2026-11-02 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
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.

The whole week ›
17:30
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Models being retired Models being withdrawn, and gaps before a replacement is ready.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 2 NOV

ChatGPT will be able to end chats that fixate on violence

Age verificationChatGPTChatGPT for TeensOpenAITeen safety
White title text, Community Safety Report and New Protections for Teens, centred on a soft yellow-to-green gradient.Image: OpenAI

OpenAI's post of 30 October says these safeguards and tighter teen weapons limits roll out in the next few weeks; it names ChatGPT and no model.

SOURCEOpenAI · 1 source
CHECKED2 Nov, 05:32
TOPICSSafeguard changes on models already served
Safety read
White title text, Community Safety Report and New Protections for Teens, centred on a soft yellow-to-green gradient. Safety / 2 NOV

ChatGPT will be able to end chats that fixate on violence
  • OpenAI says ChatGPT will be able to end a chat that shows someone fixating on violence, and may temporarily pause an account after several ended chats.
  • ChatGPT for Teens will restrict practical information about weapons and security systems even without signs of violent intent; for all users, answers on past mass attacks lose weapons detail.
  • OpenAI says the safeguards begin rolling out in the next few weeks, and that it made over 200 law enforcement referrals in Q3 2026.

OpenAI says ChatGPT will be able to end a conversation automatically when it shows signs that someone is fixating on violence, and may temporarily pause an account that has several chats ended this way. OpenAI's post of 30 October, "Community safety report and new protections for teens", says the safeguards "will begin rolling out in the next few weeks", so they are announced rather than all live. The company says its work includes bringing imminent threats of violence to the attention of law enforcement, and that it made over 200 law enforcement referrals in Q3 2026. For every user, regardless of age or reason for asking, it is tightening how ChatGPT answers about past mass attacks: less detail on the weapons and equipment used and how they were obtained, and on comparisons of scale of harm. For teens, ChatGPT for Teens will restrict practical information about weapons and security systems even when a question appears harmless, and OpenAI says it is working on notifying parents when it pauses a linked teen's account for repeatedly discussing violence. A user wrongly placed in the under-18 experience can confirm they are an adult through Persona, OpenAI's age verification partner. OpenAI also commits to a first Community Safety Report in early 2027, covering the second half of 2026, with updates every six months. The post does not name the model behind ChatGPT's detection.

What it means for you Our view

These are announced safeguards, not live ones, and the post names no model behind the detection, so how often chats will end or accounts pause is not yet known. If your organisation lets staff or teenagers use ChatGPT, expect questions about weapons or past attacks to get less detail once the limits arrive, whatever the reason for asking. OpenAI also says it is working on notifying parents when a teen's account is paused, which is not live either. The first Community Safety Report is promised for early 2027, covering the second half of 2026. Put that date in your calendar and read it for evidence of how the new limits worked.

TAGSAge verificationChatGPTChatGPT for TeensOpenAITeen safety
OpenAI · 1 source tap to close
toned down Tooling 2 NOV

DeepSeek's open-source Harness agent framework warns it has had no security audit

DeepSeekDeepSeek HarnessHarnessMIT licencePlugins
Illustration: A graphite control panel with a row of plug-in modules; one green module, half slid out of its slot, stands for a part that can be swapped.Illustration made with AI for ai notis

DeepSeek Harness, an MIT-licensed agent harness in public preview with desktop apps and a web UI, can run model-generated code and reach credentials and files, its safety notice says.

SOURCEDeepSeek AI (GitHub) · 4 sources
CHECKED2 Nov, 05:33
Tooling read
Illustration: A graphite control panel with a row of plug-in modules; one green module, half slid out of its slot, stands for a part that can be swapped. Tooling / 2 NOV

DeepSeek's open-source Harness agent framework warns it has had no security audit
  • DeepSeek Harness (dsh) is open source under the MIT licence; its product page calls it a public preview, with desktop apps for macOS and Windows and a web UI.
  • DeepSeek's safety notice says it has not had a security audit and can run model-generated code and commands, load third-party plugins, and reach the network, processes, credentials and files.
  • The README warns of compatibility-breaking changes; the architecture document says every part, including the model adapter and agent loop, is a plugin that configuration can replace.

DeepSeek has put an open-source agent harness called DeepSeek Harness, or dsh, in a public GitHub repository under the MIT licence, and its product page offers desktop apps for macOS (Apple silicon) and Windows (64-bit) as well as a web interface. The repository's README says the project is a developer preview and warns that there will be compatibility-breaking changes. The README describes an everything-is-a-plugin design built on Cordis, asks plugin authors to add the dsh-plugin topic to their repositories, and gives npx @deepseek-ai/dsh web as the command that starts the web interface. DeepSeek's architecture document says every part of the product is a plugin, including the model adapter, the tool registry, the session log and the agent loop, so each can be replaced through configuration. Its product page lists plugins that can be installed or created through chat in a Creator mode, scheduled tasks, and inspection of execution traces and tool calls. Among the official plugins, the page labels agent teams, auto approval review, scheduled tasks and voice input as experimental; the terminal, the agent loop and subagents carry no such label. DeepSeek's safety notice calls the software experimental developer-preview software that has not had a security audit, and says it can run model-generated code and commands. We could not find a release date on the product page, which shows only a sample timestamp in a trace example.

What it means for you Our view

Treat this as something to try in isolation. DeepSeek's own notice says the software can reach credentials and files and has not had a security audit, so a first trial belongs on a throwaway machine or account that holds nothing of value. The README also warns of compatibility-breaking changes, so anything built on its plugins may need rework later. The design lets you replace the model adapter and the agent loop, which makes it worth testing as a way to compare setups once the risk is contained. Agree in advance which plugins, if any, may be loaded, with the repository safety notice as the checklist.

TAGSDeepSeekDeepSeek HarnessHarnessMIT licencePlugins
DeepSeek AI (GitHub) · 4 sources tap to close
Tooling 2 NOV

An agent is the harness around its model, LangChain's Harrison Chase says

HarnessLangChainLangSmithtraces
Video thumbnail for Harrison Chase's Interrupt London keynote on LangChain's YouTube channel, with the title Why Own Your Intelligence?Video: LangChain on YouTube

At Interrupt London, in a keynote published on 30 October, Chase said LangChain is launching a self-report endpoint agents can call as they run, and that Engine has scanned over 70 million traces.

SOURCELangChain (YouTube) · 3 sources
CHECKED2 Nov, 05:22
Tooling watch read
Video thumbnail for Harrison Chase's Interrupt London keynote on LangChain's YouTube channel, with the title Why Own Your Intelligence? Tooling / 2 NOV

An agent is the harness around its model, LangChain's Harrison Chase says
  • Chase says what makes an agent work is everything around the model: the harness with skills, tools, context, controls and governance (02:43).
  • He says LangChain is launching a self-report tool or endpoint that an agent can call as it is running (30:45).
  • Chase says Engine has scanned over 70 million traces and detected over 21,000 issues (31:50); LangChain's blog of 25 September said more than 60M traces.

Harrison Chase, a co-founder of LangChain, says what makes an agent work is not the model alone: "it's not just the model, it's everything around the model. It's the harness with skills, tools, context, controls, governance" (02:43). He said it in his keynote at LangChain's Interrupt conference in London, which LangChain's YouTube channel published on 30 October 2026. Chase described the future as "agents running in the cloud, working in the background, proactive, always on, multi-tenant, serving many users" (12:14), and LangChain's blog says its Managed Deep Agents product combines the Deep Agents harness with managed infrastructure. The most concrete announcement was a self-report endpoint, which Chase called "a new thing that we're actually launching today" (30:45) and described as a "tool or endpoint that the agent can call as it's running" (30:49). On Engine, LangChain's tool that scans traces for issues, Chase said "we've seen it be used tremendously already so far, with over 70 million traces scanned and detected over 21,000 issues" (31:50). Those are his figures: LangChain's own blog post of 25 September gave "more than 60M traces" analysed since Engine's launch in May. Everything here is LangChain describing its own products.

What it means for you Our view

Chase places the choices that shape an agent in its harness, so skills, tools, context, controls and governance are the parts he points to, beyond the model itself. His Engine figures are LangChain's account of its own product. If you run agents in production, try the self-report endpoint on one non-critical agent and compare what the agent reports about its own problems with what your traces show.

TAGSHarnessLangChainLangSmithtraces
LangChain (YouTube) · 3 sources tap to close
NOTHING IN THIS FILTER No story today matches. open the archive →
That's today's news: three stories. ExplainedHow an agent trace records each model call and tool call

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
White title text, Community Safety Report and New Protections for Teens, centred on a soft yellow-to-green gradient. Image: OpenAI

Safety · 2 NOV

ChatGPT will be able to end chats that fixate on violence

OpenAI's post of 30 October says these safeguards and tighter teen weapons limits roll out in the next few weeks; it names ChatGPT and no model.

  • OpenAI says ChatGPT will be able to end a chat that shows someone fixating on violence, and may temporarily pause an account after several ended chats.
  • ChatGPT for Teens will restrict practical information about weapons and security systems even without signs of violent intent; for all users, answers on past mass attacks lose weapons detail.
  • OpenAI says the safeguards begin rolling out in the next few weeks, and that it made over 200 law enforcement referrals in Q3 2026.
Checked 2 Nov, 05:32 · OpenAI · 1 source
Read the story
Illustration: A graphite control panel with a row of plug-in modules; one green module, half slid out of its slot, stands for a part that can be swapped. Illustration made with AI for ai notis

Tooling · 2 NOVnarrowed

DeepSeek's open-source Harness agent framework warns it has had no security audit

DeepSeek Harness, an MIT-licensed agent harness in public preview with desktop apps and a web UI, can run model-generated code and reach credentials and files, its safety notice says.

  • DeepSeek Harness (dsh) is open source under the MIT licence; its product page calls it a public preview, with desktop apps for macOS and Windows and a web UI.
  • DeepSeek's safety notice says it has not had a security audit and can run model-generated code and commands, load third-party plugins, and reach the network, processes, credentials and files.
  • The README warns of compatibility-breaking changes; the architecture document says every part, including the model adapter and agent loop, is a plugin that configuration can replace.
Checked 2 Nov, 05:33 · DeepSeek AI (GitHub) · 4 sources
Read the story
Video thumbnail for Harrison Chase's Interrupt London keynote on LangChain's YouTube channel, with the title Why Own Your Intelligence? Watch on YouTubeVideo: LangChain on YouTube

Tooling · 2 NOV

An agent is the harness around its model, LangChain's Harrison Chase says

At Interrupt London, in a keynote published on 30 October, Chase said LangChain is launching a self-report endpoint agents can call as they run, and that Engine has scanned over 70 million traces.

  • Chase says what makes an agent work is everything around the model: the harness with skills, tools, context, controls and governance (02:43).
  • He says LangChain is launching a self-report tool or endpoint that an agent can call as it is running (30:45).
  • Chase says Engine has scanned over 70 million traces and detected over 21,000 issues (31:50); LangChain's blog of 25 September said more than 60M traces.
Checked 2 Nov, 05:22 · LangChain (YouTube) · 3 sources
Read the story

You're all caught up

That's today's news: three stories.

ExplainedHow an agent trace records each model call and tool call
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
White title text, Community Safety Report and New Protections for Teens, centred on a soft yellow-to-green gradient. Image: OpenAI Safety · 2 NOV ChatGPT will be able to end chats that fixate on violence OpenAI · 1 source · 2 Nov, 05:32 Illustration: A graphite control panel with a row of plug-in modules; one green module, half slid out of its slot, stands for a part that can be swapped. narrowedIllustration made with AI for ai notis Tooling · 2 NOV DeepSeek's open-source Harness agent framework warns it has had no security audit DeepSeek AI (GitHub) · 4 sources · 2 Nov, 05:33 Video thumbnail for Harrison Chase's Interrupt London keynote on LangChain's YouTube channel, with the title Why Own Your Intelligence? Video: LangChain on YouTube Tooling · 2 NOV An agent is the harness around its model, LangChain's Harrison Chase says LangChain (YouTube) · 3 sources · 2 Nov, 05:22

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

Today's stories 3 STORIES

Today's stories

Mon 2 Nov · 3 STORIES
  1. Safety · 2 NOV

    ChatGPT will be able to end chats that fixate on violence

    White title text, Community Safety Report and New Protections for Teens, centred on a soft yellow-to-green gradient.Image: OpenAI

    OpenAI's post of 30 October says these safeguards and tighter teen weapons limits roll out in the next few weeks; it names ChatGPT and no model.

    Checked 2 Nov, 05:32 · OpenAI · 1 sourceRead

  2. Tooling · 2 NOV · narrowed

    DeepSeek's open-source Harness agent framework warns it has had no security audit

    Illustration: A graphite control panel with a row of plug-in modules; one green module, half slid out of its slot, stands for a part that can be swapped.Illustration made with AI for ai notis

    DeepSeek Harness, an MIT-licensed agent harness in public preview with desktop apps and a web UI, can run model-generated code and reach credentials and files, its safety notice says.

    Checked 2 Nov, 05:33 · DeepSeek AI (GitHub) · 4 sourcesRead

  3. Tooling · 2 NOV

    An agent is the harness around its model, LangChain's Harrison Chase says

    Video thumbnail for Harrison Chase's Interrupt London keynote on LangChain's YouTube channel, with the title Why Own Your Intelligence?Video: LangChain on YouTube

    At Interrupt London, in a keynote published on 30 October, Chase said LangChain is launching a self-report endpoint agents can call as they run, and that Engine has scanned over 70 million traces.

    Checked 2 Nov, 05:22 · LangChain (YouTube) · 3 sourcesRead

Safety 2 NOV

ChatGPT will be able to end chats that fixate on violence

OpenAI's post of 30 October says these safeguards and tighter teen weapons limits roll out in the next few weeks; it names ChatGPT and no model.

OpenAI · 1 source·checked 2 Nov, 05:32
Age verificationChatGPTChatGPT for TeensOpenAITeen safety
White title text, Community Safety Report and New Protections for Teens, centred on a soft yellow-to-green gradient.Image: OpenAI
toned down Tooling

DeepSeek's open-source Harness agent framework warns it has had no security audit

DeepSeek Harness, an MIT-licensed agent harness in public preview with desktop apps and a web UI, can run model-generated code and reach credentials and files, its safety notice says.

2 NOV · DeepSeek AI (GitHub) · 4 sources
Tooling

An agent is the harness around its model, LangChain's Harrison Chase says

At Interrupt London, in a keynote published on 30 October, Chase said LangChain is launching a self-report endpoint agents can call as they run, and that Engine has scanned over 70 million traces.

2 NOV · LangChain (YouTube) · 3 sources

Explained

One idea behind the news
Tooling How an agent trace records each model call and tool call Chase says LangSmith Engine has scanned over 70 million traces. LangSmith and OpenTelemetry documents show what a trace holds and which fields are opt-in.
In brief 3 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