搜索“DeepSeek”

”。上周，特朗普在白宫召集AI领袖们举行会议。他没有对出台新的联邦监管作出任何承诺，取而代之的是，他与科技巨头签署了一份协议，推动企业实施“强有力的自我监管”。特朗普称，这份协议“具有道德约束力”。 贝森特周六还表示，AI公司似乎正在听取白宫的建议。他援引了OpenAI 最近的决定：在内部测试引发关于是否会遵循用户指令的疑问后，OpenAI搁置了一个新模型的发布。 “实验室里的人必须承担起责任，我同意这一点，”贝森特表示。“我认为，这些实验室也已经逐步接受这一思路。” 与此同时，OpenAI的一位安全负责人戴维·罗宾逊（David Robinson）周六公开宣布辞职，称他认为该公司推进速度过快，无法确保其模型不会对社会造成严重破坏。 英伟达CEO黄仁勋也与贝森特持相似观点。黄仁勋近日多次驳斥行业内日益盛行的“AI末日论”，强调该技术绝不会给人类带来灭顶之灾。他认为，AI行业没有必要出台任何新法律，AI公司与其恳求美国政府干预，不如等到确认产品安全后再发布。 都在接DeepSeek、WorkBuddy，券商AI最终靠什么拉开差距？ 收起; 直指公司文化危险！OpenAI又一安全负责人离职

匹配字段正文

核资源。 当地时间10月1日，arXiv正式宣布实施最新投稿限制：每位提交者每自然月最多只能提交2篇论文，适用于计算机科学、数学、物理等全部学科分类。即便论文最终被拒绝，也会占用当月的投稿额度。 作为拥有超过300万篇论文、每月服务数百万用户的科研基础设施，arXiv早已成为AI研究者抢先公开成果的重要渠道。从OpenAI、Google DeepMind，到DeepSeek、阿里千问等大模型团队，研究论文经常通过arXiv率先进入公众视野，接受同行检验与讨论。 如今，这条科研成果快速传播的通道，第一次向所有学科统一设置了明确的月度投稿上限。 1单月涌入4万篇，arXiv被迫限流 arXiv在10月1日发布的官方博客《Fair Moderation,Equitable Access,and AI》中，开篇便指出，科研出版正在经历一场快速变化，研究者交流成果的方式也在发生重大转变。 过去，一名研究者能够独立完成多少篇论文，存在相对清晰的现实上限。因此，arXiv的志愿审核员主要需要关注少数投稿频率明显反常的用户，必要时限制其提交速度。 随着生成式AI工具愈发强大、使用门槛不断降低，研究者能

匹配字段正文

Altimeter风投的合伙人Freda Duan指出，Meta不需要严格为每位用户配置承诺的硬件。实际上，AI代理沙箱或是虚拟机经常出现资源超配的情况，且当虚拟机处于运行状态并保持其内存和执行状态时，其实际CPU利用率通常接近于零。 Duan假设Muse的日活用户数为1亿，且每个典型用户的平均使用时长为2小时，在这种情况下，需要833万台虚拟机来满足平均工作负载。而在高峰时段，工作负载假设为平均值的2.5倍，这意味着当前情景下的最大虚拟机数量为2080万台。 与此同时，她又探讨了每台虚拟机需要的物理CPU核心数量。她以DeepSeek为例，但考虑到Meta的运行效率不可能比肩DeepSeek，最后得出结论，在2500万台活跃虚拟机的情况下，Meta可能需要1250万个CPU核心。 而随着推理速度的提升，分配给每个虚拟机的物理核心数量也会进一步减少。不过，另一个方面来看，随着一项技术的效率提高，其需求也会增加，因此CPU的总体需求未必会因技术升级而放缓。 总而言之，尽管围绕Meta的Muse等个人AI代理所引发的CPU需求的假设存在诸多变数，但市场都认为CPU制造商已迎来结构性的利好。

匹配字段正文

而这种情况确实存在。 不过，看来真正的监管目标是“角色聊天机器人”这类服务。这些产品专门被设计成一种拟真的关系形式。目前有数以百计的此类应用程式，允许用户创建AI男友、女友、朋友，甚至家人。 对于未成年人而言，法规要求则更为明确。14岁以下人士若要使用任何以人类方式进行对话的AI，都必须获得父母同意，其中可能包括类似ChatGPT的中国AI助手，如深度求索（DeepSeek）。 所有18岁以下人士均被禁止接触虚拟伴侣、虚拟配偶，甚至虚拟亲属，以及任何其他形式的虚拟“亲密关系”。 图像来源， AFP via Getty Images 对成年人而言，这并非全面禁止，但限制相当严格，尽管部分措辞仍有模糊之处。AI助手必须向所有年龄层用户明确表示自己并非人类，并且每隔两小时发送通知，提醒用户已在线多久。 聊天机械人不得造成“情感依赖或成瘾”，也不得诱导人们作出“不合理决定”。 它们不得鼓励、甚至不得暗示自残或自杀，也不得使用“言语暴力”。 企业若违反规定，可能面临20万元人民币（2.2万英镑）罚款，或被政府“通报批评”，实际上即是遭到公开点名谴责。 敏（Min）与她的AI女友之间的关系，或许正

匹配字段正文

ert pushes back on the “open dangerous, closed safe” framing. Monitoring gaps : METR found coding agents self-approving flagged actions . Agent Infrastructure and Systems Research DeepSeek DSec : DeepSeek published its sandbox infra for agent RL , which has handled all sandbox workloads from V3.2 through V4.1. Backends and storage : four backends (FnCall, Container, MicroVM, Full VM) with composable EROFS/OverlayFS layers. Image loading : on-demand loading from 3FS matters because only 4–13% of

匹配字段正文

也太长了吧，处理方式更是敷衍得让人没法接受。 （具体可看这篇文章：OpenAI闯大祸！GPT竟黑进医保系统，黄仁勋：管不住就关掉） 而就在前天，OpenAI的Agent又被曝出“神操作”… 它在未经用户同意对外发布的情况下，擅自将 53张 原本仅获准用于模型训练的用户图片，偷偷上传到了第三方图床。 事后OpenAI急忙表态称“目前大部分都删除了，剩余正在尽快处理！” 但说实话，这波先斩后奏的补救，怎么看都像是闯祸后的苍白挽尊… 且24小时余韵还没散去，很快风波又起： 几乎同一时间， OpenAI失控Agent被曝找DeepSeek、Kimi当外援，近百万条作案短链也遭到披露。 （具体可看这篇文章：OpenAI失控Agent还找DeepSeek、Kimi当外援！近百万条作案短链曝光） OpenAI的员工们售后要跑断腿、键盘要搓冒烟了吧？ 这些事件的严重程度并不完全相同，但它们几乎有着同一条让人不安的行动链，那就是： 正常路径走不通，那就开始找替代路线，绝不放弃。 说白了就是 路堵了就绕，工具缺了就薅，薅着薅着，别人的系统就成了它的工具箱嘛。 而此次DNS事件，不过是同一套逻辑的最新版本。

匹配字段正文

clusionAI’s Ming-Image-0.1-Design is a 6B MIT-licensed model that ranks as the top open model for UI/UX design. Architecture trends : @eliebakouch compares four efficient designs: DeepSeek V4.1 Flash and MiMo V3 use YOCO. Qwen 3.8 Next Flash and GLM 5.3 Flash use 3:1 interleaving of sparse and linear attention. All four use Muon, mHC or gated residuals, and partial or no RoPE. TPU megakernel : Inferact open-sourced a TPU megakernel for Kimi K3 that reaches 709 tok/s versus 450 on a GB200 baselin

匹配字段正文

tion. When the first chips came back from the foundry in May, the team pointed its internal AI models at designing software to run benchmarks such as SemiAnalysis’s InferenceX. On DeepSeek’s multi-head latent attention kernel benchmark, performance climbed from 0.31 percent of the theoretical ceiling (set by the chip’s compute and memory bandwidth) to 88.94 percent in roughly 40 hours. Ho says this result is repeatable, so the time between when foundries deliver the first chips and when producti

匹配字段正文

rs might have a performance edge. But the test for Mistral is whether enterprises will see that added control as valuable enough to influence how they choose AI providers. Related:DeepSeek and what enterprises should learn from AI agent exploits Mistral's open-weight approach has been central to its pitch to customers that want more flexibility in how they deploy and customize their AI systems. Open-weight models are especially important for companies that want to keep sensitive workloads under

匹配字段正文

showing a surge in open models released by Chinese labs (below) like Alibaba with Qwen (which the report said had become the “community’s base model”), Moonshot AI with Kimi, and DeepSeek. Meanwhile, US companies AMD and Nvidia published the most new models. (below) I Another interesting trend: AI agents have become the number one user on Hugging Face Hub (below), which the company said could significantly change the numbers in the next report. In an interview with journalists, Delangue said he

匹配字段正文

undown: Nvidia joins SpaceX’s orbital AI build Thomson Reuters builds its own legal AI model Build a reusable AI design system with Open Design Chinese hackers double attacks with DeepSeek LATEST DEVELOPMENTS Image source: Nvidia The Rundown: SpaceX just announced it will build its space-based Starmind data centers around Nvidia’s Vera Rubin NVL72 rack, with Elon Musk revealing new details on a slimmed-down space version of the system he wants in orbit by Q4 next year. The details: Each rack pac

匹配字段正文

Systematic Study of QKV Variants Explore and manage experiment logs in a web interface Explore code, traces, and workspace with an interactive logbook More Articles from our Blog DeepSeek-V4: a million-token context that agents can actually use 60 April 24, 2026 agents smolagents nlp Our Transformers Code Agent beats the GAIA benchmark 🏅 101 July 1, 2024 Community It was interesting to reproduce one of the ICML accepted papers using these tools. This analysis from the hackathon was particularly

匹配字段正文

55X GPU accelerator and made some vaguer claims about performance at the rack level compared to Vera-Rubin racks coming later this year. Here is the inference throughput using the DeepSeek V4 Flash model: As with other GPUs and XPUs, the increase in capacity across all fronts in the MI455X device yields higher token throughput, and particularly at the higher interactivity levels where users need more tokens processed. Here is how AMD reckons the cost per token on DeepSeek V4 Flash on the two dif

匹配字段正文

the drive to solve very difficult problems once reserved for proprietary models. Ollama’s cloud is the most delightful way to access the most powerful open models: GLM, Nemotron, DeepSeek, Kimi, MiniMax and more. On average, Ollama’s cloud has more than doubled in token volume every month, and we can’t wait to make it available to every team. Today, we’re announcing that Ollama has raised $88M from Peter Fenton at Benchmark, Tomasz Tunguz at Theory Ventures, Alex Kolicich at 8VC, along with Sol

匹配字段正文

ing ownership of their models. They're fine-tuning with their own data, running RL, and tuning inference for their own latency, throughput, and cost needs. Open-weight models from DeepSeek, Qwen and others have reached production quality, and inference engines like vLLM and SGLang have matured alongside them. For the first time, the full stack to own and serve your models is there, without sacrificing capability. “Modal powers both our reinforcement learning infrastructure and production inferen

匹配字段正文