2026-10-28 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 ArchiveModelsAboutCorrectionsThe numbersRSS
My notis
TODAY'S NEWS
October 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 31

Days with a dot have news. See the archive.

The whole week ›
10:28
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Models being retired Models being withdrawn, and gaps before a replacement is ready.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 28 OCT

Anthropic's unreleased research model found five attack chains, since patched or addressed

AnthropicChromeFrontier Red TeamLinuxVulnerability disclosureWindows
Illustration: A sealed inspection cabinet with a row of test bays; its green bays stand for the five attack chains the unreleased model found.Illustration made with AI for ai notis

Anthropic does not name the internal model. Its post says the Chrome, Safari, Android, Windows and Linux flaws were fixed or addressed, and does not say any was exploited.

SOURCEAnthropic · 1 source
CHECKED28 Oct, 05:40
Safety read
Illustration: A sealed inspection cabinet with a row of test bays; its green bays stand for the five attack chains the unreleased model found. Safety / 28 OCT

Anthropic's unreleased research model found five attack chains, since patched or addressed
  • Anthropic's 27 October 2026 post describes five attack chains that led to remote code execution and other critical impacts, found with an unreleased internal model it does not name.
  • Anthropic says 62 Chrome CVEs were patched as of Chrome 156. For the Windows kernel QUIC flaw, CVE-2026-33096, it estimates likely several million such servers listen on the internet.
  • The post does not say whether any flaw was exploited in the wild. Anthropic's policy gives developers 90 days to patch and a further 45 days before it discusses details.

Anthropic published a post on 27 October 2026 saying it used an unreleased, internal research model, which it does not name, to find critical vulnerabilities in widely used software. It shares details of five attack chains that led to remote code execution and other critical impacts. In Chrome, it says 62 CVEs have been assigned and patched as of Chrome 156. In Windows, a kernel QUIC flaw that can crash servers answering HTTP/3 requests was resolved by Microsoft as CVE-2026-33096, and Anthropic estimates there are likely several million such servers listening on the internet. A Linux flaw, CVE-2026-98432, was patched upstream and backported to all maintained stable kernel branches. Apple acknowledged the Safari report the same day it was submitted, and Anthropic says Google has addressed the Android lock-screen issue. Anthropic says it gives developers 90 days to patch and a further 45 days before it discusses details. The post does not say whether any of the flaws was exploited in the wild.

What it means for you Our view

Anthropic reports that each of the five chains has been patched or addressed, so the practical question is whether your own systems carry the fixes. The post does not say any flaw was exploited and does not name the model, so the account of what the model did rests on Anthropic alone. The figure of several million Windows servers is Anthropic's own estimate. If you run Windows servers that use QUIC, or Linux kernels, ask your vendor to confirm that the fixes for CVE-2026-33096 and CVE-2026-98432 are installed, and check that your Chrome browsers are on version 156 or later.

TAGSAnthropicChromeFrontier Red TeamLinuxVulnerability disclosureWindows
Anthropic · 1 source tap to close
Compute 28 OCT

Anthropic offers governments 40% off Claude through a new GSA agreement

AnthropicClaude EnterpriseClaude for GovernmentFedRAMPGSAUS
Illustration: A bench scale with a rack of sealed units beside it; its green price dial stands for the 40% government discount.Illustration made with AI for ai notis

The discount covers Claude Enterprise, Claude for Government and the Claude API, has no minimum spend, and reaches state, local, tribal and territorial governments, Anthropic says.

SOURCEAnthropic · 1 source
CHECKED28 Oct, 05:40
Compute read
Illustration: A bench scale with a rack of sealed units beside it; its green price dial stands for the 40% government discount. Compute / 28 OCT

Anthropic offers governments 40% off Claude through a new GSA agreement
  • Anthropic says the agreement gives government partners 40% off Claude Enterprise, Claude for Government and the Claude API, available through GSA's Multiple Award Schedule today.
  • Agencies are not bound by a minimum spend commitment and pay for the usage they consume. The discount also extends to state, local, tribal and territorial governments.
  • Anthropic estimates the discounts will save the government $1 billion, and says approximately 500 US federal, state and local government organizations now use Claude.

Anthropic announced on 27 October 2026 a new agreement with the US General Services Administration that gives government partners a 40% discount on Claude Enterprise, Claude for Government and the Claude API. The discount also extends to state, local, tribal and territorial governments. Agencies are not bound by a minimum spend commitment and pay for the usage they consume. Anthropic estimates the discounts will save the government $1 billion. Agencies can buy Claude at the discount through GSA's Multiple Award Schedule today. Anthropic says Claude for Government, which carries a FedRAMP High authorization, became generally available on 30 September, and that Claude on Amazon Bedrock and on Google Cloud's Vertex AI is also authorized at FedRAMP High. Claude Enterprise is in process for FedRAMP Moderate through the FedRAMP 20x program. Anthropic says about 500 US federal, state and local government organizations now use Claude, and that it is launching a program to help agencies implement it.

What it means for you Our view

If you buy AI for a government body, ask whether your Claude purchase can move onto GSA's Multiple Award Schedule at the 40% discount, which Anthropic says agencies can do today. The $1 billion saving and the count of approximately 500 organizations are both Anthropic's own figures. Authorization differs by product: Anthropic says Claude for Government is offered at a FedRAMP High authorization level, while Claude Enterprise is only in process for FedRAMP Moderate. Before you order, ask Anthropic or the reseller for each product's current FedRAMP status in writing, and check it against what your security team requires.

TAGSAnthropicClaude EnterpriseClaude for GovernmentFedRAMPGSAUS
Anthropic · 1 source tap to close
toned down Research 28 OCT

Dario Amodei says models will regularly do what seemed impossible in 12 months

AnthropicGPT-2The Builders
Dario & Daniela Amodei: The Rise of Superintelligent AI, the End of Human Control & Humanity's Fate: a video from The Builders on YouTubeVideo: The Builders on YouTube

In a Walter Isaacson interview on The Builders, published 27 October 2026, Anthropic's chief executive also describes his reaction to GPT-2 and his fortnightly talks to the company.

SOURCEThe Builders (YouTube) · 2 sources
CHECKED28 Oct, 05:41
Research watch read
Dario & Daniela Amodei: The Rise of Superintelligent AI, the End of Human Control & Humanity's Fate: a video from The Builders on YouTube Research / 28 OCT

Dario Amodei says models will regularly do what seemed impossible in 12 months
  • On GPT-2, Dario Amodei said he was "super freaked out by it": scaling it up a factor of a million would make it "better than humans at everything" (13:20).
  • He said: "But we just announced that, you know, we're going to be developing drugs for, you know, neglected rare diseases" (14:36).
  • Hard things would be easy in six months, he said, and "12 months from now, the things that seemed impossible, the models are going to be doing them regularly" (14:59).

The Builders, a new interview series hosted by Walter Isaacson, published an interview with Anthropic's co-founders Dario Amodei (chief executive) and Daniela Amodei (president) on YouTube on 27 October 2026. Asked about the moment GPT-2 appeared, Dario Amodei said he was "super freaked out by it" and that he thought scaling it up another factor of a million would make it "better than humans at everything" (13:20). He said: "But we just announced that, you know, we're going to be developing drugs for, you know, neglected rare diseases" (14:36). He also said that in six months the things that seemed hard will be easy and in 12 months the models will do regularly what seemed impossible (14:59). He also described fortnightly talks to the company, which he calls DVQs, and said nothing he says in them is an actual decision (22:43). The captions are YouTube's automatic ones, so wording is as the captions give it.

What it means for you Our view

These are the views of the chief executive of a company that sells AI models, taken from YouTube's automatic captions. Amodei gives no benchmark or date for what he expects to become possible, so the 12-month statement is a forecast to note, not a basis for a budget. If you plan AI work for next year, list the tasks your team now considers too hard and set a date to test them again against current models, at six months and at 12.

TAGSAnthropicGPT-2The Builders
The Builders (YouTube) · 2 sources tap to close
NOTHING IN THIS FILTER No story today matches. open the archive →
That's today's news: three stories. Today's explainer was dropped and does not run.

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 sealed inspection cabinet with a row of test bays; its green bays stand for the five attack chains the unreleased model found. Illustration made with AI for ai notis

Safety · 28 OCT

Anthropic's unreleased research model found five attack chains, since patched or addressed

Anthropic does not name the internal model. Its post says the Chrome, Safari, Android, Windows and Linux flaws were fixed or addressed, and does not say any was exploited.

  • Anthropic's 27 October 2026 post describes five attack chains that led to remote code execution and other critical impacts, found with an unreleased internal model it does not name.
  • Anthropic says 62 Chrome CVEs were patched as of Chrome 156. For the Windows kernel QUIC flaw, CVE-2026-33096, it estimates likely several million such servers listen on the internet.
  • The post does not say whether any flaw was exploited in the wild. Anthropic's policy gives developers 90 days to patch and a further 45 days before it discusses details.
Checked 28 Oct, 05:40 · Anthropic · 1 source
Read the story
Illustration: A bench scale with a rack of sealed units beside it; its green price dial stands for the 40% government discount. Illustration made with AI for ai notis

Compute · 28 OCT

Anthropic offers governments 40% off Claude through a new GSA agreement

The discount covers Claude Enterprise, Claude for Government and the Claude API, has no minimum spend, and reaches state, local, tribal and territorial governments, Anthropic says.

  • Anthropic says the agreement gives government partners 40% off Claude Enterprise, Claude for Government and the Claude API, available through GSA's Multiple Award Schedule today.
  • Agencies are not bound by a minimum spend commitment and pay for the usage they consume. The discount also extends to state, local, tribal and territorial governments.
  • Anthropic estimates the discounts will save the government $1 billion, and says approximately 500 US federal, state and local government organizations now use Claude.
Checked 28 Oct, 05:40 · Anthropic · 1 source
Read the story
Dario & Daniela Amodei: The Rise of Superintelligent AI, the End of Human Control & Humanity's Fate: a video from The Builders on YouTube Watch on YouTubeVideo: The Builders on YouTube

Research · 28 OCTnarrowed

Dario Amodei says models will regularly do what seemed impossible in 12 months

In a Walter Isaacson interview on The Builders, published 27 October 2026, Anthropic's chief executive also describes his reaction to GPT-2 and his fortnightly talks to the company.

  • On GPT-2, Dario Amodei said he was "super freaked out by it": scaling it up a factor of a million would make it "better than humans at everything" (13:20).
  • He said: "But we just announced that, you know, we're going to be developing drugs for, you know, neglected rare diseases" (14:36).
  • Hard things would be easy in six months, he said, and "12 months from now, the things that seemed impossible, the models are going to be doing them regularly" (14:59).
Checked 28 Oct, 05:41 · The Builders (YouTube) · 2 sources
Read the story

You're all caught up

That's today's news: three stories.

Start again See the whole week

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 sealed inspection cabinet with a row of test bays; its green bays stand for the five attack chains the unreleased model found. Illustration made with AI for ai notis Safety · 28 OCT Anthropic's unreleased research model found five attack chains, since patched or addressed Anthropic · 1 source · 28 Oct, 05:40 Illustration: A bench scale with a rack of sealed units beside it; its green price dial stands for the 40% government discount. Illustration made with AI for ai notis Compute · 28 OCT Anthropic offers governments 40% off Claude through a new GSA agreement Anthropic · 1 source · 28 Oct, 05:40 Dario & Daniela Amodei: The Rise of Superintelligent AI, the End of Human Control & Humanity's Fate: a video from The Builders on YouTube narrowedVideo: The Builders on YouTube Research · 28 OCT Dario Amodei says models will regularly do what seemed impossible in 12 months The Builders (YouTube) · 2 sources · 28 Oct, 05:41

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

Today's stories 3 STORIES

Today's stories

Wed 28 Oct · 3 STORIES
  1. Safety · 28 OCT

    Anthropic's unreleased research model found five attack chains, since patched or addressed

    Illustration: A sealed inspection cabinet with a row of test bays; its green bays stand for the five attack chains the unreleased model found.Illustration made with AI for ai notis

    Anthropic does not name the internal model. Its post says the Chrome, Safari, Android, Windows and Linux flaws were fixed or addressed, and does not say any was exploited.

    Checked 28 Oct, 05:40 · Anthropic · 1 sourceRead

  2. Compute · 28 OCT

    Anthropic offers governments 40% off Claude through a new GSA agreement

    Illustration: A bench scale with a rack of sealed units beside it; its green price dial stands for the 40% government discount.Illustration made with AI for ai notis

    The discount covers Claude Enterprise, Claude for Government and the Claude API, has no minimum spend, and reaches state, local, tribal and territorial governments, Anthropic says.

    Checked 28 Oct, 05:40 · Anthropic · 1 sourceRead

  3. Research · 28 OCT · narrowed

    Dario Amodei says models will regularly do what seemed impossible in 12 months

    Dario & Daniela Amodei: The Rise of Superintelligent AI, the End of Human Control & Humanity's Fate: a video from The Builders on YouTubeVideo: The Builders on YouTube

    In a Walter Isaacson interview on The Builders, published 27 October 2026, Anthropic's chief executive also describes his reaction to GPT-2 and his fortnightly talks to the company.

    Checked 28 Oct, 05:41 · The Builders (YouTube) · 2 sourcesRead

Safety 28 OCT

Anthropic's unreleased research model found five attack chains, since patched or addressed

Anthropic does not name the internal model. Its post says the Chrome, Safari, Android, Windows and Linux flaws were fixed or addressed, and does not say any was exploited.

Anthropic · 1 source·checked 28 Oct, 05:40
AnthropicChromeFrontier Red TeamLinuxVulnerability disclosureWindows
Illustration: A sealed inspection cabinet with a row of test bays; its green bays stand for the five attack chains the unreleased model found.Illustration made with AI for ai notis
Compute

Anthropic offers governments 40% off Claude through a new GSA agreement

The discount covers Claude Enterprise, Claude for Government and the Claude API, has no minimum spend, and reaches state, local, tribal and territorial governments, Anthropic says.

28 OCT · Anthropic · 1 source
toned down Research

Dario Amodei says models will regularly do what seemed impossible in 12 months

In a Walter Isaacson interview on The Builders, published 27 October 2026, Anthropic's chief executive also describes his reaction to GPT-2 and his fortnightly talks to the company.

28 OCT · The Builders (YouTube) · 2 sources
In brief 2 more today

Today's explainer was dropped and does not run.

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