November 19, 2026 Understand what changes. Find what to use.

Updated today at 7:08 AM

Your daily AI edition

What deserves your attention.

What's new, the context, and what you can do with it.

1000 stories · Updated today at 7:08 AM

Thursday, October 22

SecurityNews2 sources

Meta re-enables Muse sharing after flaw made links easy to find

Meta re-enabled Muse sharing after Claroc researchers showed that public links to conversations could be easily discovered. The company says it fixed the issue, changed URLs for shared artifacts, and had paused the feature while reviewing the report.

Full summary

AgentsNews1 source

1Password launches tool for AI agents to log in without seeing passwords

1Password launched 1Password for Agents, which lets AI agents access personal accounts without receiving users’ credentials. Users are alerted and must confirm each login attempt.

Full summary

ResearchNews1 source

Study says AI could estimate cancer mortality risk from a face photo

A study described by Xataka evaluates FAHR-Face, an AI that uses facial photos to estimate cancer patients’ biological age and mortality risk. In the group classified as highest risk, mortality was 3.23 times higher than in the lowest-risk group.

Full summary

CreationNews2 sources

Canva launches Canva Craft, its first image generation model

Canva launched Canva Craft, its first in-house image generation model, globally for all users. The system plans a composition before generating images of up to 2K resolution, with support for more than 30 aspect ratios.

Full summary

APIsNews1 source

OpenAI makes Decisions API widely available amid decision-model competition

OpenAI has made the Decisions API available to all customers, entering competition with TypeSafe’s Jev model. OpenAI’s Nikunj Handa said the API was optimized to reduce latency and cost, while adding voice and image capabilities.

Full summary

CodeNews1 source

Anthropic hiring data show junior hiring falling from 27% to 9%

Anthropic hiring data analyzed by Xataka show that technical junior roles fell from 27% to 9% of new hires between early 2025 and June 2026. The company credits AI agents with productivity gains.

Full summary

ResearchNews3 sources

Microsoft says MAI-DxO outperformed doctors in diagnostic tests

Microsoft says its MAI-DxO system correctly diagnosed 83% of complex medical cases, versus 36% for doctors with reference materials, in a study published in Science. The technology still needs hospital validation and has no commercial release date.

Full summary

SecurityNews1 source

Axios: US banks restrict AI agents accessing customer accounts

According to Axios, US customers are already allowing AI agents to access their bank accounts and act on their behalf. Banks including Morgan Stanley and JPMorgan are restricting this access over concerns about errors, unauthorized transfers and difficulty identifying who is making each decision.

Full summary

SecurityNews1 source

Modder says they jailbroke the Switch 2 using an old game and AI

Sammy Hajhamid, known as “pepsipu,” says they jailbroke the Nintendo Switch 2 with an exploit that uses SmileBASIC 4. They showed a modified Mario Kart World, and say OpenAI’s Codex helped with the process.

Full summary

SecurityNews1 source

Microsoft says China-aligned hackers spied on about 2,250 organizations

Microsoft researchers say China-aligned hackers used AI agents after breaching ConnectWise’s cloud backup service. Microsoft says the group attempted to access about 2,250 organizations at nearly 3,000 sign-in attempts per minute.

Full summary

RegulationNews1 source

Meta removes MC Cabelinho video about Flávio Bolsonaro, and singer reposts it

Meta removed an Instagram video by MC Cabelinho and the Revelando portal that addressed an allegation involving Flávio Bolsonaro. The company told g1 the removal followed electoral rules from Brazil’s TSE for content produced or altered by artificial intelligence; the singer later reposted the video.

Full summary

CreationNews1 source

Travis Knight says generative AI is ‘antithetical to everything that we are’ at Laika

Travis Knight, Wildwood’s director and Laika’s CEO, told The Verge that generative AI is “antithetical to everything that we are” at the studio. He distinguishes it from technology used by artists and says human stories must come from people.

Full summary

ResearchNews1 source

Hugging Face releases OlmoEarth v1.3 with global satellite embeddings

Hugging Face released OlmoEarth v1.3 and precomputed annual embeddings for the entire planet, along with a browser tool for exploring maps. The organization says the model ranked first in 20 of 27 evaluations in its 128-dimensional version.

Full summary

BusinessNews2 sources

Uber and OpenAI announce ChatGPT Work for more than 30,000 employees

Uber and OpenAI announced that ChatGPT Work will be made available to the company’s entire global workforce of more than 30,000 people. The expansion follows a pilot with 1,000 employees.

Full summary

RegulationNews1 source

Federal disclosures show record lobbying quarters for OpenAI, Anthropic and Nvidia

Federal disclosures show that OpenAI, Anthropic and Nvidia each had their highest quarterly lobbying spending between July and September. Nvidia spent $2.2 million, followed by Anthropic at $2.1 million and OpenAI at $1.5 million.

Full summary

AgentsNews1 source

Spotify says it wants to be the music service AI agents choose

Co-CEO Gustav Söderström says Spotify wants AI agents to find music, podcasts and audiobooks through its service. The company says it reached 788 million monthly active users and nearly €5 billion in revenue in the third quarter.

Full summary

MarketNews1 source

NVIDIA launches Minerva for end-to-end observability in AI factories

NVIDIA launched Minerva, an AIOps platform that combines infrastructure and AI workload telemetry to detect anomalies, predict failures and help teams investigate incidents. NVIDIA says IREN is already using the platform, while other companies plan to adopt it.

Full summary

AgentsNews1 source

Microsoft says AI agents help TK Elevator service technicians

Microsoft says TK Elevator uses AI agents to brief technicians before maintenance visits and record their spoken reports. The tool combines equipment data, service histories and digital manuals to reduce the time spent searching for information.

Full summary

BusinessNews1 source

Nirva announces $33 million raise for an AI pendant

Nirva announced a total raise of $33 million to develop a $249 AI pendant that, according to the company, filters surrounding voices to listen only to the wearer. Orders begin in November, with deliveries to early users planned by the end of the year.

Full summary

BusinessNews1 source

Bloomberg: DeepSeek reportedly raises $15 billion after U.S. restrictions

DeepSeek reportedly raised about $15 billion, according to Bloomberg, cited by TechCrunch, two days after the U.S. ordered federal civilian agencies to remove its technology over national security concerns. Potential investors include CATL and Tencent.

Full summary

One story, many sources

Coverage of the same story gathered in one item, with a link to each original.

Summary and interpretation apart

What the sources say stays in the summary; editorial context is labeled separately.

Always with the source

Every story links to the original articles, and the full content belongs to them.