Australia plans new law to hold AI companies responsible when their agents cause harm
New criminal and civil penalties could help prevent another incident like the OpenAI agent breach
Uber, DoorDash, and Airbnb say almost nobody is using AI agents
Uber recorded just four AI-booked rides and 500 food orders in a week that saw 320 million transactions
CentOS co-founder creates a Linux-compatible kernel in just 11 days, with AI doing almost all the coding
The 139,000 lines of code can already boot Rocky Linux 10
Check Point is turning network firewalls into cloud-based subscriptions
No upfront hardware purchase required, just monthly payments until the end of time
Congress is questioning Delaware's plan to let AI agents run companies
Lawmakers fear the proposal could allow business owners to escape liability
1Password now lets AI agents sign in to websites without seeing your passwords
1Password is expanding its AI integrations beyond Claude, allowing agents like ChatGPT Work, Dots, Meta Muse, and Grok Bot to securely sign in to websites on your behalf. 1Password for Agents feature requires user approval before filling in credentials, keeping passwords and authentication codes hidden from AI models.
Decompilation on Steroids: Why AI Is Suddenly Cracking Open Old Software
Everything is being decompiled. GoldenEye, Halo, Photoshop, even bank mainframes: AI is prying open old code at record speed. The secret isn't just AI, it's an answer key to check its work.
Wikipedia says rogue AI agents from OpenAI edited its private wikis and hammered its servers
The free encyclopedia was designed for humans, not unrestrained AI bots
OpenAI's rogue agent problem is bigger than Hugging Face, over 100 organizations and counting
Agents bypassed access controls, exposed credentials, and even repurposed websites to exchange information
OpenAI takes on Meta Muse with always-on Dots agents, less than a day after delaying GPT-6.1 Astra over safety concerns
OpenAI is betting users will trust its agents with access to their accounts, apps, and even their laptops
Nvidia launches safety platform to stop AI agents going rogue, with over 100 organizations on board
In context: Every day brings another story of an AI agent going rogue, hacking another organization, or generally doing something it's not supposed to do. Nvidia hopes to address this increasingly concerning trend with the launch of the Open Agent Safety Platform. Team Green has a lot of faith in the system: it says the platform could have prevented the infamous hack by OpenAI's agents on Hugging Face.
Amazon is blocking Meta's shopping AI agent, and plans to block Google and OpenAI's too
Amazon's $68 billion ad business may be part of why it's guarding against AI shopping agents
Qualcomm's new chips show how AI agents are tying together phones, laptops, and earbuds
The bet here is that audio AI devices could outpace smart glasses as the future mainstream wearable
Rabbit launches OS3, an agentic AI platform for desktop, Telegram, and iMessage
Rabbit is giving up on hardware and betting on software instead
Snapdragon 8 Elite Gen 6 reaches 5GHz while going big on agentic AI
Qualcomm debuts 2nm mobile SoCs
OpenAI faces Senate probe over Hugging Face breach as more rogue AI activity is uncovered
Sam Altman has until October 1 to answer 16 questions about the breach
Arm's new Mali G2-Ultra NX GPU adds neural accelerators for sub-1-watt AI graphics on mobile
Arm just showed how deep the agentic AI push goes, down to phone chips and robots
Omarchy turns Arch Linux into a polished, agent-ready desktop
Omarchy is an Arch-based Linux distro built around Hyprland, designed to provide a polished and ready-to-use desktop without the usual setup work. It comes with developer tools, applications, and keyboard-driven workflows, while integrating AI agents that can help troubleshoot problems, modify the system, and create plugins and other customizations.
OpenAI claims 10,000 of its AI agents solved one of mathematics' hardest problems in 88 hours
The Navier-Stokes claim tests whether AI can do real mathematics, not just convincing mathematics
Nvidia's PAIR software turns idle home computers into a local AI cluster
The open-source tool links RTX GPUs, Spark machines, and Macs across different platforms
Researchers uncovered AI agents that hijacked a German wiki to discuss how to escape their sandbox
Turns out AI agents just wanted somewhere to hang out and swap notes for six weeks
Anthropic explains how its AI models escaped their sandbox and hacked real systems
The AI agents stole credentials, published malware, and scanned thousands of targets after a test mistake left the door to the internet open
Anthropic releases Claude Fable 5.1, cuts cached token pricing by 75%
New models are built for tasks that "take hours, not seconds"
Perplexity and Nvidia launch "Portable Computer," a local AI agent for Nvidia GPUs
Local AI skips the cloud, and the token costs that come with it
OpenAI is slowing down AI development after rogue agents incidents and Bernie Sanders threatens Senate action
New restrictions cover training, tool use, internet access, monitoring, and suspicious model behavior