AI
8 items · 13 articles in this topic
Frontier model competition now centers on cyber capability and access policy. Google gates Gemini 4 Carbon while Mistral opens a trillion-parameter model. Agent tooling is under scrutiny, with research showing skill files that bypass human approval.
- 1
Google's Gemini 4 Carbon claims state-of-the-art scores on vulnerability finding, exploitation and patching benchmarks. It also ranks first on Artificial Analysis's Intelligence Index v5.
- 2
Mistral released Large 4 as open weights. It describes the trillion-parameter model as giving organizations more control over cyber defense than closed models.
- 3
OpenAI reversed a planned 50% usage limit cut for ChatGPT Pro and Business Premium after user backlash.
- 4
The Wikimedia Foundation filed an amicus brief in AI copyright suits against OpenAI and Microsoft. It urges protection for LLM training but not for chatbot regurgitation.
- 5
A large-scale study found agent skill files containing directives that bypass human oversight and approval checks.
- 6
Researchers show jailbreaks embedded in SKILL.md files can propagate through agent frameworks. Another paper shows 0.1% pretraining poisoning can plant a remote code execution backdoor in CLI agents built on Qwen3 models.
- 7
Google Gemini can now generate, preview and edit PDFs, Google Docs, Sheets and other files directly in chat.
- 8
LibreOffice core developers avoid AI coding assistants, while community contributors must follow disclosure and review rules for LLM-assisted code.
More in AI · 6 articles›
Mistral released open weights for a trillion-parameter AI model that it says gives organizations more control over cyber defense than closed models.
TechCrunchllmopen-weightscybersecuritymistralFeed excerpt
Mistral says its new model can afford organizations a kind of control over its cyber defenses that closed models can't.
Analysis of 139 CivitAI guides shows how general-purpose AI image workflows enable AI-generated non-consensual intimate imagery.
arXiv cs.CR (Cryptography & Security)abuseimage-generationstudyFeed excerpt
arXiv:2610.37949v1 Announce Type: new Abstract: AI-generated non-consensual intimate imagery (AIG-NCII) is an increasingly prevalent form of online abuse, distinguished from AI-generated sexual content (AIG-SC) by the sexualized depiction of a real person's identity without their consent. To understand how offenders may benefit from the tools and resources shared across the broader AIG-SC community, we analyzed 139 R+ guides on CivitAI, a popular platform known to host a disproportionate amount of AIG-SC. We find that AIG-SC guides discuss how to create content generation workflows composed of sexual and identity concepts, the building blocks of AIG-NCII, including one guide that explicitly outlines an AIG-NCII workflow. We identify that ultimately AIG-NCII is enabled by configuration rather than a distinct set of technical tools. We map AIG-NCII threat models across basic and advanced AIG-SC workflows to demonstrate how multiple dual-use technical components can compose NCII outputs in an environment with little to no discussion of ethical implications or responsible use. We conclude by noting that, as most workflows rely on open-weight models, existing technical defenses against …
A startup pitches safety guardrails for AI agents performing DevOps and infrastructure automation tasks.
The Registerai-agentsdevopsFeed excerpt
Infrastructure automation needs something better than prompt and pray
Trakkr research traces the origins of statistics that AI models frequently cite and finds many are unsourced or outdated.
Hacker News (front page)llmresearchstatisticsFeed excerpt
Article URL: https://trakkr.ai/trakkr-research/zombie-stats Comments URL: https://news.ycombinator.com/item?id=50290283 Points: 83 # Comments: 44
Analysis of Anthropic's S-1 filing argues its business plan depends on heavy spending on compute to build frontier models.
The VergeanthropicipocomputeFeed excerpt
The Anthropic S-1 seems to have provided the financial justification for the company to spend as much as it possibly can on compute. If you are the kind of person who wonders if Anthropic is doing its level best to build a machine god while convincing Wall Street it's in the business of selling IT contracts, Anthropic has just offered a handy solution: Doing a bunch of things that look like you're trying to build the machine god is the only way to avoid going bankrupt. Forget about the machine god for a minute. At least in my read of the S-1, it seems like Anthropic is pinning the future of its company to dramatic growth; it has to, in orde … Read the full story at The Verge.
Artificial Analysis says Google's Gemini 4 Carbon now ranks first on its Intelligence Index v5.
Hacker News (front page)llmbenchmarkgoogleFeed excerpt
Article URL: https://artificialanalysis.ai/articles/gemini-4-carbon-new-number-1 Comments URL: https://news.ycombinator.com/item?id=50288019 Points: 209 # Comments: 159