AI Pulse

Morning digest โ€” 2026-11-21

Read today's executive briefing

โ†’

Live feed

Product3h ago

Signal restricts mobile screenshots as it confronts โ€˜existential threatโ€™ of AI surveillance

Signal is rolling out screenshot restrictions on its Android and iOS apps as part of efforts to counter what it calls an 'existential threat' from AI-driven surveillance.

Read article โ†—DetailsTechCrunch AI
Product7h ago

Here are the top AI agents that can live in your text messages

A roundup article lists notable AI agents accessible via text messaging, covering general assistants as well as specialized agents for families, travel, and work.

Read article โ†—DetailsTechCrunch AI
Policy4h ago

The UN fails to agree on rules regulating โ€˜killer robotsโ€™

UN diplomats meeting in Geneva failed to reach agreement on regulations for autonomous weapons ('killer robots'), only agreeing to continue discussing options after a largely unproductive week.

Read article โ†—DetailsThe Verge AI
News3h ago

A tale of two Nvidias

Nvidia's latest earnings report presented a strong financial picture, contrasting with vocal criticism from some tech executives calling for CEO Jensen Huang's removal.

Read article โ†—DetailsTechCrunch AI
News8h ago

Is selling an AI version of your face to this microdrama startup the future of Hollywood?

An actor has licensed an AI-generated version of his face to a microdrama startup, reflecting a trend already popular in China where AI likenesses are used in short-form drama content.

Read article โ†—DetailsTechCrunch AI
News9h ago

Can Anthropic afford to put safety first?

The Verge's AI podcast discusses Anthropic's recent IPO and explores whether the company can maintain its safety-first approach as a public company.

Read article โ†—DetailsThe Verge AI
News10h ago

Chess.comโ€™s next move is a new hub for classic games, from mahjong to sudoku

Chess.com is expanding beyond chess by launching a hub featuring 10 additional classic games including mahjong, sudoku, backgammon, checkers, go, hearts, shogi, solitaire, xiangqi, and a Scrabble-like word game.

Read article โ†—DetailsThe Verge AI
Research1d ago

Researcher plants 2m canaries into the AI supply chain. A privacy disaster surfaced.

A researcher embedded two million unique 'canary' secrets into prompts across the AI supply chain and found that supposedly private data leaked and persisted, revealing a privacy disaster in how AI systems handle data.

Read article โ†—DetailsArs Technica AI
Researchโ–ฒ Trending1d ago

Combatting adversarial distillation

OpenAI detailed its layered strategy to combat adversarial distillation, combining training safeguards, detection, and enforcement to prevent unauthorized model capability extraction.

Read article โ†—DetailsOpenAI Blog
Product1d ago

Google rolls out new search homepage with AI chat history and no search button

Google has redesigned its search homepage to feature AI chat history and remove the traditional search button, making AI tools more prominent and accessible.

Read article โ†—DetailsArs Technica AI
Product23h ago

New in AWS Quick: More capable flows, an updated web experience, and agents that work together

AWS announced updates to its AWS Quick AI workspace, including improved business-process flows, a redesigned web app, multi-agent collaboration, and new data source integrations like Meta ads.

Read article โ†—DetailsAWS Machine Learning Blog
Policy1d ago

Meta reconsiders 2030 net-zero goal amid record emissions from AI data centers

Meta is reportedly re-evaluating its 2030 net-zero sustainability commitment as emissions from its AI data centers reach record levels.

Read article โ†—DetailsArs Technica AI
Product1d ago

Introducing managed cluster extensions for observability and user management in Amazon SageMaker HyperPod Slurm

AWS announced managed cluster extensions for observability and user management in Amazon SageMaker HyperPod Slurm, enabling declarative, API- and console-driven configuration of software on Slurm cluster nodes.

Read article โ†—DetailsAWS Machine Learning Blog
Product1d ago

Introducing Amazon SageMaker HyperPod Global Inference Gateway: One endpoint across Regions and clusters

AWS launched SageMaker HyperPod Global Inference Gateway, a managed endpoint that routes inference requests across clusters and regions with automatic failover and IAM/SigV4 authentication.

Read article โ†—DetailsAWS Machine Learning Blog
Product1d ago

Announcing Amazon Textract Insights: Verifiable generative AI for documents

AWS announced general availability of Amazon Textract Insights, a generative AI feature that extracts structured data from documents using natural language field descriptions, with confidence scores and page-level grounding.

Read article โ†—DetailsAWS Machine Learning Blog
Product1d ago

Bringing 327 Genspark Tools to DeepSeek Harness: Introducing dsh-genspark and the Trick for Porting to Muse

A blog post introduces 'dsh-genspark,' a plugin that ports 327 Genspark tools into the DeepSeek Harness framework, including a method for adapting them to a system called Muse.

Read article โ†—DetailsDeepSeek (news)
Research1d ago

The most misunderstood graph in AI just got a major update

AI evaluation nonprofit METR released an updated report showing some AI capabilities continue to improve exponentially, though models lag in tasks requiring judgment compared to coding tasks.

Read article โ†—DetailsMIT Technology Review AI
Funding1d ago

Daniel Ek backs AI researcherโ€™s fusion power startup. Hereโ€™s what theyโ€™re building.

Spotify founder Daniel Ek is backing Europa Fusion, a fusion power startup founded by an AI researcher.

Read article โ†—DetailsTechCrunch AI
Research1d ago

AI claimed a solution to the 167 year-old Riemann Hypothesis โ€” and a human did, too

Mathematicians are debating competing claimed solutions to the Riemann Hypothesis, one produced with AI assistance and one by a human mathematician.

Read article โ†—DetailsTechCrunch AI
Research1d ago

Quoting Ravi Vakil

A blog post quotes mathematician Ravi Vakil discussing Alessandro Masullo's proposed proof of the Generalized Riemann Hypothesis, which Brian Conrad has been reviewing for months.

Read article โ†—DetailsSimon Willison's Weblog
Policy1d ago

Trump threatens โ€˜tremendous taxโ€™ for moving AI data centers outside the US

President Trump threatened to impose a steep tax on companies that build AI data centers outside the US, amid growing state and local restrictions on new data center construction domestically.

Read article โ†—DetailsThe Verge AI
Product1d ago

Signal is blocking Gemini and Siri AI from reading your screen

Signal is adding features on Android and iOS to block screenshots and prevent AI assistants like Gemini and Siri from reading on-screen message content.

Read article โ†—DetailsThe Verge AI
Research1d ago

C-13 How to Verify Abnormal Operations Caused by External Documents: Lessons from a DeepSeek API Chatbot Case Study

An article using a DeepSeek API chatbot case study to walk through verifying abnormal operations caused by indirect prompt injection via external documents.

Read article โ†—DetailsDeepSeek (news)
Product1d ago

Six AI-powered gadgets that could make your home office smarter

A roundup article highlights six AI-powered gadgetsโ€”such as notetakers and webcamsโ€”intended to enhance home office productivity.

Read article โ†—DetailsTechCrunch AI
Research1d ago

Anthropic mathematician claims to have solved prestigious Millennium Prize problem with Claude

Anthropic mathematician Levent Alpรถge claims to have proved the Riemann Hypothesis, a Millennium Prize problem, with assistance from Claude.

Read article โ†—DetailsThe Verge AI
Research1d ago

Quoting Levent Alpรถge

A quote from mathematician Levent Alpรถge stating he believes he has found a proof of the Riemann Hypothesis (and the Generalized Riemann Hypothesis) while working with Claude.

Read article โ†—DetailsSimon Willison's Weblog
Research1d ago

CertiK and Columbia University Researchers Formally Verify Production AI Kernels From DeepSeek and FlashAttention

CertiK and Columbia University researchers formally verified production AI kernels from DeepSeek and FlashAttention, aiming to ensure correctness and security of these widely used AI computation components.

Read article โ†—DetailsDeepSeek (news)
News1d ago

Security event: Prompt leakage for requests in June and July of 2026

Together AI disclosed a security incident in which a subset of prompts submitted to its serverless inference service during June and July 2026 were leaked, and the company is notifying affected customers.

Read article โ†—DetailsTogether AI Blog
Model release2d ago

Welcome Inkling 2 Mini, Cutting-edge Performance for your Local Setup

Hugging Face announced the release of Inkling 2 Mini, a new open-source model optimized for local/on-device performance.

Read article โ†—DetailsHugging Face Blog
News1d ago

Chips can survive the space environment. Can their financing?

The article examines challenges in financing data center hardware deployed in space, noting lenders are wary of collateral risk when assets like chips are launched into orbit by companies such as SpaceX.

Read article โ†—DetailsTechCrunch AI