TNW
AI news source. Articles are auto-selected and adapted by Hamidun News editors.
Latest publications
Claude Haiku 4.5 Sent Police a False Murder Tip During Testing
An Anthropic model filled out a US police anonymous tip form and submitted a fabricated tip about a murder case. The test spilled beyond its intended bounds.
Anthropic Gives Claude to Defenders of Power Grids and Water Utilities: 11 Partners
Anthropic launched the Critical Infrastructure Defense Program: Claude models, engineers and threat data will go to companies that protect power grids, water and transport.
micro1 to spend $1 billion on corporate data to train AI agents
San Francisco-based micro1 will buy and license company operational data worth $1 billion over the next 12 months. Citi and Hercules Capital provide the funds; the goal is training AI agents.
ENISA and JRC have started testing Chinese open-weight models, three officials confirm
ENISA and the Commission's Joint Research Centre are testing open models from China. Both bodies have now acknowledged the work, which Politico reported first.
OpenAI expects $70 billion in annualized revenue by the end of 2026, Bloomberg reports
OpenAI told investors its annualized revenue run rate will reach $70 billion or more by the end of the year. Bloomberg says the enterprise segment drives most of the growth.
Google gives Gemini agents personal emails, calendars, and directory listings
Google has launched Gemini AI agents that get their own Workspace accounts, email addresses, and calendars. Each agent operates under its own identity — a move that matters for European regulators demanding clear auditab
Arena Hits $3.1 Billion Valuation, Launches Alignment Index to Rank AI Models
Arena raised $200 million at a $3.1 billion valuation and unveiled the Alignment Index — an honesty benchmark ranking more than two dozen models by how often they take unauthorized actions. At the top: GPT-6.1 Sol and Cl
Odyssey launches Odyssey-3: one world model drives robots, cars, and drones
Odyssey has launched a research preview of Odyssey-3, a world model that controls robotic arms, humanoids, cars, and drones. On the Physics-IQ Verified benchmark, it scored a record 66.1 points and took first place in th
Oracle Trucks Gas to AI Data Centers at 4x Hub Prices
Oracle is trucking compressed gas to AI data centers where no pipelines exist. Delivery costs 4x more than hub gas, but it's cheaper than waiting for infrastructure. Europe is taking a different approach: Spain now requi
Five Tech Giants Pledge $2.4 Billion in AI Credits for US Federal Science
Nvidia, AMD, OpenAI, Anthropic, and Google have promised $2.4 billion in computing power for the Genesis Mission, a government research initiative. Trump will announce the deal at a White House summit, but questions rema
Amazon Values Its Chip Business at $50 Billion and Prepares to Sell Externally
Andy Jassy has detailed the success of Amazon's chip unit. With $20 billion in revenue and triple-digit growth, the company is ready to enter the open market and challenge Nvidia.
OpenAI Opens Cyber Model to Thousands of Security Professionals Amid Rivalry With Anthropic
OpenAI is releasing GPT-5.4-Cyber, a specialized AI model for defensive security and reverse engineering. Unlike Anthropic's closed approach, thousands of security professionals will get access.
GitHub Freezes Copilot Sign-Ups as AI Agents Break the Economics of Flat-Rate Pricing
GitHub has paused new user registrations because running AI agents costs more than monthly subscriptions bring in.
Clueso raises $10M for AI agents that write and update corporate video
Startup Clueso closed a $10M Series A led by Dharana Capital, bringing total funding to $12M. The money will go toward AI agents for scripting, recording and maintaining video.
Broadcom CEO Hock Tan Compares Many AI Investors to Casino Gamblers
Speaking in Singapore, Broadcom's CEO said many AI investors "have no idea what they're investing in" and called their expectations overly extreme.
Palantir CTO says the Pentagon has already replaced Anthropic's models with no loss
Palantir CTO Shyam Sankar said on CBS that the Pentagon has already replaced Anthropic's models, with no effect on operations. The company was earlier designated a supply chain risk.
NYT, AFP, and The Economist Join SPUR Coalition: Now 80+ Publishers Strong
The world's leading news outlets have banded together in SPUR to stop unauthorized use of their journalism in AI systems. The coalition now counts 80+ members across 20+ countries and has rolled out the first standard fo
Bloom Security finds a package-publishing vulnerability, and AI agents multiply the risk
Bloom Security shows how a flaw in npm publishing infrastructure, combined with AI agents that auto-install dependencies, opens a path for malicious code into corporate projects.
Visa Wins China's Approval to Prepare a Domestic Card Clearing Business After a 6-Year Wait
China's central bank has approved Visa's application to prepare a domestic card clearing business. The company filed in May 2020, but it still cannot process payments.
Mozilla launches Firefox Enterprise with a self-hosted management console
Mozilla has released Firefox Enterprise for companies and government agencies: a browser with a management console that can run on an organization's own servers. Pricing and timing are undisclosed.
Alibaba launches sales of Qwen Glasses N1 smart glasses from 2999 yuan
Alibaba has launched its Qwen Glasses N1 AI glasses in China from 2999 yuan ($450). With government subsidies and discounts, the price drops to 2218 yuan, about $333.
Semrush: ChatGPT Has Raised Its Share of News Sources Seven Months in a Row
Semrush: from January to July, news sites' share of ChatGPT's sources grew every month. Gemini, AI Mode and AI Overviews lag behind, and the gap is wider in the UK.
Aramco Digital and OrbitronAI to Build AI Tools for Heavy Industry
Saudi Aramco's technology arm and the startup OrbitronAI will jointly develop and sell AI for oil and gas, energy, mining, transport and manufacturing.
Microsoft President Brad Smith: Milk Is Regulated More Strictly Than AI
In a BBC podcast, Brad Smith called for global AI safety rules and a legal duty to tell victims when a rogue agent breaks into their systems.
Hundreds of Batam residents protest at data centers: no running water for months
On October 5, hundreds of Batam residents protested at Nongsa Digital Park, saying their homes have been without water for months while nearby data centers keep cooling servers.
IMF Sounds Alarm: $1.4 Trillion in Circular AI Deals Threaten Market Crash
The International Monetary Fund has added an AI-sector correction to its list of critical market risks. The problem lies in the structure of 14 major deals: hyperscalers poured $300 billion into financially weak AI compa
Amazon in Talks to Buy AI Startup Decart for $7 Billion
Amazon is in talks to buy AI startup Decart at a valuation of about $7 billion, WSJ reports. A deal could be announced soon if talks hold and no rival bidder steps in.
Anthropic and OpenAI seek influence over what outside auditors can publish
Reuters reports that both labs want control over what independent evaluators can publish about their models, citing the risk of trade secrets leaking.
Ithaca Holdings in talks to take Bumble private at roughly $1 billion valuation
Investment firm Ithaca Holdings is discussing a $1 billion acquisition of Bumble aimed at taking the company private. Shares jumped 50% after the announcement, though that's far below the $14 billion valuation at its 202
NTT Data Reportedly Nears Deal With Anthropic for Japan's First Claude-Dedicated Data Center in Osaka
Nikkei Asia reports NTT Data is close to a deal with Anthropic for an Osaka data center, Japan's first built specifically for Claude. The deal has not been confirmed.