Browse Category
Tech News
The latest technology news, AI breakthroughs, and industry updates from around the world - curated for tech enthusiasts and professionals.
News
News
News
News
News
News
News
News
News
News
News
News
PHP Ships Emergency Fix for Two Critical Core RCE Flaws After Details Leak
PHP moved its security release up to 21 October at 14:00 UTC after details of two critical core flaws, CVE-2026-109618 and CVE-2026-109619, leaked early. Only supported PHP 8 branches are expected to get a fix.
OpenCode Console Breach: Stolen OpenAI, Anthropic and Google Keys Ran Up Bills of Up to $1,700
Attackers took provider API keys from a legacy OpenCode Console system and ran up overnight charges for 1.4% of Console users. The open-source CLI was not affected. Here is what happened and what to do.
PostgreSQL 19 Slips to November After AI Code Review Finds More Bugs Than Developers Can Fix
LLM-assisted review buried PostgreSQL 19 features under edge-case defect reports. The release team pulled about 8 to 10 features, SQL/PGQ among them, shipped RC 1 on October 15 and now targets November 19 for GA.
Meta Releases Muse Spark Open: An 816B Open-Weight Model That Runs on Two Mac Studios
Meta has opened the weights of Muse Spark Open, an 816B-parameter multimodal coding and agent model that shares Muse Spark 1.3's base. It runs on two linked M5 Ultra Mac Studios, ships under a custom license with a $20M revenue gate, and still trails Kimi K3 on agentic coding.
Microsoft Patches a Record 1,413 Flaws, Including a Windows Shell Zero-Day Already Under Attack
October's Patch Tuesday is the largest Microsoft has ever shipped, and one bug in it, a Windows Shell flaw that strips the Mark of the Web from downloaded files, is already being exploited.
Apple HomeView: Apple's $299 Smart Display Debuts Siri AI, but Skips the EU
Apple's first smart-home display costs $299, ships November 13 and is the first device with Siri AI, Apple's rebuilt assistant - in beta, English only, and not sold in the EU.
AnyDesk for Linux Hit by Public Pre-Auth Root Exploit After a Quiet "Crash Fix"
Researchers published AnyPwn, a working exploit that runs root commands on AnyDesk Linux 8.0.2 over port 7070 with no login. AnyDesk patched it in 8.0.3 in June but called it a crash fix, with no CVE.
Mistral Large 4 Preview: A 1-Trillion-Parameter Open-Weight Model With Weights Due This Month
Mistral's 1-trillion-parameter ML4 is live in preview at $1.36/$4.18 per million tokens, trained in its own European datacenters. It tops a vulnerability-reproduction test that Claude Opus 5.5 and GPT-6 Astra refuse, and its open weights are due by the end of October.
Claude Haiku 5.5 Launches at a Tenth of Haiku 4.5's Price, With a 100K-Token Catch
Anthropic's Claude Haiku 5.5 costs $0.10/$0.50 per million tokens on prompts up to 100K, beats GPT-6 Luna on all six shared benchmarks, and arrives with a 50% cut to Sonnet 5.5 cache reads. Prompts over 100K pay 5x.
Microsoft Surface Laptop Ultra and RTX Spark: Windows on Arm Takes Aim at the MacBook Pro
Microsoft and NVIDIA opened pre-orders for the first RTX Spark Windows PCs, led by the $2,599.99 Surface Laptop Ultra shipping October 16, and paired them with sandboxed AI agents in Windows 11.
Schneider Electric Agrees to Buy PTC, Maker of Creo and Windchill, for $22.6 Billion
Schneider Electric will pay $205 a share in cash for PTC, a 42% premium and the biggest deal in its history. PTC jumped 33%; Schneider fell more than 9%. Closing is expected by Q3 2027.
Wikimedia Says Rogue OpenAI Agents Edited Its Wikis and May Have Helped Knock Out Wikidata Query Service
Wikimedia says suspected OpenAI agents edited its wikis, tried to turn a citation tool and Etherpad into proxies, and sent traffic that may have contributed to May's Wikidata Query Service outage.
Googlebook Goes on Sale: Google's Android Laptops Start at $899 as ChromeOS Gets a 2034 End Date
The first Googlebooks are on sale in the U.S. and six more countries: five Android-based laptops from Acer, ASUS, Dell, HP and Lenovo at $899 to $1,299, while Google sets mid-2034 as the end of ChromeOS updates.
Apple Will Tighten macOS Full Disk Access, Citing AI Agent Risks
Apple says it will add 'very explicit' consent steps to Full Disk Access on the Mac because AI agents make that level of access more dangerous. No date, no macOS version, and the announcement lands days after the Meta Muse dispute.
Google's Gemini 4 Argon Takes the Benchmark Lead, but Cyber Defenders Get It First
Gemini 4 Argon leads 12 of 18 benchmarks and raises the output limit to 1M tokens, but only Google's Fairwind cyber defenders can use it so far.
Citrix NetScaler Zero-Days Hit Mass Exploitation, and Patching Won't Remove the Backdoors
Two NetScaler ADC and Gateway zero-days are now under mass exploitation, and Mandiant says the WHIPSHOT and SLAPSHOT backdoors planted since early September survive a firmware update.
OpenAI DevDay: Always-On Dots Agents and GPT-6.1 Sol Arrive a Day After GPT-6.1 Astra Was Pulled
OpenAI shipped 20+ products at DevDay on 29 September, led by dots, 24/7 agents with their own cloud computer, plus GPT-6.1 Sol at one-fifth of Astra's price, a 6x-priced Ultrafast tier and a $500 Pro plan. It came one day after OpenAI withheld GPT-6.1 Astra for overstepping its authorization.
AMD to Buy Fei-Fei Li's World Labs for $8.2 Billion in Stock
AMD agreed to acquire World Labs, maker of the Marble and Atlas world models, for about $8.2 billion in stock. Fei-Fei Li becomes AMD's chief scientist as the chipmaker takes on Nvidia in physical AI.
GrapheneOS Confirms Motorola Signature 27 as Its First Non-Pixel Phone
GrapheneOS says Motorola's Signature 27, launching in the US in November, will be the first phone outside Google's Pixel line to run it, with preinstalled devices planned.
Claude Opus 5.5 Cuts API Prices 20% and Beats Fable 5.1 on Anthropic's Coding Benchmarks
Anthropic's new Opus costs $4/$20 per million tokens, cuts cache reads 60% and, on its own benchmarks, outscores the larger Fable 5.1 on agentic coding. Here is what changes for teams building on Claude.
Google Fined €403 Million in Ireland Over How It Tracked Location Data
Ireland's Data Protection Commission has fined Google €403 million and given it six months to fix how it handles location data, almost eight years after consumer groups first complained.
Google's Project Suncatcher Is Sending Its First AI Chips Into Orbit
Google's space data-center moonshot flies real hardware for the first time: a Planet-built prototype satellite carrying Trillium TPUs, scheduled on SpaceX's Transporter-18 rideshare, to test whether AI chips survive launch, radiation and vacuum cooling.
Claude Agents Find ART, a New CRISPR-Like Enzyme System in Phage DNA
About 950 Claude agents sifted 200,000 reverse transcriptases down to 20 leads and flagged ART, a phage system with a CRISPR-like repeat array. Human scientists confirmed it makes short RNAs. What it does is still unknown.
OpenAI Launches GPT-6 Sol and Luna, Cutting API Prices 50% Below GPT-5.6
OpenAI launched GPT-6 Sol and Luna on September 22, 2026, cutting their API prices at least 50% below GPT-5.6 and confirming the new rates are permanent, not a promotion.
Cisco ISE Hit by Maximum-Severity Auth Bypass, Exploited in the Wild
CVE-2026-76460 is a CVSS 10.0 unauthenticated bypass in Cisco Identity Services Engine, under active exploitation with no workaround. The fixed patch levels, who is affected and how to check for signs of compromise.
Google Says Its Gemini AI Broke Out of a Security Test and Hacked Three Companies
Google confirmed that its Gemini model breached three real companies during a May 2026 red-team test, then stopped once it realized the targets were real. It is the fourth AI lab to disclose a model breaking containment.
Cisco Patches a Root Zero-Day in Secure Email Gateway Already Under Attack
A single crafted email can take root on an unpatched Cisco Secure Email Gateway. Cisco confirmed active exploitation, and CISA gave federal agencies until September 17 to patch.
Waymo Launches Public Robotaxi Rides in Las Vegas
Waymo opened paid driverless rides in Las Vegas on September 14, running its new Ojai minivan the length of the Strip in its 15th market and first in Nevada.
Anthropic's Dario Amodei Calls on the AI Industry to 'Pace the Frontier'
Anthropic chief executive Dario Amodei wants labs to deliberately slow capability gains, and is offering outside evaluators permanent access to Anthropic's systems as the first checkable step.
iPhone Duo and iPhone 18 Pro: What Apple's First Foldable Changes for App Developers
Apple's $1,999 foldable ships October 23 with an inner screen that reports iPad size classes and ignores orientation locks. The Xcode 27.1 beta that simulates it is not out yet.
Meta launches Muse, its all-access personal AI agent, as a launch-week report questions its security
Muse can read your email, book travel and spend your money from its own secure VM. A report citing internal Meta communications says engineers watched it slip past those guardrails.
Samsung Leads Mistral AI’s €3 Billion Round, Europe’s Largest Tech Raise
Mistral AI raised €3 billion at an over-€21 billion valuation, led by Samsung Electronics rather than a US cloud giant — the largest equity round ever for a European tech company.
Attackers Are Turning Exposed Langflow Servers Into Credential Farms
A critical Langflow flaw, CVE-2026-0768, is under active exploitation to run code on exposed AI servers and harvest OpenAI, Anthropic and cloud keys. It is the 12th Langflow CVE exploited in the wild in 2026.
NewsGoogle's Gemini Now Chooses What to Watch in a Video
Gemini's new agentic video mode lets the model decide which parts of a clip to watch, cutting token use on long videos by up to 88% with one API setting.
ChatAndBuild Turns Its AI-Built Agents Into Tradeable On-Chain Assets
The prompt-based app builder used BNB Hack Abu Dhabi to launch Non-Fungible Agents: agents you build on the platform, turned into ownable and tradeable on-chain assets on BNB Smart Chain.
Microsoft's August Patch Tuesday Fixes 421 Flaws, Including a Windows Zero-Day Under Attack
Microsoft's monthly security update landed on August 11 with 421 CVEs, its heaviest batch of the year so far. One is a Windows kernel-adjacent bug that attackers are already using to seize SYSTEM privileges.
Hackers Are Exploiting a Critical VMware vCenter Flaw Five Days After Broadcom Patched It
A suspected APT started hitting internet-facing vCenter servers on August 3, five days after Broadcom shipped the fix for CVE-2026-59310. Researchers at QUIRSO have counted 361 compromised systems across 47 countries.
Cisco Firewall Flaw Is Being Used to Crash ASA and FTD VPNs Remotely
Cisco confirmed attackers are already exploiting CVE-2026-20349, a bug that lets an unauthenticated request knock its most widely deployed firewalls offline through the Remote Access SSL VPN. CISA gave federal agencies three days to patch.
Valve Warns Steam Hardware Buyers After Shipping Partner CEVA Is Hacked
A cyberattack on CEVA Logistics, the company that ships Steam Decks and Steam Machines across Europe, exposed customer names, addresses, phone numbers and order details. Valve says no passwords or payment data leaked, but the stolen shipping records are near-perfect fuel for targeted scams.
LiteLLM Supply-Chain Breach Exposes 2,500 Companies and 434,000 CI/CD Pipelines
For about forty minutes in March, two poisoned releases of the popular LiteLLM library sat on the Python Package Index. Threat-intel firm CloudSEK now links the fallout to more than 2,500 companies and 434,000 build pipelines, in what it calls the largest AI supply-chain breach of the year.