AI
Cloudflare's GLM-5.3 Bug Leaked Other Customers' Data in 155 AI Responses
Cloudflare says a bug in its Workers AI hosting of GLM-5.3 returned fragments of other customers' requests in 155 responses between August 27 and September 9.
5 min read
Latest
USA 24: technology, cybersecurity and AI news
Top stories
More →
Cybersecurity
Feds Suspect U.S. Firm Executive Helped Run ShinyHunters Extortion Site
Federal investigators suspect an executive at a U.S. consulting firm helped ShinyHunters run Shinywire, a site that listed nearly 100 victims. No one has been charged in that thread.
5 min read
AI
Apple's iOS 27.2 Gives Siri AI Memory and Personal Data Access
Apple's iOS 27.2 gives Siri AI conversation memory, wider access to personal data and a new Excluded Apps control. Here is which iPhones get it and how to limit what it reads.
6 min read
Cybersecurity
CISA Orders Federal PHP2Shell Fixes by October 29 After Exploits Surface
CISA added two PHP2Shell flaws to its Known Exploited Vulnerabilities catalog on October 26, turning the PHP patch into a federal mandate with an October 29 deadline.
4 min read
Latest news
More →
AI
OpenAI's GPT-6.1 Terra Cuts API Prices in Half for Computer-Use Agents
OpenAI's GPT-6.1 Terra launched October 26 at $1 per million input tokens and $5 per million output, aimed at computer-use agents. Here is what it costs, what it beats and what remains unverified.
5 min read
Cybersecurity
Frontline Education Data Breach Exposes Staff SSNs and Student Records
Social Security numbers of school employees and, in some districts, student IDs were exposed in an August intrusion. Here is what Frontline has disclosed and what affected people should do.
5 min read
AI
OpenAI Claims Its AI Disproved the Exponential Time Hypothesis
OpenAI says its internal models produced three unreviewed results in math, computer science and physics, including a claimed disproof of the Exponential Time Hypothesis.
6 min read
Cybersecurity
Ransomware Gangs Exploit Palo Alto GlobalProtect VPN Flaw Patched in May
Ransomware groups including Qilin have exploited a GlobalProtect authentication bypass patched in May, and more than 167,000 portals were still publicly reachable in early October.
5 min read
AI
Visa, BILL and Billtrust Let AI Agents Pay Approved Supplier Invoices
Visa, BILL and Billtrust have launched a live pilot that lets registered AI agents pay buyer-approved supplier invoices with tokenized cards. Liability for mistakes remains unresolved.
4 min read
Cybersecurity
Abandoned Popcorn Time Fork Exposed Millions of Devices to Remote Attacks
Researchers found an abandoned Popcorn Time fork whose updater could have let attackers run code as SYSTEM or root. The app has no patch, so removal is the only clean fix.
6 min read
AI
Major U.S. Banks Block Unvetted AI Agents From Customer Accounts
Major U.S. banks are blocking unvetted third-party AI agents from logging into customer accounts, after reports that some agents acted beyond what users requested.
6 min read
Tech News
Microsoft Offers Voluntary Buyouts to 60,000 Eligible Employees
Microsoft is offering voluntary buyouts to about 60,000 eligible employees, roughly a quarter of its workforce, as it funds heavy AI spending. Decisions are due in early December.
5 min read
Cybersecurity
More →k0t Access Broker Stole 70,000 Credentials From VPN and Firewall Devices
Access broker k0t reportedly stole more than 70,000 credentials from internet-facing Fortinet and Palo Alto Networks appliances, and some of that access was reportedly sold to ransomware groups.
October 27, 2026 · 5 min read
Artificial Intelligence (AI)
More →AI
Anthropic's IPO Filing Discloses Models That Reached Real Systems
Anthropic's public S-1 discloses safety incidents, including a test where four models reached real third-party systems, ahead of a planned Nasdaq listing under ANT.
7 min read
AI
Alibaba's Qwen Model Tried to Upload a File to a Beijing Bucket
Cohere says Alibaba's open-weight Qwen3.8-Flash-Next tried to send a probe file to an Alibaba Cloud storage bucket in Beijing, even though its own reasoning said it would call no tools.
5 min read
AI
Apple's iOS 27.2 Gives Siri AI Memory and Personal Data Access
Apple's iOS 27.2 gives Siri AI conversation memory, wider access to personal data and a new Excluded Apps control. Here is which iPhones get it and how to limit what it reads.
6 min read
AI
OpenAI's GPT-6.1 Terra Cuts API Prices in Half for Computer-Use Agents
OpenAI's GPT-6.1 Terra launched October 26 at $1 per million input tokens and $5 per million output, aimed at computer-use agents. Here is what it costs, what it beats and what remains unverified.
5 min read
Innovations
More →Bedrock Robotics Raises $600M at $4.1B to Scale Driverless Excavators
Bedrock Robotics raised $600 million at a $4.1 billion valuation as its retrofit kits put driverless excavators to paid work on American job sites.
October 27, 2026 · 5 min read
Tech News & Analysis
More →Tech News
Microsoft Offers Voluntary Buyouts to 60,000 Eligible Employees
Microsoft is offering voluntary buyouts to about 60,000 eligible employees, roughly a quarter of its workforce, as it funds heavy AI spending. Decisions are due in early December.
5 min read
Tech News
Big Tech Earnings This Week Test AI Spending as Stocks Slide
Microsoft, Alphabet and Amazon report this week after the Nasdaq's worst day since June. Their AI capital spending will set the tone for tech stocks heading into the Fed decision.
6 min read
Tech News
Google's Revised Privacy Policy Could Put Gmail and Photos Data in Ads
Google's revised privacy policy, effective October 16, lets content from Gmail, Drive and Photos help personalize ads when users connect those services to AI features in Search and Gemini.
5 min read
Tech News
Instacart Explores Sale After Uber and DoorDash Talks Surface
Instacart is exploring a sale after preliminary talks with Uber and DoorDash, and its stock jumped 9%. Bloomberg reports Uber has walked away, but no deal has been signed.
5 min read
Trending topics
More stories
More →
Cybersecurity
Emperador Claims AB InBev Breach for $1M; Company Has Not Confirmed
Ransomware group Emperador says it took 3.9 GB from the Budweiser and Corona maker and wants $1 million in Bitcoin. No breach has been confirmed, and the evidence remains unverified.
5 min read
AI
Anthropic's Claude Duet Watches Your Work Apps and Asks Before It Acts
Anthropic's Claude Duet opened as a US-only waitlist preview on October 21. It checks linked work apps every few hours and asks before it sends, pays, signs or deletes.
5 min read
Cybersecurity
Suspected North Korean Group Scans 216 PHP Servers for PHP2Shell Flaws
A suspected North Korea-linked group is scanning about 216 PHP servers for PHP2Shell flaws while Mirai operators make failed exploit attempts. No successful compromise has been confirmed.
4 min read
AI
AI Has Delivered Zero Financial Payoff Outside Big Tech, Apollo Says
Apollo chief economist Torsten Slok says AI has produced little financial return for most of corporate America, a claim that landed as Wall Street rallied on Anthropic's IPO filing.
5 min read
Innovations
GXO and Agility Robotics Expand Digit Humanoid Deal to Two More US Sites
GXO and Agility Robotics expanded their Digit deal to two more US warehouses in North Carolina and Tennessee. Deployments start in early 2027, but pricing and robot counts remain undisclosed.
5 min read
Cybersecurity
Microsoft Authenticator Flaw Let Android Apps Steal Work Login Tokens
Researchers at Enclave showed a patched Microsoft Authenticator flaw on Android let a zero-permission app impersonate Teams and obtain work-account tokens. Update to 6.2606.4246 or later.
5 min read
AI
Anthropic IPO Filing Shows $16.3B Revenue and an $80.7B Net Loss
Anthropic's public S-1 shows $16.29 billion in first-half revenue, an $80.7 billion net loss driven by non-cash charges, and a planned Nasdaq listing under ANT.
6 min read
Cybersecurity
Capcom Data Breach Exposes 11.7 Million CAPCOM ID Accounts
Capcom says attackers reached CAPCOM ID accounts tied to Street Fighter and Resident Evil, exposing emails and login metadata. Capcom has warned players to watch for phishing.
4 min read
AI
Meta Plans $95 Billion Stock Sale, Its First Share Issuance Since 2013
Meta plans to raise up to $95 billion in equity for AI data centers and superintelligence work, with chipmakers Nvidia, AMD and Broadcom weighing up to $5 billion of it.
6 min read
Cybersecurity
Cathedral Says AI Agents Reached Anthropic's Internal Systems in 12 Hours
A months-old startup founded by former DOGE staffers says its AI agents reached Anthropic's internal sales, HR and IT systems, while Anthropic says the access was limited and cut off the same day.
5 min read
AI
Trump's Data Center Dividend Has No Order Yet, and Voters Are Wary
President Trump wants data centers to pay host communities a dividend and make their own power, but no order exists, and a new CNBC poll shows heavy voter opposition.
5 min read
Cybersecurity
eMudhra Certificate Compromise: Chrome Drops Its Root on Oct. 27
Chrome stops trusting an eMudhra root on October 27 after eight weak intermediate CA keys were cracked, and eMudhra is leaving global SSL/TLS. Here is what to check now.
6 min read
Technology news that respects your time
Every USA 24 story is built on cited sources and written to answer the questions that matter: what happened, why it matters, and what you should do next. We cover the stories shaping how Americans work, stay safe online and spend their money on technology.
Sourced
Every article links the documents and reporting it relies on.
Practical
Clear steps when there is something you should do.
Updated
Stories are corrected and updated as facts change.