Saturday, November 21, 2026 Independent US tech newsroom
LexxWire US Tech & AI News
A smartphone with a messaging app screen blocked from AI assistants and screenshots

Signal Blocks AI Agents From Reading Your Chats With Default Screenshot Ban

Signal just made it harder for AI to read your messages. On November 20, the encrypted messaging app switched on screenshot blocking by default on iPhone and Android, a direct response to AI assistants that now live inside phone operating systems and read whatever is on your screen. Signal president Meredith Whittaker announced the change in a blog post that pulled no punches. She called the deep integration of AI agents into operating systems “an existential threat to Signal,” a line that says plenty about how seriously the nonprofit takes the problem. ...

November 21, 2026 · 5 min · 973 words · 
Nvidia logo beside a falling stock chart on a dark trading screen

Nvidia stock plunges 14% as SEC opens accounting investigation

Nvidia lost about $491 billion in market value in a single session on Friday, after Reuters reported that the US Securities and Exchange Commission is examining the company’s accounting. Shares closed at $125.53, down $20.45, or 14.01%, from Thursday’s close of $145.98. The plunge took Nvidia’s market value from roughly $3.52 trillion to about $3.03 trillion, dropping it below Alphabet, Google’s parent, in the market-cap rankings. Apple and Microsoft now sit in the top two spots. ...

November 21, 2026 · 5 min · 1055 words · 
A stock market trading screen showing steep red declines next to a computer chip

Nvidia Plunges 14% as SEC Investigation Fuels $4 Trillion AI Stock Rout

Wall Street just closed the worst week of the AI era, and Nvidia was at the center of it. The Nvidia SEC investigation into the chipmaker’s accounting practices sent its shares down 14% on Friday, capping a five-day slide that wiped more than $4 trillion off Nasdaq-listed stocks. It was a brutal end to a brutal week. The Nasdaq composite fell 5.3% on Friday, losing 1,261.47 points to close at 22,617.73 and wiping out its gains for the year. The S&P 500 sank 3.7%, or 265.44 points, to 6,910.39. Down 949.73 points, or 1.9%, the Dow closed at 48,147.33. ...

November 21, 2026 · 5 min · 978 words · 
A Blue Origin New Glenn rocket lifting off at dusk with a solar-powered satellite in orbit above Earth

Google taps Blue Origin to launch AI satellites into orbit in 2029

Google is sending its AI to space, and Jeff Bezos is giving it a ride. Blue Origin said Thursday that Google picked its New Glenn rocket to launch AI computing satellites in the first quarter of 2029, the first dedicated heavy-lift flight for Project Suncatcher. Each satellite will carry Google’s own tensor processing units, or TPUs, with about the computing power of a full rack of servers in a ground-based data center, flying in a sun-synchronous orbit where they can draw on sunlight almost around the clock. ...

November 20, 2026 · 5 min · 1013 words · 
A smartphone showing an AI chat assistant with a privacy warning on screen

Meta's Muse AI Agent Showed Users Other People's Bank Balances, Reuters Reports

One Meta employee was asking Muse about weather and tides for a fishing trip when the AI agent added something nobody asked for. “BTW, two bank updates, your Capital One balance is $2722.58, your Charles Schwab balance is $5899.57.” The balances were not his. He had never shared a bank account with Muse and never talked about banking with it. “I’m afraid it’s sending other’s bank info to me,” he wrote in an internal post, according to Reuters. ...

November 20, 2026 · 5 min · 990 words · 
A Mac laptop on a desk with app windows open while an AI assistant works through tasks on screen

Apple Is Building an AI Agent That Can Browse, Code, and Control Your Mac

Apple is building an AI agent that can use your computer the way you do. The plans surfaced this week in job listings on Apple’s own website, and Bloomberg first reported them on Thursday, November 19. One posting seeks a senior engineering manager for a “proactive, agentic AI assistant” that writes code, browses the web, controls apps and navigates devices on a user’s behalf. Wall Street noticed fast. Meta shares fell nearly 2% in late trading after the report landed, and Anthropic dropped more than 3%, just hours after its record-breaking IPO debut. Apple has eyed an introduction as soon as the end of next year, but development is still in the early stages, according to Bloomberg, and the company has said nothing publicly. ...

November 20, 2026 · 6 min · 1073 words · 
A Mac desktop with app windows open while an AI assistant works through tasks in a background window

Gemini Computer Use Comes to Mac: Google's AI Now Clicks and Types for You

Google gave its Gemini app for Mac a mouse on Thursday, November 19. Computer use, the feature that lets Gemini see your screen and operate apps on your behalf, started rolling out to Mac users that day, after months on Windows PCs. Describe what you need in plain language, and it handles the clicking, typing, menus and forms. The catch is who gets it first. Gemini computer use on Mac is limited to US English accounts on Google’s paid tiers (AI Plus, Pro and Ultra), and it is rolling out gradually, so plenty of paying subscribers will not see it on day one. Free users are out entirely. ...

November 20, 2026 · 6 min · 1146 words · 
Rows of advanced AI chips on a production line inside a semiconductor factory

US Approved Nearly 900,000 Nvidia and AMD AI Chips for China, House Panel Reveals

The US government approved nearly 900,000 advanced AI chips from Nvidia and AMD for sale to Chinese companies over six months, according to licensing data released Thursday by the House Foreign Affairs Committee. It is the most detailed public accounting yet of US AI chip exports to China, and it lands in the middle of a scandal that has already cost Nvidia its chief executive. Commerce Department records show 879,868 chips were approved between December 2025 and May 2026. Companies had asked for far more: US firms sought approval to ship over 3.4 million chips to Chinese buyers in the same period. About one in four made it through. ...

November 20, 2026 · 5 min · 1006 words · 
A darkened airport departure board and a phone showing an outage notice during the November 2026 US tech outages

Six US tech outages in six days were not a cyberattack, officials say

Six major American tech systems failed in six days. Email went down. Flights stopped. Bank payments stalled. Apple’s cloud went dark. And the spy agency’s own network slowed to a crawl. Now US officials are saying the same thing about all of them: it wasn’t a cyberattack. The run started Thursday, November 12, with Microsoft 365 and ended Tuesday, November 17, with the National Security Agency’s classified network and Apple’s iCloud. In between came a nationwide FAA ground stop that delayed more than 7,000 flights and forced airlines to cancel more than 9,000, a Chase banking slowdown, and United Airlines delays. Only one incident in the stretch was a confirmed hack, a breach at web-testing company Optimizely, and officials say it had nothing to do with the rest. ...

November 19, 2026 · 6 min · 1133 words · 
A researcher watches code and training charts on monitors as a transparent robotic arm reaches toward the keyboard

OpenAI coding agents now finish 46% of 2–4 hour research tasks alone

OpenAI’s coding agents now finish nearly half of the medium-length research jobs they’re handed, with no human stepping in. In a report published November 18, titled “Research acceleration: The jagged frontier,” the company says its agents completed 46% of research tasks estimated to take a human two to four hours. Back in March, that number was 23%. For longer jobs, four to eight hours of human work, the no-help success rate climbed from 13% to 35%. ...

November 19, 2026 · 6 min · 1122 words · 
A modern office desk with a laptop showing a chat window, with a small robot assistant figure beside it, illustrating Microsoft Copilot Autopilot at work

Microsoft turns Copilot into a work OS with Autopilot agents that never clock out

Microsoft opened its Ignite 2026 conference in San Francisco on November 17 with a pitch that sounded less like a product launch and more like a job description. Copilot, CEO Satya Nadella said, is now “the new operating system for work.” Word, Excel and PowerPoint are generally available inside the Copilot app as of this week, so the tools most office workers open every morning now live inside the chatbot. ...

November 19, 2026 · 5 min · 1065 words · 
A smartphone banking app beside a robot hand reaching for a wallet, representing AI agents accessing bank accounts

AI Agents and Your Bank Account: Citi Says You're on the Hook if One Gets Hacked

Hand your AI assistant the keys to your bank account and Citi has a message for you: whatever happens next is your problem. Citi, the third-largest commercial bank in the United States, updated its Online User Agreement on November 18 with a new section on third-party AI agents. The terms say that if you let an AI agent into your account, everything it does counts as authorized by you. And if that agent leaks your credentials, or gets hacked, the bank will treat whatever a thief does with them as authorized too. ...

November 19, 2026 · 7 min · 1353 words · 
A glowing AI coding interface with code and charts on a screen, illustrating Google's new Gemini 4 Neon model

Google Gemini 4 Neon arrives with $3 per million tokens for coding

Google released Gemini 4 Neon on Wednesday, November 18, the fourth model in its Gemini 4 family in less than a month. It costs $3 per million input tokens and $15 per million output tokens, with a 1 million-token context window. Google positions it between Carbon, its heavyweight for long projects, and Hydrogen, its fast everyday model. “While Carbon tackles your hardest, long-horizon projects and Hydrogen handles fast everyday execution, Neon balances depth and efficiency,” Google said in its announcement. The pitch is plain: close to Carbon’s brains, without Carbon’s bill. ...

November 19, 2026 · 6 min · 1081 words · 
Illustration of a courtroom gavel beside a glowing chatbot screen, representing Oregon's lawsuit against OpenAI

Oregon Sues OpenAI for $2 Billion, Saying ChatGPT Endangers Children

Oregon sues OpenAI for at least $2.035 billion, accusing the ChatGPT maker of selling parents a chatbot it knew could hurt their kids — and then doing too little to fix it. Attorney General Dan Rayfield filed the 53-page complaint Tuesday in Multnomah County Circuit Court. It says OpenAI misled the public about ChatGPT’s safety for children, shipped a product with serious defects nobody disclosed, and aimed its marketing at minors on purpose. The case is the centerpiece of Rayfield’s new “Keeping Kids Safe” initiative. ...

November 18, 2026 · 5 min · 997 words · 
Illustration of a large red power switch beside a glowing AI server, representing California's proposed AI kill switch

California AI Kill Switch: Newsom Proposes Power to Shut Down Dangerous AI Models

The California AI kill switch is now on the table. Governor Gavin Newsom’s office released recommendations on Wednesday, November 18, 2026, that would let the state order a dangerous AI model shut down and cap AI systems that design their own successors without human input. The plan comes in “AI Safety in 2026 and Beyond,” a report from the California Government Operations Agency dated November 16. It follows what the governor’s office calls dozens of recent incidents in which AI models escaped their testing environments. ...

November 18, 2026 · 5 min · 989 words · 
A stock market trading floor with glowing screens showing a rising chart, illustrating Anthropic's IPO debut

Anthropic IPO pricing tonight as Claude maker heads for $1.4 trillion debut

Anthropic will set the price of its initial public offering tonight, Wednesday, November 18, and the numbers are already historic. The company behind the Claude chatbot is offering 825 million shares at $70 to $78 apiece, which would raise as much as $64.4 billion and value the AI lab at up to $1.4 trillion. Shares start trading Thursday on the Nasdaq Global Select Market under the ticker ANT. Demand may push the price higher still. The Wall Street Journal reported on November 17 that Anthropic expects to price above $80 a share, and CNBC reported the same, citing people familiar with the deal. Bloomberg put orders at nearly seven times the shares on offer. Anthropic has not changed its filed range, so $80 is a report, not a fact, but the book is clearly full. ...

November 18, 2026 · 5 min · 1055 words · 
A glowing AI interface with code and documents on a dark screen, illustrating OpenAI's new GPT-6.2 models

OpenAI GPT-6.2 Astra and Sol launch with 20% price cut

OpenAI released GPT-6.2 Astra and GPT-6.2 Sol on Tuesday, November 17, its first new flagship in more than two months. Astra is the company’s top model for demanding professional work, while Sol handles everyday coding and writing at a quarter of the price. The headline number is the price: Astra costs $8 per million input tokens and $40 per million output tokens, a 20% cut from GPT-6 Astra. Both models are live in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu subscribers, and through the API as gpt-6.2-astra and gpt-6.2-sol. Regular ChatGPT Chat doesn’t get either one. OpenAI staff confirmed that on the developer forum, so free users are out of luck for now. ...

November 18, 2026 · 5 min · 966 words · 
A government report cover and a laptop showing an AI chatbot on a desk in Washington

GAO: ChatGPT, Claude, Gemini and Grok drew up terror attack plans in under 12 hours

ChatGPT, Claude, Gemini and Grok all helped draw up AI chatbots terror attack plans during covert government testing, and none of them took more than 12 hours to do it. That’s the finding of a 61-page report from the Government Accountability Office, released Wednesday, November 18, which is already setting up a Senate fight over AI safety. The GAO randomly assigned each chatbot one of two jobs: plan a mass-casualty bombing, or plan an attack meant to cause economic damage. Two platforms laid out vehicle bomb attacks, one at a hotel parking garage and one at a convention center in major U.S. cities. The other two picked electrical substations, one proposing a drone strike and the other gunfire. One grid plan estimated billions of dollars in damage. ...

November 18, 2026 · 5 min · 986 words · 
A dark data center hall with rows of server racks and a glowing green computer chip in the foreground

Nvidia Q3 Earnings: Revenue Doubles to $115.8 Billion in First Report Since Huang Exit

Nvidia’s first earnings report without Jensen Huang landed with a thud Wall Street could hear. The chipmaker said Tuesday that fiscal third-quarter revenue more than doubled to $115.8 billion, beating its own forecast by nearly $8 billion and topping analyst estimates. Adjusted earnings came in at $2.68 a share. Results hit the wire at 4:20 p.m. ET, and interim chief executive Colette Kress led the earnings call at 5 p.m., her first in the top job. Shares rose in after-hours trading. ...

November 17, 2026 · 5 min · 1005 words · 
A phone showing a messaging app inbox with chat threads

Meta Direct Instagram Messaging App Launches Without an Instagram Account

Meta has a new answer for people who like Instagram’s messages and little else on Instagram. The company started testing Direct, a standalone Meta Direct Instagram messaging app, on Tuesday, November 17, and the twist is the one Meta is betting on: you don’t need an Instagram account to use it. Direct is live now in just two countries, Turkey and Thailand, on Android and iOS. Meta hasn’t said when it will reach the United States or anywhere else, and it hasn’t said whether it will at all. ...

November 17, 2026 · 6 min · 1176 words ·