

Most read this week

read on screen, counted from 15 Oct
1Mon 9 Nov · 11 readsClaude Code Projects is in beta on Pro, Max and Ultra plans only2Mon 9 Nov · 4 readsGoogle Cloud's Auth Manager for AI agents is now generally available3Mon 9 Nov · 1 readNvidia's board opens an investigation after reports a foreign intelligence effort targeted Huang

Monday 9 November

Good morning. Google Cloud says its Auth Manager, which automates OAuth sign-in and token refresh for AI agents, became generally available on 5 November 2026. Here is what that covers, and two more stories worth your time.

Open the edition →
Safety and security Google Cloud's Auth Manager for AI agents is now generally available 4
Agents and apps Claude Code Projects is in beta on Pro, Max and Ultra plans only Most read · 11
Money and compute Nvidia's board opens an investigation after reports a foreign intelligence effort targeted Huang 1
Laws and rulings What the H200 export rule requires before China licences get case-by-case review
Research Nvidia reports its Reshard call cuts DeepSeek-V3 weight refit from 14.9 to 2.6 seconds in its own tests.
Money and compute Nvidia, Rebellions and SK Telecom plan a Korean inference system, with R200 systems expected in 2028.
Agents and apps Eligible Claude Enterprise organizations can now tag projects with cost codes and see most spend by tag, in beta.
Models and releases Nvidia publishes Dynamo recipes for four Korean open models, which it says deliver up to 2x performance.
2026-11-09 notis
OTHER NOTIS SITES AIYOU ARE HERE AI models, prices and agents. ainotis.com Hiking Trail openings, hut bookings, and route conditions. hiking.notis.news Photography Cameras, lenses and firmware. photo.notis.news More areas when we can cover them every day.
CATEGORY Join
Show Read as Find Sections Models and releasesAgents and appsResearchMoney and computeLaws and rulingsSafety and securityAI at work More The monthlyArchiveModelsAboutCorrectionsThe numbersRSS
My notis
THIS WEEK'S NEWS
November 2026 3 STORIES
MTWTFSS
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Days with a dot have news. See the archive.

Show one day ›
03:38
0/3
WEATHER THEME CARD SPACING

Today's news

today

My page

Kept in this browser only.

SAVED FOLLOWING COLLECTIONS READING SINCE
SAVED STORIES
Nothing saved yet. Tap the bookmark on a card to keep it here. read today
CASES YOU FOLLOW all cases

No case followed yet.

COLLECTIONS
Price changes Stories where a published price went up or down.
Models being retired Models being withdrawn, and gaps before a replacement is ready.
THE MORNING EMAIL The day's top stories. Sign up
Safety 9 NOV

Google Cloud's Auth Manager for AI agents is now generally available

Agent identityAuth ManagerCodeMenderGoogleModel ArmorOAuth
Illustration: A wall-mounted access gate with a slim card reader and a sliding lock bolt; the green reader and bolt stand for automated sign-in for AI agents.Illustration made with AI for ai notis

Google says Auth Manager automates OAuth sign-in and token refresh for AI agents; JWT support for Agent Identity, a Model Armor topic detector and CodeMender sandboxing are in preview.

SOURCEGoogle Cloud · 1 source
CHECKED9 Nov, 05:26
TOPICSAgent identity in directories and OAuth
Safety read
Illustration: A wall-mounted access gate with a slim card reader and a sliding lock bolt; the green reader and bolt stand for automated sign-in for AI agents. Safety / 9 NOV

Google Cloud's Auth Manager for AI agents is now generally available
  • Google Cloud says Auth Manager became generally available on 5 November 2026; it automates OAuth 2.0 and PKCE flows and refreshes user-delegated tokens for AI agents.
  • Google says Agent Identity JWT support, in preview, issues OIDC tokens so agents can authenticate across AWS, Azure, on-premises backends and MCP servers without static, long-lived API keys.
  • Also in preview: Workload Identity Federation for GKE in VPC Service Controls, Cloud Armor exploit-protection rules, a Model Armor custom topic detector and sandboxed CodeMender verification.

Google Cloud made Auth Manager generally available on 5 November 2026. It automates OAuth 2.0 and PKCE flows and refreshes user-delegated tokens for AI agents, which Google says reduces consent fatigue and the theft of credentials.

The release comes in a Google Cloud security blog post of the same date. Alongside Auth Manager, Agent Identity JWT support is in preview. Google says it issues standard OIDC tokens so agents can authenticate across AWS, Azure, on-premises backends and MCP servers without static, long-lived API keys.

Two network controls are generally available: IAM roles in VPC Service Controls ingress and egress rules, and the Cloud NGFW advanced malware sandbox, which Google says is enabled by Palo Alto Networks WildFire. In preview are Workload Identity Federation for GKE in VPC Service Controls, Cloud Armor managed rules for exploit protection, and a Model Armor custom topic detector that lets teams enforce topic adherence for agents.

Google also lists new CodeMender capabilities in preview. CodeMender can now run vulnerability verification and exploit simulation inside an operating-system-level sandbox or an isolated virtual machine, and can scan a folder, a whole repository or only files changed since the last scan.

What it means for you Our view

If you build or buy agents that act on a user's behalf in Google Cloud, Auth Manager is the part of this release Google says is generally available, and it handles the sign-in and token refresh work that Google says otherwise causes consent fatigue and credential theft. The Agent Identity JWT support, Model Armor detector and CodeMender sandboxing are previews, so they are worth testing but not yet worth building production on. Check which of your agents authenticate with static, long-lived API keys today, and list which of those flows the preview JWT support would replace.

TAGSAgent identityAuth ManagerCodeMenderGoogleModel ArmorOAuth
Google Cloud · 1 source tap to close
Tooling 9 NOV

Claude Code Projects is in beta on Pro, Max and Ultra plans only

AnthropicClaude CodeClaude Code Projects
Illustration: A dispatch console with a long central tray and small drawers on rails; the green tray stands for one long-running project coordinating separate threads of work.Illustration made with AI for ai notis

Anthropic's documentation says Projects in Claude Code, where Claude acts as coordinator for threads, is rolling out gradually and uses plan limits faster than other sessions.

SOURCEAnthropic · 1 source
CHECKED9 Nov, 05:26
Tooling read
Illustration: A dispatch console with a long central tray and small drawers on rails; the green tray stands for one long-running project coordinating separate threads of work. Tooling / 9 NOV

Claude Code Projects is in beta on Pro, Max and Ultra plans only
  • Anthropic's Claude Code documentation says Projects is in public beta on Pro, Max and Ultra plans and rolling out gradually; Team and Enterprise plans do not have it yet.
  • A project is one long-running session in which Claude acts as coordinator and decides what becomes a thread; each thread is a separate session with its own context window.
  • A cloud thread works on its own branch and opens a pull request when the work calls for one, and keeps going after you close your laptop.

Anthropic's Claude Code documentation says Projects is in public beta on the Pro, Max and Ultra plans and is rolling out gradually; Team and Enterprise plans do not have it yet.

A project is one long-running conversation in which Claude acts as coordinator. It decides what becomes a thread and tracks each one, while each thread is a separate session with its own context window. A cloud thread works on its own branch and opens a pull request when the work calls for one.

Anthropic says cloud threads keep going after a laptop is closed, and that a task needing something only the user's computer has can be run as a thread on that computer through Remote Control. The documentation says a project draws on the same plan limits as other Claude Code sessions and uses them faster.

What it means for you Our view

If you are on a Pro, Max or Ultra plan, Projects is a way to give Claude Code a larger piece of work and let it split that work into threads, each on its own branch. Anthropic's documentation says a project draws on the same plan limits as other Claude Code sessions and uses them faster, so a big job will take more of your allowance. Team and Enterprise plans do not have it yet, and the rollout is gradual, so not every eligible account will see it today. Try it on one contained task, and compare how much of your plan limit it uses against a normal session on similar work.

TAGSAnthropicClaude CodeClaude Code Projects
Anthropic · 1 source tap to close
toned down Compute 9 NOV

Nvidia's board opens an investigation after reports a foreign intelligence effort targeted Huang

Corporate governanceNvidia
Illustration: A sealed inspection cabinet with a glass front and an arm-mounted lamp; the green lamp stands for an independent review opened by a company board.Illustration made with AI for ai notis

Independent directors will oversee it, and the board has no reason to believe national security was compromised. Nvidia calls itself "the world leader in AI and accelerated computing".

SOURCENVIDIA · 3 sources
CHECKED9 Nov, 05:40
Compute read
Illustration: A sealed inspection cabinet with a glass front and an arm-mounted lamp; the green lamp stands for an independent review opened by a company board. Compute / 9 NOV

Nvidia's board opens an investigation after reports a foreign intelligence effort targeted Huang
  • Nvidia's board statement of 8 November 2026 says a committee of independent directors, assisted by outside independent counsel, will oversee the investigation.
  • The board says it has no reason to believe that national security or Nvidia secrets were compromised, and that its inquiry will examine all available information.
  • The statement names no country, does not describe the reporting and gives no timetable; the board says it will provide further updates as appropriate.

Nvidia's board has started an independent investigation after reporting that chief executive Jensen Huang was the target of a foreign intelligence effort, the board said in a statement published on 8 November 2026.

A committee of independent directors, assisted by outside independent counsel, will oversee the investigation. The board says it has no reason to believe that national security or Nvidia secrets were compromised, but that its inquiry will examine all available information.

The board says Huang supports the review and has committed to cooperating, that it has confidence in him and his leadership team, and that it will provide further updates as appropriate. Huang said the same in an email to employees that Nvidia published on its newsroom the same day: he supports the investigation and will cooperate fully.

The board's statement does not name a country, describe the reporting it refers to or give a timetable for the investigation.

What it means for you Our view

For customers and partners this is, so far, a governance matter at one supplier. The board has said it has no reason to believe national security or Nvidia secrets were compromised, and it has also named no country, described no reporting and set no timetable, so the facts are not yet established and a statement of confidence is not a finding. Nothing in the statement changes a purchase, a contract or a supply plan. Jensen Huang supports the review and the board has promised updates as appropriate, so watch Nvidia's newsroom for the next one and read it for scope and timing before drawing conclusions.

TAGSCorporate governanceNvidia
NVIDIA · 3 sources tap to close
NOTHING IN THIS FILTER No story today matches. open the archive →
That's today's news: three stories. ExplainedWhat the H200 export rule requires before China licences get case-by-case review

The morning email

On the mornings we publish: the three top stories and up to four short ones. Free.

We email you a link to confirm. An issue may include one sponsor, always labelled Sponsored · Advertisement. Our emails count opens and clicks, not who made them. Unsubscribe in one click. What we keep

1 of 3
Illustration: A wall-mounted access gate with a slim card reader and a sliding lock bolt; the green reader and bolt stand for automated sign-in for AI agents.Illustration made with AI for ai notis

Safety · 9 NOV

Google Cloud's Auth Manager for AI agents is now generally available

Google says Auth Manager automates OAuth sign-in and token refresh for AI agents; JWT support for Agent Identity, a Model Armor topic detector and CodeMender sandboxing are in preview.

  • Google Cloud says Auth Manager became generally available on 5 November 2026; it automates OAuth 2.0 and PKCE flows and refreshes user-delegated tokens for AI agents.
  • Google says Agent Identity JWT support, in preview, issues OIDC tokens so agents can authenticate across AWS, Azure, on-premises backends and MCP servers without static, long-lived API keys.
  • Also in preview: Workload Identity Federation for GKE in VPC Service Controls, Cloud Armor exploit-protection rules, a Model Armor custom topic detector and sandboxed CodeMender verification.
Checked 9 Nov, 05:26 · Google Cloud · 1 source
Read the story
Illustration: A dispatch console with a long central tray and small drawers on rails; the green tray stands for one long-running project coordinating separate threads of work.Illustration made with AI for ai notis

Tooling · 9 NOV

Claude Code Projects is in beta on Pro, Max and Ultra plans only

Anthropic's documentation says Projects in Claude Code, where Claude acts as coordinator for threads, is rolling out gradually and uses plan limits faster than other sessions.

  • Anthropic's Claude Code documentation says Projects is in public beta on Pro, Max and Ultra plans and rolling out gradually; Team and Enterprise plans do not have it yet.
  • A project is one long-running session in which Claude acts as coordinator and decides what becomes a thread; each thread is a separate session with its own context window.
  • A cloud thread works on its own branch and opens a pull request when the work calls for one, and keeps going after you close your laptop.
Checked 9 Nov, 05:26 · Anthropic · 1 source
Read the story
Illustration: A sealed inspection cabinet with a glass front and an arm-mounted lamp; the green lamp stands for an independent review opened by a company board.Illustration made with AI for ai notis

Compute · 9 NOVnarrowed

Nvidia's board opens an investigation after reports a foreign intelligence effort targeted Huang

Independent directors will oversee it, and the board has no reason to believe national security was compromised. Nvidia calls itself "the world leader in AI and accelerated computing".

  • Nvidia's board statement of 8 November 2026 says a committee of independent directors, assisted by outside independent counsel, will oversee the investigation.
  • The board says it has no reason to believe that national security or Nvidia secrets were compromised, and that its inquiry will examine all available information.
  • The statement names no country, does not describe the reporting and gives no timetable; the board says it will provide further updates as appropriate.
Checked 9 Nov, 05:40 · NVIDIA · 3 sources
Read the story

You're all caught up

That's today's news: three stories.

ExplainedWhat the H200 export rule requires before China licences get case-by-case review
Start again

You can join ai notis soon, to follow cases and keep stories on every device.

Follow us on Instagram and Threads

The morning email

On the mornings we publish: the three top stories and up to four short ones. Free.

We email you a link to confirm. An issue may include one sponsor, always labelled Sponsored · Advertisement. Our emails count opens and clicks, not who made them. Unsubscribe in one click. What we keep

No story today matches. Open the archive
Illustration: A wall-mounted access gate with a slim card reader and a sliding lock bolt; the green reader and bolt stand for automated sign-in for AI agents. Illustration made with AI for ai notis Safety · 9 NOV Google Cloud's Auth Manager for AI agents is now generally available Google Cloud · 1 source · 9 Nov, 05:26 Illustration: A dispatch console with a long central tray and small drawers on rails; the green tray stands for one long-running project coordinating separate threads of work. Illustration made with AI for ai notis Tooling · 9 NOV Claude Code Projects is in beta on Pro, Max and Ultra plans only Anthropic · 1 source · 9 Nov, 05:26 Illustration: A sealed inspection cabinet with a glass front and an arm-mounted lamp; the green lamp stands for an independent review opened by a company board. narrowedIllustration made with AI for ai notis Compute · 9 NOV Nvidia's board opens an investigation after reports a foreign intelligence effort targeted Huang NVIDIA · 3 sources · 9 Nov, 05:40

That's today's news: 3 stories.Archive

Today's stories 3 STORIES

Today's stories

Week 46 · 3 STORIES
  1. Safety · 9 NOV

    Google Cloud's Auth Manager for AI agents is now generally available

    Illustration: A wall-mounted access gate with a slim card reader and a sliding lock bolt; the green reader and bolt stand for automated sign-in for AI agents.Illustration made with AI for ai notis

    Google says Auth Manager automates OAuth sign-in and token refresh for AI agents; JWT support for Agent Identity, a Model Armor topic detector and CodeMender sandboxing are in preview.

    Checked 9 Nov, 05:26 · Google Cloud · 1 sourceRead

  2. Tooling · 9 NOV

    Claude Code Projects is in beta on Pro, Max and Ultra plans only

    Illustration: A dispatch console with a long central tray and small drawers on rails; the green tray stands for one long-running project coordinating separate threads of work.Illustration made with AI for ai notis

    Anthropic's documentation says Projects in Claude Code, where Claude acts as coordinator for threads, is rolling out gradually and uses plan limits faster than other sessions.

    Checked 9 Nov, 05:26 · Anthropic · 1 sourceRead

  3. Compute · 9 NOV · narrowed

    Nvidia's board opens an investigation after reports a foreign intelligence effort targeted Huang

    Illustration: A sealed inspection cabinet with a glass front and an arm-mounted lamp; the green lamp stands for an independent review opened by a company board.Illustration made with AI for ai notis

    Independent directors will oversee it, and the board has no reason to believe national security was compromised. Nvidia calls itself "the world leader in AI and accelerated computing".

    Checked 9 Nov, 05:40 · NVIDIA · 3 sourcesRead

Safety 9 NOV

Google Cloud's Auth Manager for AI agents is now generally available

Google says Auth Manager automates OAuth sign-in and token refresh for AI agents; JWT support for Agent Identity, a Model Armor topic detector and CodeMender sandboxing are in preview.

Google Cloud · 1 source·checked 9 Nov, 05:26
Agent identityAuth ManagerCodeMenderGoogleModel ArmorOAuth
Illustration: A wall-mounted access gate with a slim card reader and a sliding lock bolt; the green reader and bolt stand for automated sign-in for AI agents.Illustration made with AI for ai notis
Tooling

Claude Code Projects is in beta on Pro, Max and Ultra plans only

Anthropic's documentation says Projects in Claude Code, where Claude acts as coordinator for threads, is rolling out gradually and uses plan limits faster than other sessions.

9 NOV · Anthropic · 1 source
toned down Compute

Nvidia's board opens an investigation after reports a foreign intelligence effort targeted Huang

Independent directors will oversee it, and the board has no reason to believe national security was compromised. Nvidia calls itself "the world leader in AI and accelerated computing".

9 NOV · NVIDIA · 3 sources

Explained

One idea behind the news
Laws What the H200 export rule requires before China licences get case-by-case review BIS's final rule moves exports of Nvidia's H200 and equivalent chips to China from a presumption of denial to case-by-case review, if the exporter certifies set conditions.
In brief 4 more today
0 of 3 read
Drag a card to start Drag left or right · tap to read more Hover to lift a card · tap to bring it forward Tap again to put it back Tap a card to open it Tap again to close it Tap a headline to read it as a card