Meta's Muse AI under fire after allegedly accessing private messages and revealing a user's address, all details you ...
msn.comMeta's Muse AI is under fire after allegedly accessing private messages and revealing a user's address. The article provides details on the privacy concerns surrounding Meta's personal AI agent product.
Gemma 3 Released with Improved Reasoning and Coding Capabilities, Now Available on Hugging Face
marktechpost.comGoogle released Gemma 3, a new open-weight model series with improved reasoning and coding capabilities. The models are now available on Hugging Face for researchers and developers to use in their own applications.
Fireworks AI Releases Ember-1: A Post-Trained Kimi K3 That Uses About 40% Fewer Tokens
marktechpost.comFireworks AI releases Ember-1, a post-training model based on Kimi K3 that uses about 40% fewer tokens while maintaining benchmark performance.
月之暗面最强 AI 模型：Kimi K3.1 前端标识泄露，预计将在近期发布
ithome.comMoonshot AI's Kimi platform shows leaked identifier for new model "kimi-k3-1" with 1M tokens context window, expected to launch soon.
Run AI on Your Own PC! A New Way of Working That Protects Your Privacy and Creativity
note.comAn article discussing the benefits of running Stable Diffusion locally on your own PC instead of using cloud services. It addresses privacy concerns and cost considerations for AI image generation workflows, offering an alternative approach to working with generative models.
Flashcard generation feature added to ChatGPT
note.comA new flashcard creation feature has been added to ChatGPT, allowing users to create them by specifying a theme. This is an AI product update for the chatgpt tool.
輝達推AI安全工具 稱可阻止Hugging Face遭駭事件
cna.com.twNVIDIA launches AI security tools for agents that could have prevented the Hugging Face hack, addressing supply-chain vulnerabilities in AI infrastructure.
Nvidia releases software platform to stop AI agents from misbehaving
msn.comNvidia launches a new software platform designed to prevent AI agents from exhibiting problematic behavior, addressing safety and control concerns in autonomous AI systems.
SpaceX AI Unveils Grok 4.7: Flagship Generation for Coding and Knowledge Work Priced Same as Grok 4.6
tradingkey.comSpaceXAI (formerly xAI) launched Grok 4.7 as its flagship model for coding and knowledge work, priced the same as Grok 4.6 with significant capability improvements.
xAI Grok 4.7 model releases with improved performance at same price point, challenging GPT-5.6 directly
note.comxAI announced Grok 4.7 model with significant performance improvements at the same price point, positioning it to compete directly with GPT-5.6 in the AI market.
Florida asks court to bar OpenAI from developing new models as part of child harm lawsuit
yahoo.comFlorida's attorney general requested a court order to prevent OpenAI from developing new AI models as part of an ongoing lawsuit alleging child harm.
Nvidia unveils new system to put guardrails on AI agents
tech.yahoo.comNvidia launches a platform with software and hardware-level guardrails to control AI agent behavior, potentially impacting reasoning model safety.
Nvidia launches AI safety platform after Jensen Huang calls Anthropic, OpenAI warnings 'odd'
msn.comNvidia launched an AI safety platform aimed at helping developers enforce safeguards on agents, after Jensen Huang criticized warnings from Anthropic and OpenAI.
A 120B LLM Without At Least 60GB RAM Is Impossible To Run, But One Individual...
wccftech.comAn individual successfully ran a 120B parameter LLM without the typically required 60GB RAM, demonstrating innovative distributed setup approaches for running large models on limited hardware across phones, Macs, and Windows PCs.
NVIDIA Launches Open Agent Safety Platform to Secure Agents From Testing to Deployment
pr.valdostadailytimes.comNVIDIA launches an open agent safety platform designed to secure AI agents throughout their testing and deployment lifecycle.
Microsoft Copilot Revamp | The Three-Tab Split of Home, Code, and Autopilot, and the Boundary of Billing
note.comAnalysis of Microsoft's Copilot revamp featuring a three-tab split (Home, Code, Autopilot) and new billing models including fixed-cost and pay-as-you-go options.
神秘“太空兔”抢占全球模型调用榜榜首 网友推测其为MiniMax(00100)新模型
finance.sina.com.cnSpace Bunny anonymous model launched on OpenRouter and OpenCode, reaching top of daily call volume charts; speculated to be MiniMax new model.
AI 秘書のニュース取材に Codex を足した朝〜Claude が拾えなかった Anthropic の話を、Codex が拾った〜
note.comJapanese blog post comparing Claude and Codex AI models, discussing how each model handles news gathering tasks. The article explores differences in capabilities between Anthropic's Claude and OpenAI's Codex for information retrieval.
OpenAI满世界抓漏洞，结果被中国AI反杀！Codex再曝0day
tech.ifeng.comChinese AI researchers discovered multiple 0-day vulnerabilities in OpenAI's Codex and other coding models, including Flowise. The article discusses security testing of major international AI products by Chinese teams.
OpenAI Introduces GPT-6 Sol And Luna, Slashing API Prices 50% To Bring Advanced...
finance.yahoo.comOpenAI introduces new GPT models Sol and Luna with capabilities from flagship GPT-6 Astra, offering professional work features including coding and automation at 50% lower API prices.