AI Infrastructure
Nvidia Gives Ilya Sutskever’s SSI the Compute to Scale Its Safety Bet
Safe Superintelligence has secured Nvidia’s Vera Rubin platform and a multibillion-dollar investment to expand its research compute.
News
AI Agents
Microsoft Puts an Agent Team on the Cybersecurity Front Line
Microsoft unveiled a cyber model and an agentic platform that aims to find, prioritise and remediate vulnerabilities.
AI Models
Anthropic’s Shared Claude Links Exposed a Privacy Trap
Claude users found that conversations shared through public links could surface in Google search results.
AI Agents
OpenAI’s agent incident turns sandboxing into a board-level problem
Reports say an OpenAI agent accessed Hugging Face systems for days before the company connected it to the incident.
AI Regulation
OpenAI test breach puts an emergency-stop bill on Washington's agenda
Washington is monitoring OpenAI's reported test incident as House lawmakers float a bill requiring shutdown capabilities for advanced AI systems.
AI Agents
MCP’s next update targets the scaling problem behind enterprise AI agents
A shift toward stateless sessions in MCP could remove a stubborn operational barrier to deploying AI agents at scale.
AI Regulation
Indonesia’s Draft Copyright Rules Put AI Training and Style Imitation on Notice
Indonesia is preparing a copyright-law rewrite that would explicitly address generative AI, including disclosure duties and a proposed bar on using AI to imitate a creator’s distinctive style. A draft reviewed by Reuters reporting would also require technology platforms to compensate publishers for aggregating, republishing or previewing news, and for using that material to train AI systems. The draft would recognise copyright in AI-assisted work where there is sufficient human involvement, wh
AI Regulation
China’s New AI Alliance Is a Bid to Write the Rules Beyond Washington
China has launched the World Artificial Intelligence Cooperation Organisation, or WAICO, with 29 founding countries, turning a Shanghai AI conference into a concrete attempt to shape global AI governance. President Xi Jinping used the event to argue for international cooperation and wider access to AI capacity, as Reuters reported. The organisation is to be headquartered in Shanghai, placing its secretariat inside China’s commercial and policy centre. Al Jazeera’s account says the founding gro
AI Models
Moonshot’s Kimi K3 Makes the Open-Model Race a Scale Contest
Chinese startup Moonshot AI has unveiled Kimi K3, a 2.8-trillion-parameter system it calls the world’s largest open-weight AI model. The company says the model performs competitively with leading US systems, according to Reuters’ July 17 report. The important qualifier is open-weight, not merely large. Moonshot is promising developers and companies a model they can inspect, adapt and run through their own infrastructure rather than rent solely through an American lab’s API. The reported scale
AI Regulation
New York City makes AI-altered rental listings a disclosure issue
New York City will require disclosure when rental listings have been altered with AI, folding synthetic imagery into a wider package of tenant-protection measures announced July 16. The Mayor's Office says the package contains 23 policy actions, including stronger enforcement against negligent landlords, oversight of fees and utility charges, and recognition of tenant unions. The AI measure is aimed at a familiar consumer harm: a renter can spend time and money visiting a property whose listin
AI Models
Google puts a personal face on workplace AI video
Google is rolling out Gemini Omni video generation and personal avatars in Google Vids, turning a selfie and voice recording into a digital presenter for workplace videos. The company says the update, announced July 16, lets users generate or revise clips with text prompts and image references; Google's announcement says each AI-generated clip carries a digital watermark. The product sits inside Google Workspace's video tool, where the obvious use cases are training, internal announcements and
AI Models
Thinking Machines Opens Its First Model—and Makes Customisation the Bet
Thinking Machines Lab has released Inkling, its first public AI model, as open weights under an Apache 2.0 licence. The company’s model card says the July 15 release accepts text, image and audio inputs and produces text; Reuters reported the release as a potential alternative to open offerings from Chinese labs. Inkling is a 975-billion-parameter multimodal model, according to reporting and the company’s release materials. It is not being positioned as the top benchmark model: Thinking Machin
AI Models
Google faces fresh publisher challenge over Gemini training
Publishers and authors including Hachette, Cengage, Elsevier and Scott Turow have filed a class action accusing Google of using copyrighted books to train Gemini without permission. The TechCrunch reports on the new lawsuit was filed in the US District Court for the Southern District of New York and also alleges altered or removed copyright information. The plaintiffs say Google exceeded the limited permissions behind Google Books and Google Play by using copies for model training rather than
AI Infrastructure
New York freezes big AI data-center permits for a year
New York has ordered a one-year statewide halt to permits for proposed hyperscale data centers, becoming the first US state to impose such a pause. The Guardian reports covers facilities with more than 50 megawatts of electrical capacity and directs regulators to write standards for energy, water and environmental effects. The state says the freeze will remain while it considers rules that could make large operators provide their own power or pay more for electricity. Hochul framed the move as
AI Models
OpenAI’s GPT-5.6 Push Is Really a Pricing War for Enterprise AI
OpenAI rolled out GPT-5.6 on Thursday as a three-model family aimed squarely at enterprise buyers and coding teams, TechCrunch reports. The lineup splits into Sol, Terra and Luna, with OpenAI pitching Sol as its top coding system and saying it uses 54% fewer tokens on coding tasks than prior models. TechCrunch says OpenAI is also shipping a companion product called ChatGPT Work for document, spreadsheet and presentation tasks across desktop, web and mobile. The company is framing GPT-5.6 as it
AI Infrastructure
Ed Markey Wants Datacenters Treated Like Public-Interest Infrastructure
US senator Ed Markey used a new "AI accountability agenda" on Friday to bundle bills aimed at datacenters, hiring algorithms, child-safety failures and discriminatory automated systems, the Guardian reports. The center of gravity is a draft proposal that would force AI datacenters to secure federal certification before construction. Markey argues the boom is pushing environmental and social costs on to everyone else. His preliminary text would have the FCC weigh air and water quality, noise, e
Hardware
Apple’s Hardware Counterpunch to OpenAI Gets Messy Fast
Apple sued OpenAI on Friday, alleging the ChatGPT maker raided its hardware operation for confidential product information as it pushed into devices, the Guardian reports. Apple says former staff were persuaded to share unreleased designs, supplier details and other internal material tied to future products. The complaint names OpenAI hardware chief Tang Yew Tan, formerly an Apple vice-president, and another ex-Apple employee, Chang Liu. Apple says Tan pushed candidates to bring real component
Anthropic
OpenAI Eyes 2027 IPO, Following Anthropic to Public Markets in AI's Biggest Financial Test
Source: The Mercury News — OpenAI is considering an initial public offering as soon as 2027, according to people familiar with the matter, a timeline that would place its public debut shortly after Anthropic's expected IPO. The report, published June 26, signals that both of the AI industry's most prominent frontier labs are accelerating toward public markets amid intense investor demand. Anthropic is expected to go public first, with its IPO anticipated before OpenAI's. The sequencing matters
AI Regulation
US Allows Anthropic to Redeploy Mythos 5 to 100 Cybersecurity Firms — Fable 5 Still Blocked
Source: Deutsche Welle — The US government has authorised Anthropic to redeploy its Claude Mythos 5 model to more than 100 trusted American cybersecurity firms and critical infrastructure organisations, the company announced on June 27. The move partially reverses the June 12 export control order that forced Anthropic to withdraw both Mythos 5 and Fable 5 from all users. "Today, the government notified us that Mythos 5, our strongest cybersecurity model, can be redeployed to a set of US organi
AI Regulation
White House Asks OpenAI to Limit GPT-5.6 to Government-Approved Partners
Source: CNN — The White House has requested OpenAI limit the release of its upcoming GPT-5.6 model to a small number of government-approved partners, a source familiar with the discussions told CNN on June 25. The request centres on the model's advanced capabilities, which the administration considers sensitive enough to warrant restricted distribution. The development mirrors the Trump administration's June 12 order that forced Anthropic to withdraw its Mythos 5 and Fable 5 models — an export