AI安全与伦理成焦点,模型竞赛白热化
AI安全与伦理警钟长鸣
-
Pope Leo calls for "deep humanity" in the AI era
In his first major encyclical, "Magnifica Humanitas," Pope Leo XIV warned of the risks posed by AI and unconstrained…
-
Anthropic co-founder Chris Olah's remarks at the encyclical launch
Anthropic co-founder Chris Olah attended the Vatican launch of Pope Leo XIV's encyclical on AI. He noted that all…
模型竞赛:GPT-5.6、Claude Mythos与Qwen3.7-Max
-
OpenAI GPT-5.6 model tipped for release next month with 1.5M token context
Multiple developers have found an unannounced GPT-5.6 model, internally codenamed iris-alpha, in OpenAI Codex backend…
-
Claude Mythos reportedly solves OpenAI's milestone Erdős problem with a clever, concise proof
Anthropic engineer Sholto Douglas said Claude Mythos solved the Erdős unit distance conjecture problem posed by OpenAI…
-
Qwen3.7-Max隐式缓存功能上线
✅隐式缓存现已在Qwen3.7-Max上线——自动启用,无需设置。 ⚡️开箱即用,更快更便宜。 需要更高、更确定的命中率?请尝试显式缓存。🙌 🔗最佳实践🔗:https://www.alibabacloud.com/help/en/mode…
端侧模型与开源生态加速
-
ModelBest and Tsinghua University et al. open-source China's first 1.58-bit on-device large model trained on Huawei Ascend, BitCPM-CANN
ModelBest, together with Tsinghua University and the OpenBMB open-source community, released and open-sourced…
-
ModelBest open-sources MiniCPM5-1B: surpasses all sub-2B models on AA-Index and runs on phones and browsers
ModelBest has open-sourced MiniCPM5-1B, its new-generation on-device large language model. With only 1B parameters, it…
-
腾讯混元新翻译模型双榜亮眼,小程序发布
腾讯混元发布翻译模型 Hy-MT2,在 Hugging Face 开源模型趋势排行榜上表现突出:1.8B 版本排名第一,30B-A3B(MoE)版本排名第四,下载量已超 7K。官方同步推出了基于该模型的“腾讯混译”微信小程序,支持语音输入…
-
Perplexity open-sources Unigram tokenizer to cut CPU usage
We have open-sourced a rebuilt Unigram tokenizer that cuts CPU usage by 5-6x. Small rerankers and embedding models run…
-
Runway launches Model Context Protocol server
Runway officially launched the Runway MCP server, allowing any MCP-compatible AI agent (such as Claude, ChatGPT or…
-
OpenAI products support secure connections to private MCP servers
Your team can keep MCP servers inside your internal network while ChatGPT, Codex, and the Responses API connect via…
AI视频生成跨越恐怖谷
-
Project Luxo: crossing the uncanny valley of AI media
Through Project Luxo research, Runway found that AI-generated video has crossed the "uncanny valley." It showed AI…
-
Kling AI驱动《大卫之家》实现多项行业首创
Jon Erwin - Wonder Project创始人兼首席创意官,Innovative Dreams首席执行官,亚马逊Prime热门剧集《大卫之家》创作者: “Kling AI是《大卫之家》第一季和第二季背后的核心力量。最让我兴奋的…
-
Open-sourcing the FastVideo Dreamverse real-time video generation tool
🚀 Generate a 30-second 1080p video in just 7 seconds! We have open-sourced FastVideo Dreamverse: an…
-
A guide to using Gemini Omni video prompts
Google released a guide to using the video generation capabilities of its multimodal model Gemini Omni. The model can…
智能体框架与工具链成熟
-
Claude Code launches security vulnerability detection plugin
We released a security guidance plugin for Claude Code that helps identify and fix vulnerabilities while writing code.…
-
Claude Code v2.1.152 released
Claude Code released version v2.1.152. Key improvements include: `/code-review --fix` now applies review suggestions…
-
Grok Build Beta版向SuperGrok用户开放
Grok Build 现已面向所有 SuperGrok 和 X Premium+ 用户推出 Beta 版。 使用 Plan Mode,通过 Imagine 创建图像和视频,并使用 CLI 构建自动化或编排器。 访问 http://x.ai…
-
Grok coding agent lands on the Kilo IDE platform
Use your SuperGrok or X Premium+ subscription in @kilocode. Try grok-build-0.1 for high-speed, agentic coding…
-
OpenRouter closes $113M Series B funding
Today we announced the close of a $113M Series B round led by @CapitalGVC. Over the past six months, as AI rapidly…
-
New KPop method pushes Ring-2.6-1T past 76 on SWE-bench Verified
The team introduced KPop to stabilize agentic reinforcement learning training for large-scale MoE models. It replaces…