模型发布/更新
-
Qwen3.8-Max released: most powerful open-source coding and collaboration model with 2.4T parameters
Qwen officially released Qwen3.8-Max, the most powerful model in the Qwen family to date, with 2.4T parameters (95B…
-
SenseTime releases SenseNova U1.5-Lite-Preview open-source model
SenseTime introduced SenseNova U1.5-Lite-Preview, a lightweight native unified multimodal model based on the NEO-Unify…
产品发布/更新
-
Cloudflare 推出 Billable Usage API:为自助账户提供按产品与计费周期的程序化成本可见性
Cloudflare 发布 Billable Usage API,为自助账户提供单一端点,一次调用即可返回按产品和计费周期拆分的用量与成本,覆盖 Workers、R2、D1、Workers AI、Vectorize、Images 和 St…
-
OpenRouter launches Ori Eval to simplify evaluation
Introducing Ori Eval: the easiest way to write your first evaluation. There is no absolute best model, only the model…
-
Cloudflare 推出 @cloudflare/computer 预览版:为智能体提供虚拟文件系统与多执行环境
Cloudflare 发布 @cloudflare/computer 早期预览版,这是一个开源智能体运行时,为每个智能体提供虚拟文件系统,并支持在 isolate、容器沙箱或浏览器中执行代码。
-
GPT-Live real-time audio architecture released
GPT-Live is a new architecture and stack for real-time audio: GPT-Live can listen while it speaks. To make this…
-
Microsoft open-sources Orchard agent training framework
Orchard is an open-source framework for the research community to train and evaluate AI agents across task types. It…
-
Cloudflare Workers 与 Containers 现已支持入站 TCP 连接和 gRPC
Cloudflare 在 Agents Week 期间推出 Workers 运行时新处理器 connect(socket),可直接接受 Spectrum 提供的入站 TCP 套接字,并支持将套接字转发至 Durable Objects 或…
-
Databricks 推出 Variant 通用可用版本,加速半结构化数据摄取
Databricks 宣布 Variant 正式全面可用(GA),用于更快速、高效地摄取 JSON、XML、CSV 等半结构化数据。该功能旨在简化传统上复杂的数据导入流程,提升处理效率。
行业动态
-
EU AI Act transparency rules take effect, violations face fines up to €15 million
New transparency obligations under the EU AI Act took effect on August 2, requiring companies to disclose when users…
-
Databricks 完成对 Panther 的收购,加速安全湖仓时代
Databricks 宣布正式完成对安全数据平台 Panther 的收购,旨在加速安全湖仓(Security Lakehouse)时代。此次收购将 Panther 的安全分析能力整合进 Databricks 平台,帮助企业在统一的数据架构…
论文研究
-
多模态大语言模型对齐的全面研究:Apple 团队独立拆解偏好对齐各环节
Apple 研究团队系统梳理了多模态大语言模型(MLLM)中的偏好对齐方法,将算法分为离线(如 DPO)与在线(如 online-DPO)两类,并发现两者结合可在特定场景下提升模型性能。团队还提出无需额外标注或外部模型的新型多模态偏好数据…
技巧与观点
-
After strong quarter, Palantir CEO Alex Karp calls AI industry "Marxist"
In his quarterly shareholder letter, Palantir CEO Alex Karp warned that frontier AI labs are too untrustworthy for…
-
Kimi Work slide creation tutorial released
Making slides with Kimi Work - Tutorial #1. Kimi Slides handles the entire slide creation workflow: - Clear structure…
-
EA chief strategy officer on how generative AI enters playable live game worlds
EA Chief Strategy Officer Mihir Vaidya believes games are a testing ground for AI, but generative AI entering games…
-
Behind Google Agent Skills: how to build, test and scale
The Google Agent Skills team detailed the build and governance process for its open-source skills library: the project…
-
关于 Astra 与数学的两则重要更新:Anthropic 数学家 24 小时复现 OpenAI 半数结果
Gary Marcus 称 OpenAI 的 Astra 可能并非其宣传的突破,Anthropic 数学家 Levent Alpöge 用已公开的 Fable 模型在 24 小时内复现了 OpenAI 半数结果,质疑其真实进展。
-
Paramount CTO Phil Wiser:AI 属于史上最伟大技术趋势,但“iPhone 时刻”尚未到来
Paramount 首席技术官 Phil Wiser 认为 AI 应跻身人类史上最伟大技术趋势前五,其影响堪比火的使用。他主张等待依赖条件成熟、以“aha 时刻”引导采用,而非以技术为先导;并警告行业巨头过度分析将错失窗口期,这一窗口可能…