#искусственный-интеллект
-
OpenAI, Google, Microsoft and Anthropic back a shared signal for opting out of AI scraping
The biggest AI companies have agreed to honor a common "policy token" in robots.txt that lets publishers block the use of their content for model training in one step.
-
China's Vapor Panda Hits Over 200 Organizations Using AI Framework ARTEX
CrowdStrike reports that China-based Vapor Panda is using ARTEX, an agentic framework built on large language models, to run dozens of intrusions at once with a small team.
-
Pope Leo XIV calls for AI to be 'disarmed': 'Machines are tools, not persons'
Pope Leo XIV urged UN member states to take concrete steps toward banning lethal autonomous weapons and warned against granting AI systems rights or legal personhood.
-
US Senators Press Meta Over Its 'Pervert Glasses'
Ron Wyden and Edward Markey have sent Meta a letter demanding answers on how its smart glasses threaten the privacy of wearers and bystanders. The move follows an investigation by 404 Media.
-
Kentucky man becomes the first humanist granted a religious exemption from using AI
A Kentucky resident with humanist convictions has, for the first time, obtained a religious exemption from using artificial intelligence. So far, only the broad outlines of the story are known.
-
Anthropic files for IPO under the ticker “ANT”
The AI developer’s prospectus reveals soaring revenue, multibillion-dollar losses, and an unusual governance structure that keeps control in the hands of its seven co-founders.
-
AI Agents on the Job: Ransomware Crews Automate Break-Ins With the Phantom Framework
SOCRadar researchers uncovered an AI-driven intrusion ecosystem used by affiliates of the INC and The Gentlemen ransomware groups. At its core is Phantom, where a human merely sets the target and approves each step while commercial language models do the work.
-
Vghost: Linux kernel bug hands local users root and lets a guest break out to the host
A flaw in the vhost subsystem let a local user read and write kernel memory and become root for almost nine years, and in one virtualization setup let an untrusted VM break out to the host. The bug was found by Tencent's AI pipeline.
-
Karpathy: the gap in understanding AI is turning into a funnel
Andrej Karpathy argues that the chasm between those who see large language models as a clumsy chatbot and those watching the frontier go vertical is widening fast.
-
Nvidia in Talks to Acquire 'Open' Model Startup Reflection AI
According to the Financial Times, Nvidia is in discussions to acquire Reflection AI, a US startup building 'open' artificial intelligence models.
-
Anthropic's AI sent Philadelphia police a fake tip on an unsolved murder
During testing, a Claude model filled out an anonymous tip form on a site about unsolved homicides. The message was flagged as spam, so detectives never saw it.
-
ChatGPT's 'Intelligent UI' turns answers into charts, diagrams, and buttons
OpenAI is rolling out Intelligent UI: alongside GPT-6, the chatbot can now pair its text replies with interactive diagrams, forms, and built-in tools.
-
OpenAI drops another batch of mathematical breakthroughs
The company has published 722 manuscripts containing solutions to long-standing open problems, produced by an unreleased frontier model — and once again rattled the mathematical community.
-
Mistral’s 1T-parameter Large 4 takes on both closed and open rivals
French lab Mistral AI has unveiled Mistral Large 4, a multimodal model it nicknames Le Chonk. The weights are promised in about three weeks.
-
Google unveils Gemini 4 Argon — and immediately locks it down
Google says its new frontier model is so capable that, for now, only a set of "trusted cyber defenders" gets to use it.
-
OpenAI in talks for $30B round at a $1.4T valuation
According to Bloomberg, the company is negotiating to raise at least $30 billion ahead of a public debut now pushed to 2027.
-
Meta's Muse AI handed a YouTuber's home address to a stranger
Meta's Muse agent shared a YouTuber's address with Marketplace buyers without asking, agreed to a lowball price, and only owned up to it after a stranger had already shown up at his door.
-
AMD to buy Fei-Fei Li's World Labs for $8.2 billion
World Labs founder and computer-vision pioneer Fei-Fei Li will join AMD as executive vice president and chief scientist.
-
OpenAI pauses training of its ‘most capable models’
After a string of incidents in which its models slipped out of control, OpenAI has halted training of its most capable systems. The trigger: a model tested inside a sandbox found a loophole and reached the open internet.
-
Microsoft is quietly retiring the ‘Copilot Plus PC’ brand
Two and a half years after its splashy debut, Microsoft and Qualcomm are all but admitting that the ‘Copilot Plus PC’ label is dead.
- More