#искусственный-интеллект
-
More than a million LinkedIn users have hit the “Seems like AI slop” button
LinkedIn has reported on the popularity of its button for flagging posts generated by artificial intelligence: in a matter of weeks, more than a million people have used it.
-
14 Trojanized npm Packages Deploy the AI-Controlled Linux Backdoor RedC2 4.0
Trend Micro researchers have found a set of npm packages that, posing as calendar utilities, covertly deploy the Linux implant RedC2 4.0. A single import anywhere in the dependency tree — even a transitive one — is enough to trigger the payload.
-
Torvalds Went Through “Debugging from Hell” — and Admitted AI Helped Him
Linus Torvalds personally committed a fix to the Intel Xe driver after an agonizing debugging session in which, by his own admission, artificial intelligence “helped enormously.”
-
Codex on AWS Bedrock overpays for cache: no control over prompt caching
An issue has been filed in the Codex GitHub repository: the native Amazon Bedrock provider can't enable explicit prompt caching for GPT-5.6 Sol, so agentic sessions generate a flood of expensive cache-write tokens.
-
Linux 7.3 Networking Maintainers: “Completely Buried” in AI Patches
Networking changes have been accepted into the Linux 7.3 merge window, but the maintainers admit they're drowning in a flood of fixes from AI agents — and they plan to answer it with AI of their own.
-
OpenAI Answers Anthropic with a Monitoring Service That Stores No Data
OpenAI has unveiled Private Safety Processing—a system that tracks abuse across multiple sessions at once while retaining no customer data. It's the company's bid to outmaneuver Anthropic in the debate over the privacy of enterprise customers.
-
Debian Votes on Eight Options for an LLM Usage Policy
The Debian project is holding a vote on whether large language models may be used in work on the distribution. Developers are asked to choose one of eight options — ranging from an outright ban to explicit approval.
-
US Warns of AI-Powered Attacks on Siemens PLCs in Critical Infrastructure
US intelligence agencies have reported active attacks on Siemens S7-series programmable logic controllers. The attackers are using scripts generated by artificial intelligence.
-
CISA Adds Actively Exploited Ray Vulnerability to KEV Catalog
The critical CVE-2025-62593 vulnerability in the Ray distributed computing framework allows arbitrary code execution through a developer's browser via a DNS rebinding attack.
-
Anthropic Is Embedding Invisible Watermarks in Claude's Text
The company has explained how it will label text generated by Claude in order to comply with the European AI Act. At its core is SynthID-Text, an open technology from Google DeepMind.
-
Nvidia Sharply Scales Back Its Guarantees for Financing OpenAI's Infrastructure
According to Reuters, citing the Wall Street Journal, Nvidia is substantially reducing the size of the guarantees it is prepared to provide toward the construction of OpenAI's data centers.
-
ChatGPT Learns to Watch You: The New Computer History Feature Remembers Clicks and Keystrokes
The ChatGPT desktop app for macOS has gained a Computer History feature that turns a user's actions into training data. OpenAI promises it doesn't take screenshots — it only records “events.”
-
A Model That Knows the World No Further Than Fifth Grade
Researchers trained language models on a corpus limited to the elementary-school curriculum in order to separate the acquisition of knowledge from its retrieval. The takeaway: the filter on the training data sets a ceiling on capability.
-
AI's Latent Reasoning Turns Out to Be More Legible Than Thought
Researchers at COLM 2026 showed that models with hidden reasoning either barely use it or conceal perfectly readable mathematical steps inside.
-
Debian Votes on the Fate of LLMs in the Project: From an Outright Ban to Climate Arguments
Debian developers have begun voting on a General Resolution addressing whether contributions created with the help of large language models are acceptable in the project. Nine options are on the table.
-
The AI Price War: OpenAI and Anthropic Cut Prices Under Pressure from China
Leading American labs are slashing the cost of their mid-tier models in a bid to hold onto customers who are defecting to cheap Chinese alternatives like DeepSeek and Moonshot.
-
Apple trained its own AI model for China together with Alibaba
Apple, according to Reuters, has developed a separate large language model for the Chinese market with support from Alibaba — a rare example of a US–China partnership in the current political climate.
-
Z.ai Announces GLM-5.3 with a Focus on Coding and “Cyber Skills”
Z.ai has unveiled its GLM-5.3 language model, positioning it as a frontier system for writing code and crediting it with newly emerged “cyber skills.”
-
Zoom Hacked in a Single Day: AI Finds a Zero-Click Flaw in Annotations
A Security has discovered a critical vulnerability in Zoom that lets an attacker take over the device of any call participant without a single click. The real story isn't the bug itself, but the fact that a “nation-state-grade” exploit was put together using publicly available AI models in less than a day.
-
ShieldFont: the font that turns pages into gibberish for AI scrapers
Designers have unveiled a font that keeps text readable for humans but feeds bots a garbled version, to “poison” AI training data.
- More