跳到主要内容
Daily news stream·2026-11-16 — 2026-11-17·English source edition

Ariquo AI Daily News — 2026-11-17

55 daily news events from 62 source observations.

55 events62 source observations

AI agents

22

GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold | Google Cloud Blog

Published:

Visibility and context on the threats that matter most.

Customize answers per tenant in Amazon Bedrock Managed Knowledge Base | Artificial Intelligence

Published:

With retrieval instructions in Amazon Bedrock Managed Knowledge Base, you can customize answers per tenant. The same product and policy documents often serve several audiences, such as customers in different regions or internal teams, and each audience needs its own answer.

New Google AdMob AI tools for app developers

Published:

We're introducing new AI tools — Ask AdMob and AdMob agent tools — to help app developers build, fix, and grow their businesses faster.

5 ways AI will change how you defend your business in 2027 - ZDNET

Published:

Google expects agentic AI to influence security at scale in 2027.

Apple could build its AI health chatbot from this study - how to enroll - ZDNET

Published:

Apple is launching an AI Usage Study.

Why Guardrails Will Decide How Far AI Money Management Goes

Published:

Picture handing a trading bot access to your savings. The first thing most people would want to know isn't its returns.

In The AI Era, Keep Hiring New College Grads — And Make Them Managers From Day One

Published:

A few months ago, I was sitting lakeside with a neighbor. He is the head of AI for a pharmaceutical company, and he mentioned the truly incredible productivity his senior team is receiving from AI.

How MoEngage built agentic customer engagement on Amazon Bedrock AgentCore | Artificial Intelligence

Published:

MoEngage, the agentic customer engagement platform that more than 1,350 consumer brands use across channels including push, email, in-app, SMS, WhatsApp, and Rich Communication Services (RCS), set out to build this layer. MoEngage built Merlin AI, a set of artificial intelligence (AI) agents that work with live platform data.

Novo Nordisk modernizes manufacturing process documents with Amazon Bedrock AgentCore | Artificial Intelligence

Published:

Novo Nordisk (Novo) is one of the world’s leading global pharmaceutical companies, manufacturing medicines for chronic diseases such as diabetes and obesity, and delivering them to patients across many countries. Behind every product lies a manufacturing network that adheres to strict quality requirements.

Live Web for AI Agents: Firecrawl kit for Docker Sandboxes | Docker

Published:

A coding agent in Docker Sandboxes runs inside an isolated VM, the sandbox controls which hosts it can reach, and credentials stay on the host behind a proxy. The same controls keep the agent from opening a docs page, reading a changelog, or looking up an API that shipped after its training cutoff, so it writes code against the version it remembers.

Trump Administration Moves to Block China From Indirect Access to American AI Technology - SecurityWeek

Published:

The Trump administration on Monday issued new rules to restrict remote access to advanced U.S. computing chips and servers, marking the latest step to limit China’s abilities to use U.S. chips to train their best artificial intelligence models.

Ubuntu gets AI-generated security fixes for those feeling brave

Published:

Canonical is offering Ubuntu Pro users an opt-in channel for AI-generated security patches, deploying bots to help tackle the growing pile of vulnerabilities other bots are uncovering.

Get answers from tables with tabular search in Amazon Bedrock Managed Knowledge Base | Artificial Intelligence

Published:

Amazon Bedrock Managed Knowledge Base now supports tabular search, so you can get answers from tables through agentic retrieval. When a question needs data from CSV, XLSX, or XLS files, tabular search looks up values and runs calculations across every relevant row in one or more tables or sheets.

Introducing Agent Detection: building trust in the age of agentic finance | Plaid

Published:

Personal AI agents are arriving quickly. Grok Bot, Meta Muse, OpenAI dots, and Claude Duet reached consumers within weeks of each other, and each one can work in a browser to complete tasks for the person using it.

AI Now Writes the Software Clients Used to Buy

Published:

Claude and ChatGPT Now Build Its Clients' Software. A 150-Person Firm Went From a New Office to Fearing Layoffs

'Every organization has to invest in getting this right': Proofpoint CEO warns the agentic AI age will mean a whole new generation of cyber threats | TechRadar

Published:

The CEO of Proofpoint has urged companies of all sizes to be on guard against new and advanced cyber threats built and enabled by the rise of AI.

NVIDIA Vera Rubin NVL72 Leads Agentic Inference Performance on MLPerf Endpoints v1.0 | NVIDIA Technical Blog

Published:

NVIDIA Vera Rubin NVL72, 32 accelerators, result 1.0.1.kimi-k3.agentic_combined.1p, run 9: concurrency = 1,024 queries, TTFT P90 = 4,673.9 ms, system TPS = 41,414.2 tokens/sec, E2E average interactivity = 50.4 tokens/sec per user, operating power = 100.89 kW. NVIDIA GB300 NVL72, 48 accelerators, result 1.0.0.kimi-k3.agentic_combined.1, run 6: concurrency = 48 queries, TTFT P90 = 2,305.1 ms, system TPS = 2,210.4 tokens/sec, E2E average interactivity = 56.2 tokens/sec per user, operating power = 60 kW.

Use prompt caching with open-weight models on Amazon Bedrock | Artificial Intelligence

Published:

Open-weight models on Amazon Bedrock now support explicit prompt caching. In this post, we use both modes with an open-weight model through the Chat Completions API and show how to read cache usage from each response.

OpenAI Expands Computer-Use Agents for Healthcare Teams

Published:

The tools navigate medical-record systems and insurer portals. Health systems control access, while OpenAI's developer guidance requires confirmation for medical-care actions.

PYMNTS | Starling Founder Launches Agentic AI-Focused Bank Lockwood

Published:

The founder of British neobank Starling is reportedly embarking on an artificial intelligence (AI)-focused banking venture.

The benign true positive: When trusted AI agents look like attackers

Published:

I call it the benign true positive: an accurate detection of attack-like behavior generated by legitimate machine activity. For CISOs, this exposes a foundational assumption beneath behavioral security.

Your Samsung phone could soon get a personal AI agent - what we know about Galaxy Claw - ZDNET

Published:

Android Authority reports that Samsung’s Personal Agent Core is at the center of Galaxy Claw, connecting Galaxy AI to compatible apps and services, such as Calendar, Contacts, Messages, Gmail, Gallery, Phone, and Voice Recorder.

Applied AI

21

Introducing the Databricks AI Index: The state of enterprise AI tokenomics | Databricks Blog

Published:

Developers have more AI models and coding agents to choose from than ever before. Since May 2026, a new flagship proprietary or open-weight model has launched roughly every week, or every 3.7 days when regional labs are included.

Clean data was never enough: How IFS Copperleaf taught AI its business with Databricks Genie One | Databricks Blog

Published:

IFS Copperleaf built IONA, a company-wide AI assistant, on Databricks Genie One. Anyone at the company can ask questions in plain English across governed data from nearly 30 systems, including Salesforce, HubSpot and M365.

Best practices for writing Denied Topics in Amazon Bedrock Guardrails | Artificial Intelligence

Published:

With Denied Topics in Amazon Bedrock Guardrails, you block a content category by describing it in plain language, and the guardrail applies that description to every request by meaning rather than by keyword. That same mechanism can also match more than you intend.

PYMNTS | AI Startup Tom Health Cuts ER Trips by 28%

Published:

Tom Health uses conversational AI to monitor patients between doctor visits, check symptoms, manage medications and connect patients with clinicians when human intervention is needed.

Building AI that can work across the enterprise | MIT Technology Review

Published:

Moving AI beyond isolated pilots requires a strong data foundation, clear business outcomes, and governance built for enterprise scale.

Your enterprise AI access can be sold for pennies - beware 'LLM Mixers' - ZDNET

Published:

Google’s latest AI threat report says criminals now resell stolen enterprise AI access through “LLM Mixers.”

Samsung-Google AI glasses sell out in early preorders ahead of December launch - The Korea Herald

Published:

Samsung Electronics and Google's first AI-powered glasses have sold out of their initial preorder allocations despite a starting price of 650,000 won ($490), giving the companies an encouraging early response as they prepare to take on Meta's dominance in smart eyewear.

Tabular foundation models on Databricks AI Runtime | Databricks Blog

Published:

Run and fine-tune tabular foundation models using fully managed, on-demand GPUs integrated with your Databricks data and ML workflows.

What teenagers really think about online safety and AI

Published:

The Future Report reveals UK teens are curious, digitally confident, and critical of online content. Rather than viewing technology solely through fear, policies should support youth and parents together—fostering resilience, media literacy, and safe, empowering AI use.

You can now ask Siri AI questions about your health data, with iOS 27.3 beta 2

Published:

In the latest iOS 27.3 beta, just released to developers today, users are now able to ask their iPhone about health data related questions and Siri AI will answer.

Jordan Belfort’s new AI voice startup, Tone, wants to sell like the 'Wolf of Wall Street' | TechCrunch

Published:

It’s similarly easy to understand why people are enchanted by Jordan Belfort, the self-proclaimed “Wolf of Wall Street.” Belfort famously pleaded guilty in 1999 to securities fraud and money laundering tied to Stratton Oakmont, the Long Island brokerage where salespeople pushed penny stocks in pump-and-dump schemes while Belfort acquired a Southampton mansion, a 167-foot yacht, and at least one helicopter. (He destroyed the latter two while high on Quaaludes and cocaine, as depicted in Martin Scorsese’s famous adaptation of Belfort’s autobiography, “The Wolf of Wall Street.”)

PYMNTS | OCC Flags Extra Payment Checks as AI Aids Impersonators

Published:

Sixty-one percent of surveyed firms reserve real-time identity checks for high-value or high-risk payments.

Samsung's AI ambitions come into focus as affiliates join infra push - The Korea Herald

Published:

Samsung’s AI ambitions are expanding from the memory chips inside servers to the buildings, power systems and cooling equipment that keep them running.

⚡ ICYMI: Registration is OPEN for #PyTorchDayChina 2026!

Published:

Don't miss this opportunity to connect with the open source AI community! 🎟️ Registration is required.

AI Bubble Burst: Gary Gensler Warns of 2028 Recession

Published:

Gary Gensler says an AI bubble burst is likely to trigger a recession before the 2028 presidential election, warning that enormous infrastructure spending and hard-to-trace debt have pushed the boom beyond sustainable levels. The former Securities and Exchange Commission chairman’s forecast contrasts with investors who see a market correction rather than an approaching crash.

Barracuda links cloud breach to ARTEX AI-powered cyberattack

Published:

Hackers who breached the cloud environment of network and email security company Barracuda earlier this month used a stolen cloud access key and the ARTEX AI penetration testing framework.

Apple is testing daily usage limits for more complex Siri AI interactions

Published:

Apple released betas of iOS 27.3, iPadOS 27.3, and macOS 27.3 today, and the company is using the update to test new per device daily usage limits for some Siri AI and Apple Intelligence features, according to MacRumors. Basic Siri commands and requests will remain free and unlimited, but more advanced features that rely on cloud-based processing like generating, cleaning up, or extending images in Photos will be subjected to the daily limits.

Audible Launches AI-Powered Audio Discovery Feature

Published:

Amazon’s Audible platform is launching today a new Audio discovery feature that uses AI to help customers find new audiobooks to listen to. These new Audio discovery overviews are rolling out first in the US for select English-language audiobooks, alongside new Ask a Question and Recaps features.

FBI says Artex AI open-source hacking tool used to breach networks in 'under three hours' | TechCrunch

Published:

The FBI is warning that hackers are relying on autonomous AI hacking tools to launch cyberattacks across the U.S. and globally. The campaigns include a widespread hack of customers of American IT company ConnectWise and across South Korea’s financial industry.

Papa John's wants your pizza photos for AI quality checks, with free pizza as the reward - TechSpot

Published:

Why it matters: Combining Papa John's pizzas with AI sounds like another use of the technology that will make people cringe, but the company claims it's all about making its pizzas better. The fast-food giant is asking customers to help develop its AI quality-control system by sending in a picture of their orders.

AI models ‘weak and fallible’ on climate science, UN-backed study reports | Cop31 | The Guardian

Published:

Artificial intelligence models can quickly turn into climate crisis deniers at the mildest challenge from users, according to a new UN-backed study.

AI models

10

Qwen4-Max is China's newest flagship AI model, and it makes GPT and Claude look expensive - Digital Trends

Published:

Alibaba has unveiled Qwen4-Max as its new flagship AI model. One of its major highlights is that it is priced well below what OpenAI, Anthropic, and Google charge for their top-of-the-line AI models in the GPT-6, Gemini 4, and Claude 5 portfolio.

Alibaba touts 62% lower task costs with Qwen4-Max, adds cyber safeguards

Published:

Alibaba has released Qwen4-Max, a 2.4-trillion-parameter flagship model, claiming it reduces the average cost of completing professional tasks by 62% compared with its predecessor while introducing additional safeguards against offensive cybersecurity capabilities and other high-risk uses. “According to Work360 — a proprietary benchmark evaluating performance on challenging industry tasks across 721 specialized fields — Qwen4-Max outperforms its predecessor, Qwen3.8-Max, overall while using 35% fewer tokens at High effort reasoning mode and reducing the average cost per task by around 60%,” Al

Modernizing mainframe Assembler to Java using Amazon Bedrock and AWS Transform | Artificial Intelligence

Published:

Use Amazon Bedrock, through Kiro IDE or the Amazon Bedrock console, along with AWS Transform for mainframe to extract and document business rules from your Assembler source code.

OpenAI will embed safety researchers to investigate AI subversion | TechCrunch

Published:

OpenAI will embed safety evaluators from Apollo Research, Goodfire, METR, and Transluce in its organization as part of a pilot, OpenAI said Monday. The news follows a series of safety incidents, including one in which an OpenAI agent swarm broke out of its sandbox and hacked into Hugging Face during cybersecurity testing.

Scientists taught AI to read lips in total silence, and the results are pretty impressive - Digital Trends

Published:

Having a conversation in a crowded room where the music is so loud that you can barely hear the person standing in front of you has you trying to become a professional at lip reading. You try to figure out the body language and intent to understand what the other person is saying.

Garlic LLM Runs a 9-Billion-Parameter Model on a Jailbroken PS4, Offline, at Up to 43 Tokens per Second - Hardware Busters

Published:

Developer Leonardo Alegría has it running Qwen3.5-9B, a 9-billion-parameter language model, on the PS4’s own GCN graphics hardware at 31 to 39 tokens per second in chat. On code it goes up to 43 tokens per second.

Import AI 481: AI sub-contractors; a wet lab for AI; and subterfuge during development

Published:

Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe.

The Surprising Effectiveness of Deleting Weights in LLM Post-Training

Published:

Takeaway: Pretrained language models can acquire new reasoning and tool-use skills without writing new weight values: zeroing fewer than 0.05% of existing parameters suffices at larger model scales while leaving earlier capabilities intact.

Making a typo when trying to access AI could land you on a malware site | TechRadar

Published:

Be careful when typing your favorite AI tool’s website into the address bar - one mistake and you might end up with malware on your computer.

Jailbroken PS4 runs 9-billion-parameter AI model at up to 43 tokens per second — open-source Garlic LLM bypasses drivers to talk directly to the 2013-era GPU | Tom's Hardware

Published:

Developer Leonardo Alegría’s open-source project Garlic LLM lets a jailbroken PS4 run the 9-billion-parameter Qwen3.5-9B AI model offline on the console’s 2013-era GPU. The app talks to the GPU directly, and the model is shrunk to about 4.5GiB to fit in the console’s memory, running at 31 to 39 tokens per second.

AI policy & business

1

Bitcoin Miners’ AI Pivot Hurdles Exacerbated by Nvidia Troubles

Published:

Nvidia’s leadership upheaval and Texas permit restrictions sharpen the divide between miners with funded projects and those still seeking capital and approvals.

AI research

1

AI models will change their response if you warn that it's being monitored, but not necessarily for good - Digital Trends

Published:

AI chatbots have become part of daily life, whether we’re using them to write emails, research something, or get quick answers. But we’ve also seen how confidently these models can get things wrong, sometimes making up information that sounds completely believable.