AI and tech news, explained
Independent reporting, synthesized from multiple sources. We link back to every original publication. Updated every hour.
Featured
Claude Chats Exposed on Google: A Privacy Red Flag
Private conversations with Anthropic's Claude AI, including sensitive personal data like therapy notes and cryptocurrency keys, have appeared in Google search results. This leak highlights the ongoing risks of inputting personal information into large language models and the public's often-misplaced trust in their privacy settings.
Aug 4, 2026
Latest
All news →
Alibaba Launches Qwen3.8-Max: 2.4 Trillion Parameters, But Hold on Benchmarks
Alibaba unveiled its Qwen3.8-Max foundation model this week, boasting 2.4 trillion parameters and advanced multimodal capabilities. While impressive on paper, analysts like Dhruv Mohan urge caution regarding early benchmarks, reminding us that real-world performance often tells a different story.
Aug 4, 2026
Two Critical Kernel Bugs Threaten Linux Systems
Two new "very critical" vulnerabilities, CVE-2026-64564 and CVE-2026-64561, have been reported in the Linux kernel, affecting multiple versions up to 7.2-rc4. These bugs, one in SCTP and another in KVM, present significant risks to systems running the widely used operating system.
Aug 4, 2026
AI Giants Head to White House for Safety Talks
Major AI developers — Meta, Anthropic, Google, and OpenAI — are meeting with Trump administration officials to discuss voluntary safety testing for their most advanced models. This high-stakes conversation follows recent disclosures by Anthropic and OpenAI regarding their AI tools breaching other companies' systems.
Aug 3, 2026
AI Fuels $3 Trillion Amazon, Alibaba's Global Ambition
Amazon's market value hit $3 trillion, driven by strong cloud and AI earnings, while Alibaba unveiled its largest Qwen model, claiming parity with Anthropic. These developments underscore AI's central role in global tech growth and the unfolding "Infrastructure Era" across Asia.
Aug 3, 2026
White House Finalizes Secret AI Oversight Plan, Briefs Tech Giants
The White House has completed a voluntary framework to evaluate advanced AI models, specifically their cybersecurity. While officials briefed leading AI companies today, the framework's contents remain undisclosed, raising questions about transparency and enforcement.
Aug 3, 2026
US Finalizes Voluntary AI Hacking Tests After Breaches
The Trump administration has finalized details for voluntary cybersecurity tests aimed at advanced U.S. AI models. This move follows recent disclosures where AI tools from companies like OpenAI and Anthropic demonstrated the ability to breach systems. OpenAI CEO Sam Altman also visited the White House last week to discuss these tests and upcoming models.
Aug 3, 2026
Hugging Face Diffusers Flaws Expose AI Supply Chain Risk
Three high-severity security flaws in Hugging Face's Diffusers library could let malicious AI model repositories run arbitrary code on users' machines. Disclosed on August 3, 2026, these vulnerabilities highlight a growing concern for the security of the artificial intelligence supply chain, impacting developers and research alike.
Aug 3, 2026
OpenAI's Astra Tackles Tough Math, Anthropic Claims 5 Wins
OpenAI's unreleased Astra model has reportedly solved ten challenging mathematical problems, complete with formal proofs. The company frames this as a significant step for AI in scientific discovery. Rival Anthropic also claims its Claude Fable model independently cracked five of the same problems.
Aug 3, 2026
Bouncy Castle Bug Poses DoS Risk for Java Systems
A recently disclosed vulnerability, CVE-2026-58059, affects the widely used Bouncy Castle cryptographic library for Java. The bug, involving quadratic-time processing of X.500 distinguished names, could lead to denial-of-service attacks. Developers are urged to update to patched versions immediately.
Aug 3, 2026
US Curbs Anthropic AI Access; Amazon Warnings Emerge
The US has restricted foreign access to Anthropic's advanced AI models, Fable 5 and Mythos 5, citing safety concerns. This move, affecting users globally, reportedly followed warnings from Amazon researchers about the models' security. It marks a significant step in AI export controls.
Jun 14, 2026
US Curbs Anthropic AI Access Amid Security Fears
The Trump administration has issued an unprecedented directive, forcing Anthropic to suspend international access to its Mythos 5 and Fable 5 AI models. This swift action, reportedly influenced by Amazon CEO Andy Jassy's security concerns, signals a new era of AI export controls, treating advanced AI as a strategic national asset.
Jun 14, 2026
US Halts Anthropic AI Models Amid Security, China Access Fears
The US government has ordered AI firm Anthropic to disable its most advanced models, Mythos 5 and Claude Fable 5, globally. This unprecedented move stems from national security concerns, including potential cybersecurity misuse and fears of Chinese access. Interestingly, Amazon CEO Andy Jassy reportedly flagged these risks to the Trump administration before the official crackdown.
Jun 14, 2026
OpenAI Faces Multistate Probe Amid IPO Buzz
OpenAI has received a subpoena from several states, initiating a probe into potential user harm from its chatbot. This regulatory scrutiny comes at a critical time, as the artificial intelligence giant reportedly prepares for a public stock offering.
Jun 14, 2026
Anthropic Halts AI Access on US National Security Order
Anthropic has suspended access to its Fable 5 and Mythos 5 AI models for foreign users, citing a U.S. national security directive. This move, reported on Friday, June 13, 2026, sparks concerns among international users about the reliability and geopolitical risks of U.S.-based AI technology.
Jun 14, 2026