
IT之家 10 月 21 日消息,据《华尔街日报》当地时间周二报道,Anthropic 上半年净亏损超过 800 亿美元 (IT之家注:现汇率约合 5,354.93 亿元人民币) ,同时释放信号:未来数年可能投入高达 7,500 亿美元 (现汇率约合 5.02 万亿元人民币) ,用于扩充算力资源。 报道援引一份首次公开募股(IPO)招股书草案称,去年同期该公…
AI 点评 · AI巨头烧钱规模远超营收,亏损与天量算力投入并存,IPO估值逻辑正遭根本性质疑。
共 386 条相关资讯 · 来自历史归档

IT之家 10 月 21 日消息,据《华尔街日报》当地时间周二报道,Anthropic 上半年净亏损超过 800 亿美元 (IT之家注:现汇率约合 5,354.93 亿元人民币) ,同时释放信号:未来数年可能投入高达 7,500 亿美元 (现汇率约合 5.02 万亿元人民币) ,用于扩充算力资源。 报道援引一份首次公开募股(IPO)招股书草案称,去年同期该公…
AI 点评 · AI巨头烧钱规模远超营收,亏损与天量算力投入并存,IPO估值逻辑正遭根本性质疑。

IT之家 10 月 21 日消息,埃隆 · 马斯克(Elon Musk)在当地时间周二的 SpaceX 电话会议上表示,SpaceX 与特斯拉合资成立的芯片制造企业 Terafab 计划在 2028 年第三季度实现量产。 谈及这份时间表时,马斯克称:“我给这个目标预留了很大的误差区间,因为我们不确定最终能否如期达成。” 该项目计划每年产出 1 太瓦算力,供给…
AI 点评 · 马斯克亲自承认时间表存疑,2028量产目标更像愿景,自研芯片能否支撑特斯拉与SpaceX算力野心值得
From GPU trading to physical settlement, independent markets, and programmable compute
AI 点评 · 算力金融化从期货交易走向实物交割,独立市场与可编程调度重塑数据中心定价权。
Google and Meta open-sourced TorchTPU, a native PyTorch backend for Cloud TPUs. It runs eager by default, compiles through torch.compile and XLA, and replaces PyTorch/XLA for new w…

IT之家 10 月 20 日消息,综合彭博社与路透社报道, 人工智能初创企业 Anthropic 正与数据中心开发商 Nscale 就一笔新的云算力租赁合同进行洽谈 。 ▲ Nscale 的 Monarch 计算园区 Nscale 正在积极建设其位于美国西弗吉尼亚州 Mason County 的 Monarch 计算园区。该设施远期规划容量达到 8GW,首期…
AI 点评 · 算力军备竞赛再升级,巨额长约折射头部模型公司对基础设施的焦虑与押注。

IT之家 10 月 20 日消息,一款全新的 AMD Gainsborough APU 疑似出现在 OpenBenchmark 平台的泄露信息中,被认为有可能是 Valve Steam Deck 2 的备选芯片。这份工程样片信息显示,这款 APU 的 CPU 拥有 6 核 12 线程,相比当前 Steam Deck 搭载的 APU 核心数增加了 50%,基础…
AI 点评 · 核心数增五成,掌机性能代际跃升或重塑便携游戏格局。

IT之家 10 月 20 日消息,当地时间 10 月 19 日,亚马逊宣布推出开源工具 TorchNeuron,借助该工具,开发人员编写的 PyTorch 代码可以在亚马逊的自研 Trainium 芯片上本地运行, 只需更改一次配置,无需重写 。 IT之家获悉,TorchNeuron 将于当地时间 10 月 20 日在加州圣何塞举行的北美 PyTorch 大…
AI 点评 · 亚马逊用开源工具降低PyTorch迁移门槛,直击CUDA生态护城河,自研芯片竞争进入软件层。

“IT早报”时间,大家好,现在是 2026 年 10 月 20 日星期二,今天的重要科技资讯有: 1. 苹果官宣北京时间 10 月 29 日举行特别活动,有望推出 OLED MacBook Pro / iPad mini 8 等机型 苹果已确认发布会时间,官方将于美国当地东部时间 10 月 28 日下午 1 时(北京时间 10 月 29 日凌晨 1:00)在…
AI 点评 · 苹果新品发布与特斯拉、小米、豆包同日回应,多条线索交织,科技圈动向密集值得关注。

Amazon Bedrock now supports structured outputs for all Anthropic Claude models on both the bedrock-runtime and bedrock-mantle endpoints. Receive schema-compliant JSON that conforms…
AI 点评 · 结构化输出让Claude在Bedrock上直接产出合规JSON,省去解析校验环节,企业集成更稳。
LLM agents that compute the right answer can still answer wrongly. Agent harnesses increasingly keep long tool outputs out of the context window, saving them to files that the agent searches or reads…
同步搭建千比特规模量子算力云平台
六年蛰伏出货80万颗
NVIDIA has released the full Run:ai GPU orchestration platform as open source under Apache 2.0. The release includes KAI Scheduler, CUDA 13.4 GPU fractioning, quota management, and…

IT之家 10 月 19 日消息,据央视财经援引英国《金融时报》16 日报道, 美国人工智能企业 Anthropic 正在以超过 10% 的收益率 ,吸引黑石集团为其芯片交易提供融资。 这笔总额为 600 亿美元 (IT之家注:现汇率约合 4,021.75 亿元人民币) 的债务融资分为两部分: 一是约 420 亿美元(现汇率约合 2,815.22 亿元人民币…
AI 点评 · 算力军备竞赛推高融资成本,AI巨头借钱买芯片的隐忧值得警惕。

IT之家 10 月 18 日消息,据科技媒体 TechSpot 今天报道,索尼最近申请一项专利,描绘以下场景:让 AI 在 GPU 完成渲染前预测下一帧画面,减少游戏画面卡顿,同时避免帧生成技术带来的额外延迟。 IT之家从原报道获悉,这项技术能够解决帧生成的缺点,让显示画面不再等待两帧画面完成渲染,才能生成插入其中的中间帧。 而是直接利用已经生成的画面预测下…
AI 点评 · AI抢在GPU前预测画面,或改写游戏帧生成规则,降低延迟值得期待。

IT之家 10 月 18 日消息,据 Tweaktown 报道,一名韩国半导体工程师指控,特斯拉在与三星签订芯片代工协议时,试图争取极为优厚的条款:既要获取敏感的制造工艺信息、审计权限,同时还要求压低报价。这位在 X 平台账号 @TTHUEONE 发文的工程师还称,特斯拉的 Terafab 项目从三星位于得州泰勒的工厂挖走了大量员工。 这些指控出现前,埃隆…
AI 点评 · 特斯拉借大客户地位挤压三星,暴露AI芯片供应链权力失衡与人才争夺暗战。
IT之家 10 月 18 日消息,据韩媒 The Elec 报道,美光正在推进面向物理 AI 的专用存储芯片 —— 边缘推理内存(Edge Inference Memory,EIM)的研发。 10 月 16 日,在首尔江南 COEX 举行的韩国半导体展(SEDEX 2026)专题论坛“驱动人形机器人的半导体”上,美光高管(常务)李哲(音译)表示:“自动驾驶、…
AI 点评 · 存储巨头切入边缘推理内存,瞄准物理AI落地瓶颈,或重塑端侧算力竞争格局。

Deepseek has released V4.1-Flash-1017 as an open-weight model that posts better benchmarks, finishes tasks faster, and produces shorter outputs than its predecessor. The company is…
AI 点评 · 小模型在智能体任务上反超更大对手且大幅降本,开源权重或改写性价比竞争格局。

IT之家 10 月 17 日消息,据韩媒 THE ELEC 报道,Intel(英特尔)高级副总裁 조민성 昨日在韩国举行的“半导体驱动人形机器人”大会上表示,该企业计划 2027 年初面向物理 AI 领域推出 "Nova Lake" 处理器。 而在 "Nova Lake" 后, 英特尔还将在 2028 年为物理 AI 带来 "Razor Lake AX" 。…
AI 点评 · 英特尔把机器人算力路线图排到2028年,意在抢占物理AI芯片主动权,与英伟达正面交锋。
Speculative decoding typically characterizes speedup by draft tokens accepted per verification round. Lossless verification preserves the target model's output distribution and thus expected completio…
Counterfeiters are getting better at replicating brands' marks and passing off dupes, known as "superfakes," as genuine items. A startup called Veritas wants to tackle the problem…
A decision library for running local LLMs on multi-GPU machines without NVLink: parallelism, communication, quantization, speculative decoding, KV cache. Every…

IT之家 10 月 16 日消息,据《商业内幕》15 日报道,在英伟达成为 AI 巨头之前,CEO 黄仁勋曾为联系台积电费尽周折:接连打电话,却始终没人回电。 在 14 日发布的一段播客访谈中,黄仁勋讲述了自己最初如何联系上台积电创始人张忠谋。上世纪 90 年代末,英伟达需要寻找代工厂生产图形芯片。当时芯片代工产能紧张,部分厂商甚至要求客户提供技术授权,才肯…
AI 点评 · 黄仁勋早年求台积电无人理,印证芯片供应链权力变迁,如今巨头反成座上宾。
AI 点评 · 中国AI公司在算力受限下仍坚持突破,这种韧性才是真正值得关注的竞争力。

IT之家 10 月 15 日消息,据路透社报道,全球最大的先进人工智能芯片制造商台积电将于周四发布财报,受科技行业基建投入热潮推动,公司有望连续第六个季度创下盈利新高。 分析师表示,市场对台积电 AI 芯片产能,以及先进芯片封装技术 CoWoS 的需求依旧旺盛。 这家亚洲市值最高的企业是英伟达与苹果的核心供应商,强劲需求推动其市值再创新高,约 2 万亿美元…
AI 点评 · AI算力需求真实反映在台积电业绩上,连续六季创新高印证产业链景气度。

IT之家 10 月 15 日消息,在伦敦接受彭博社 Emily Chang 采访时,英伟达首席执行官黄仁勋表示,假若其客户被认定正在开发不安全或不负责任的人工智能(AI)软件, 英伟达将断然停止向其供应芯片。 他坦言, 切断供应目前仍是一个“假设性问题” ,且英伟达在采取断供等极端反制措施之前,必然会先与相关客户进行深入地沟通与对话。 黄仁勋补充道:“我们并…
AI 点评 · 芯片巨头首次以断供为筹码给AI安全划红线,算力霸权正演变为伦理执法权。

IT之家 10 月 15 日消息, OpenAI 基础设施与算力负责人萨钦 · 卡蒂(Sachin Katti)昨日(10 月 14 日)在 X 平台宣布辞职,其空缺将由乌代 · 鲁达拉朱(Uday Ruddarraju)接替。 知情人士透露,卡蒂离职后预计将 继续转任顾问 ,与公司保持合作。 IT之家查询公开资料,卡蒂于 2025 年 11 月由英特尔公司…
AI 点评 · 核心算力操盘手一年即走,星际之门推进或遇变数。
Reusing the attention KV cache across requests is a standard way to cut prefill cost in long-context LLM serving, and production systems increasingly back the per-engine cache with a remote tier share…
Modern generative models produce high-quality individual outputs but struggle to sample diversely, even when explicitly instructed to do so. We introduce MINE-SAMPLE-RENDER (MSR), a test-time approach…
CMOS标准工艺兼容的1TnR技术路线加速研发。

Google has made Gemini 4 Argon generally available on the Gemini API, AI Studio and the Gemini app. The sparse MoE model writes up to 1M output tokens and leads the Vals Index at 6…

“IT早报”时间,大家好,现在是 2026 年 10 月 14 日星期三,今天的重要科技资讯有: 1. 华为首次公布麒麟 9050 Pro 芯片详细架构图:在中国半导体业界支持下成功量产,未来属于开放合作 华为半导体业务无线终端芯片产品管理部部长徐松余表示,首款逻辑折叠韬芯片 —— 麒麟 9050 系列在中国半导体产业界的支持下成功量产,但韬路径未来的发展还…
AI 点评 · 华为公开麒麟9050 Pro架构图,罕见披露国产先进制程进展,是观察中国芯片自主化真实水平的关键信号
Max Hodak, the former president and co-founder of Neuralink, detailed onstage at TechCrunch Disrupt how his company's work with PRIMA is part of its path toward a novel brain-compu…
AI 点评 · 视网膜芯片首次落地遗传性失明患者,脑机接口正从运动修复迈向感官重建。
Existing benchmarks measure solve rate and token output/cost, but don't punish excess code, which makes a codebase harder to work with in the long term. I wanted to measure this si…
AI 点评 · 首次量化AI写代码的冗余度,直击长期可维护性痛点。
Generating information-rich visual artifacts such as slides, posters, and dashboards as code is appealing because the outputs are structured, editable, and verifiable. Yet code models cannot see how t…
Large language models (LLMs) generally rely on sampling temperature to regulate the amount of randomness in their outputs. Mainstream methods for selecting temperature typically involve fixed, heurist…
算力、模型、行业,三层生意一起做

IT之家 10 月 13 日现场消息,华为今日(10 月 13 日)在上海练秋湖举办 2026 麒麟媒体沟通会,华为半导体业务无线终端芯片产品管理部部长徐松余发表演讲。 徐松余谈到了麒麟 9050 系列的一些性能规格, 并首次公布麒麟 9050 Pro 芯片详细架构图 (发布会上为简要版本)。 IT之家注意到,该芯片搭载新一代灵犀 CPU、新一代马良 GPU…
AI 点评 · 国产先进制程与自研架构真正落地,量产即证明供应链已能协同突破,开放合作才是破局关键。

IT之家 10 月 13 日消息, 微软于当地时间周一发布了新一代自研 AI 芯片 ,并计划在明年部署超百万颗该芯片,用于支撑自身内部 AI 业务。 这款芯片名为 Maia 300,是微软第三代自研芯片,专门用于 AI 数据运算,例如代码生成、代表用户执行复杂商业任务。上一代 Maia 芯片已经用于运行微软自研模型,以及微软 Azure 云平台上承载的 Op…
AI 点评 · 自研芯片规模化部署,标志云巨头正以定制算力摆脱对英伟达依赖,重塑AI成本格局。

IT之家 10 月 13 日现场消息,华为今日(10 月 13 日)在上海练秋湖举办 2026 麒麟媒体沟通会,华为半导体业务无线终端芯片产品管理部部长徐松余发表演讲。 徐松余介绍, 麒麟 9050 Pro 搭载 9 核 16 线程灵犀 CPU ,搭配鸿蒙应用并行化优化,实现用户典型场景卡顿降低 31%。 麒麟 9050 Pro 的马良 GPU 采用第四代…
AI 点评 · 华为麒麟芯片强势回归,GPU与影像ISP双双突破,国产半导体自研实力再获验证。

IT之家 10 月 13 日现场消息,华为今日(10 月 13 日)在上海练秋湖举办 2026 麒麟媒体沟通会,华为半导体业务无线终端芯片产品管理部部长徐松余发表演讲。 徐松余谈到了华为最新的韬定律和逻辑折叠路线,韬定律首次应用于麒麟 9050 系列,在性能 / 能效上获得了可观的收益。 徐松余表示,麒麟 9050 系列作为首款逻辑折叠的韬芯片, 开启未来…
AI 点评 · 30B大模型跑进手机端,端侧AI算力门槛被重新定义,华为芯片路线图值得关注。
AI 点评 · 英特尔以480GB显存杀入AI加速器战场,内存容量碾压同级竞品,但生态短板能否补齐才是决胜关键。
Knowledge distillation (KD) frameworks place teacher and student on the same GPUs, so per-GPU memory limits the teachers that fit. DisaggKD separates them across nodes: teacher replicas stream hidden…

Learn how to deploy one large language model across two GPU nodes on Amazon EKS using the multi-node Ray LLM Deep Learning Containers, with separate prefill and decode nodes behind…
Microsoft is announcing its third-generation Maia in-house AI chip today, designed to be the "primary AI inference system" for Microsoft's products and services. Maia 300 not only…

IT之家 10 月 12 日消息,根据 HPE(慧与)向 The Next Platform 等外媒分享的演示文稿,这家企业确认其 ProLiant DL585a Gen13 服务器 支持 NVIDIA(英伟达)的 Rubin PCIe GPU 。 ▲ 引自:The Next Platform NVIDIA 的现有 Rubin GPU 属于数据中心级产品,采…
AI 点评 · 英伟达把数据中心级算力做成标准PCIe插卡,企业升级门槛大幅降低。
The increasing demand for GPU-accelerated machine learning inference has exposed limitations in centralized GPU infrastructures, including high cost, limited elasticity, and long w…

Data from a16z shows a Jevons paradox in the AI market: token prices keep falling, but H100 GPU rental prices hold steady or climb. Cheaper AI drives demand faster than costs drop.…
AI 点评 · AI算力需求不降反升,杰文斯悖论在AI市场应验,英伟达的护城河比想象中更深。
马斯克连横A社,硅谷VC开始囤GPU
AI 点评 · 算力成硬通货,巨头合纵连横抢芯片,AI竞赛进入资源卡位战。
We study pretraining when compute is abundant but unique training data are finite, requiring repeated training on the same data. We first revisit infinite-compute pretraining (Kim et al. 2026) and sho…
A parallel sampler can choose at random which coordinates to generate together. For a target $p$ and the output law $q_U$ of a schedule $U$ drawn afresh for each output, we compare the schedule-averag…

A reference architecture for securely sharing one Amazon SageMaker HyperPod EKS cluster across multiple teams, using AWS IAM Identity Center for authentication, per-team SageMaker…
联想YOGA Pro 15 RTX Spark笔记本电脑,于10月8日9:00正式开启全网盲约。
Microsoft revealed the specs and price for its Surface Laptop Ultra, AI PCs that run on Nvidia chips that are designed to run AI models and agents.
AI 点评 · 英伟达芯片首次进入微软Surface,AI PC从概念走向本地推理实战,值得关注。

OpenAI is rolling out GPT-6 with "Intelligent UI," a feature that turns answers into interactive interfaces with charts, buttons, and forms. The model can now respond while still t…
AI 点评 · AI输出从文字转向可交互界面,标志大模型正从问答工具演化为应用平台。
Liquid AI has released Open d1, two open-weight multimodal models in its d1 decision model family. d1-3B reads text and images. d1-omni-600M reads text with an image, or text with…
AI 点评 · 决策模型不再生成冗长文本,直接输出动作或结果,为智能体与端侧多模态应用开辟新范式。

OpenAI is updating ChatGPT for all users with an “Intelligent UI” that’s more visual—generating interactive elements as part of the chatbot’s outputs.
Like other solutions, it is not especially reliable, and it's easy to circumvent.
AI 点评 · 欧盟用户将默认获得AI水印,但可靠性有限且易规避,凸显监管与技术的现实差距。
Reka has released Rho-1, a 19B omni-reasoning model trained from scratch. One network reads and generates text, images, video and robot actions over a shared KV cache. A distilled…

IT之家 10 月 6 日消息,@央视新闻 昨日(10 月 5 日)发布微博,我国下一代通信关键技术“空心光纤”已率先在宁夏中卫的智算中心间投入商用,传输时延与损耗直降 30% 以上, 在 AI 算力时代给光信号造出一条“磁悬浮”高速轨道。 该光纤直径不足 0.4 毫米,内部嵌入 20 根微型玻璃圆柱,可为光信号提供低损耗、高速传输通道,有望实现超低时延、超…
AI 点评 · 空心光纤商用落地,为AI算力集群间互联降时延三成,或重塑数据中心网络架构。
GPU-parallel simulation provides abundant robot interaction, but existing benchmarks rarely combine this scale with heterogeneous manipulation tasks and standardized multi-task RL evaluation. We intro…

Reka AI's Rho-1 is a 19-billion-parameter omni-model that processes and generates text, images, video, and robot control actions in a single neural network. Trained on 320 H100 GPU…
AI 点评 · 单一模型打通多模态与机器人控制,19B参数实现感知决策一体,值得关注。
An invisible, machine-readable watermark in text output is rolling out to ChatGPT and Codex, but only for users in the European Union at first. OpenAI says its textGrain watermarki…
AI 点评 · 欧盟先行试水,AI文本溯源迈出关键一步,但水印能否扛住改写仍是未知数。
We study agnostic tomography of pure bosonic Gaussian states: given copies of an arbitrary $n$-mode bosonic state $ρ$, the goal is to output a pure Gaussian state whose infidelity with $ρ$ is at most…

“IT早报”时间,大家好,现在是 2026 年 10 月 5 日星期一,今天的重要科技资讯有: 1. 乐山大佛景区回应网传“掏耳朵”养护作业视频:系 AI 合成,佛耳内并无所谓“杂物” 网传视频显示文保工人给大佛掏耳朵、掏鼻孔,还出现猴子抢安全帽等离谱画面。乐山大佛管委会回应称该视频系他人 AI 生成,并非真实场景。2026 年以来大佛已开展 5 次保养维护…
AI 点评 · AI伪造文保现场视频,暴露深度合成技术滥用风险,平台核实与标识机制亟待跟上。

IT之家 10 月 5 日消息,美国上周逮捕了一名嫌疑人,此人被控走私含有受出口管制的英伟达芯片的高端计算机服务器进入中国。 美国司法部于当地时间上周四发布新闻稿,指控 38 岁的格雷格 · 刘(Greg Lui)使用虚假文件掩盖一批货值超过 3 亿美元 (IT之家注:现汇率约合 20.16 亿元人民币) 的服务器货物;据称他明知这批货物最终将运往中国。 美…
AI 点评 · 芯片管制正从纸面走向刑事追责,此案或成对华出口红线执法的风向标。
Vision-language models (VLMs) are increasingly trained to generate structured outputs like points and bounding boxes that downstream interfaces, agents, and robots can act on, yet safety alignment of…
一种混搭的可能:英特尔继续供先进工艺,即前端用14A;后端再接台积电,来补工厂运营、良率、封装这些能力。
AI 点评 · 马斯克押注中国制造补链,折射全球AI算力供应链正从单点依赖转向混搭协作。
A free, open-source AI infrastructure roadmap: a 9-layer map of the stack, 6 stages with exit criteria, around 200 vetted open-source projects, and the math beh…

IT之家 10 月 4 日消息,Qwen3.8‑27B 是一款性能不错的人工智能模型,但前提是设备拥有足够的显存与系统内存。搭载 M4 Pro 芯片的 MacBook Pro 仅有 24GB 统一内存,在运行该 AI 模型时,预填充速度会因此受到限制。据 Wccftech 报道,有一名用户通过 USB‑C 接口将 iPhone 17 Pro Max 外接至这…
AI 点评 · 手机内存可当Mac显存扩展,苹果生态协同的算力潜力初现端倪。

IT之家 10 月 4 日消息,特斯拉首席执行官埃隆 · 马斯克(Elon Musk)做出说明,该公司的 AI5 芯片不会采用 72GB 的 DDR5 内存,而是改用 96GB 内存。这样做是为了避免只有特斯拉一家为驱动擎天柱(Optimus)人形机器人的这款芯片采用最低 72GB 的配置标准。 马斯克此前曾表示,特斯拉已经把 AI5 与 AI6 芯片的内存…
AI 点评 · 内存加码凸显特斯拉为擎天柱算力铺路,自研芯片正从车端向机器人延伸。
Jev-style decision models return categorical probability distributions over predefined options without generating free-form text, enabling software systems to act on their outputs directly. In this wo…
Test-time training (TTT) lets a model store information in its weights during inference. When the model learns from its own output, however, each update also changes the model that generates the next…
Recent GPU kernel optimization agents retain what they learn in knowledge bases or as distilled skills. Kernel benchmarks score each attempt's implementation for correctness and speed but leave the re…
Most vision-language-action models represent future motion as fixed-rate action chunks, tying temporal resolution and prediction horizon to a fixed output budget. This pointwise representation wastes…
Large language models (LLMs) progressively revise their predictions across Transformer layers, yet we typically observe only the final output, discarding the trajectory through which it is formed. We…

Nvidia’s chip-smuggling problem won’t go away as arrests continue.
AI 点评 · 芯片走私规模惊人,出口管制形同虚设,巨头供应链漏洞正被逐一揭开。

IT之家 10 月 3 日消息,Microsoft(微软)当地时间 1 日通过 DirectX 开发者博客宣布,高级着色器交付 (ASD) 将在本月扩展到高通、英特尔、英伟达硬件上。 由于 AMD 的 "RDNA" 家族已率先支持 ASD,这意味着 该特性即将实现对四大 GPU 制造商平台的全覆盖 。更多用户将可享受到直接下载云端预编译着色器带来的好处:缩短…
Split learning lets a client train a language model on a server without sending its text. The client runs the first layers itself and sends the server only their output, a vector of numbers for each t…
A benchmark score compares a system output against a reference, and methodological attention falls almost entirely on the first term. We measure the second. The retained annotation record of a six-lan…
Google launched its first advanced chip into orbit to pave the way for space data centers.
AI 点评 · 谷歌亲自验证太空算力,千次发射成本才是真正门槛。

... but you can’t try it yet unless you are “government users and trusted cyber defenders in the Fairwind Program”
AI 点评 · 百万上下文输出加政府优先准入,谷歌正把最强模型变成地缘博弈筹码。

IT之家 10 月 1 日消息,Synopsys(新思科技)当地时间 30 日宣布与 OpenAI 建立战略合作伙伴关系。两家公司签署了一份范围广泛的多年期协议, 共同开发 GPT-Synopsys 模型 。 GPT-Synopsys 结合了 OpenAI 的先进人工智能能力、Synopsys 的 EDA 工具和代理式人工智能能力, 可利用 Synopsys…
AI 点评 · AI正从辅助工具切入芯片设计核心,或将重塑EDA行业格局与研发效率。
Generalization in large language models (LLMs) is the ability to produce consistent and semantically stable outputs when the same input is expressed in different ways. Existing work typically evaluate…
On-policy distillation (OPD) trains a student on the responses it generates. Existing LLM multi-teacher OPD transfers what specialists predict through their output distributions. We introduce Latent-M…
As LLM agents undertake increasingly complex, long-horizon tasks, verifying their outputs becomes increasingly challenging. We study how verification capability can be strengthened with a fixed base m…

OpenAI and Synopsys are building GPT-Synopsys, a specialized AI model for chip design. It's meant to operate Synopsys' EDA tools like a seasoned engineer and optimize designs on it…
AI 点评 · AI开始插手芯片设计核心流程,若真能替代工程师经验,半导体研发效率或迎质变。
AI 点评 · 全双工多模态交互让Agent像真人对话般自然,昇腾原生亲和则意味着国产算力生态正加速补齐Agent基

IT之家 9 月 30 日消息,据科技媒体 Sammobile 今天报道,AMD 首席执行官苏姿丰预计将在下个月前往韩国,与三星电子高层进行会谈。 ▲ 苏姿丰:图源:AMD IT之家从原报道获悉,苏姿丰本轮预计将谈论 AI 芯片、技术合作等内容,三星有望成为 AMD HBM 4 供应商。作为参考,韩国拥有许多先进半导体生产基地,全球最大的存储芯片供应商 SK…
AI 点评 · AMD若牵手三星HBM4,将打破SK海力士独供格局,AI芯片供应链或迎变局。
AI 点评 · 国产大模型主动适配国产芯片,软硬协同破局生态短板,自主算力闭环再进一步。
Controlling the output of Large Language Models (LLMs) is a central challenge for their reliable deployment, yet a clear understanding of the involved trade-offs remains elusive. C…
Delegating complete application development to coding agents requires preserving the intended design rather than simply producing plausible outputs through naive prompting. Game development provides a…
Autoregressive (AR) grounding models serialize spatial predictions, introducing sequential latency and imposing a causal order on output tokens. We view grounding as visual evidence extraction: object…
AI 点评 · 智能体爆发催生安全隔离刚需,沙箱或成算力新入口。
Meet GPT-6.1 Sol: near-Astra intelligence for coding, computer use, and professional work at one-fifth of Astra’s standard API input and output token prices.
AI 点评 · 端侧大模型上车成刚需,芯片厂商抢滩车规级AI,谁先跑通算力与功耗平衡,谁就握住智能座舱下一轮话语权。
AI 点评 · 成本下降反而凸显延迟与隐私价值,边缘AI的战场正从省钱转向体验。
Distributed inference depends on GPU collective communication that must keep pace with evolving hardware and specialized workloads. However, existing collective implementations often couple semantics,…
Reproducibility is essential for scientific research, yet prior work shows that LLM outputs vary with hardware and batching. We identify an overlooked factor: the hidden injection of the current date…

China is reportedly mulling letting ByteDance, Alibaba buy banned Nvidia chips.
AI 点评 · 中国或松绑英伟达芯片采购,牵动美企游说与地缘博弈走向。
Pixel diffusion models generate RGB images directly, avoiding the bottleneck of an autoencoder, yet their outputs still systematically underrepresent fine-scale natural-image statistics. We show that…
Reproducibility is essential for scientific research, yet prior work shows that LLM outputs vary with hardware and batching. We identify an overlooked factor: the hidden injection of the current date…
On-policy distillation (OPD) improves language-model by supervising a student on prefixes sampled from its own policy, but its supervision remains confined to the output space. Consequently, OPD trans…

Anthropic has released Claude Sonnet 5.5, the second model in its Claude 5.5 family. It generates output more than 30 percent faster, costs up to 30 percent less per task, and near…

Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe no…
Large language models (LLMs) excel at structured, verifiable tasks, but their low-entropy bias can produce homogeneous and predictable outputs, limiting their utility for open-ended scientific ideatio…
Reinforcement learning (RL) post-training for large language models (LLMs) coordinates multiple models across generation, inference, and training on GPU clusters. Several factors may change during a r…
As large language models are deployed in increasingly autonomous long-horizon tasks, manually auditing and verifying the actions, artifacts, and outputs of models becomes more difficult. Users instead…
Video-language models are increasingly used as judges of video understanding, both for evaluating model outputs and for training reward models. Whether their judgments remain reliable when the evidenc…
A 3D scene reconstructed from a single image is most useful when represented not as a rendering or a fixed 3D output, but as an explicit scene program whose execution yields a scene that can be inspec…
Host-resident layer streaming enables full-parameter LLM fine-tuning beyond GPU memory, but data-parallel ranks compete for shared host resources. Replicated transfers amplify traffic, while strong sc…
Multi-step matching models, including flow and diffusion models, produce high-quality outputs but incur substantial inference costs and may reproduce unwanted components of their training datasets. We…
High-performance GPU kernels are essential to modern machine learning systems, yet automatically generating kernels that are both correct and efficient remains challenging. Existing LLM-based approach…
We found an approach to get Jev-like properties from standard LLMs like GLM-5.3-Flash. The core idea is to craft the input prompt so that the first output token answers the questio…
GitHub现在最火热的大模型开源小蜂鸟Colibrì是个啥?
vLLM人马创业公司团队出品

Google Deepmind researcher Robert O'Callahan has quit, saying AI's "current rate of change is far too high." He worked on chip design tools that helped make AI cheaper and faster,…
AI 点评 · 顶尖人才接连出走,说明连内部人都对失控加速深感不安,这比外部批评更值得警惕。
Sony and Universal Music Group filed yet another suit against Suno. The labels claim its new v6 model still infringes on their copyrights because it's trained on user outputs from…
At TechCrunch Disrupt 2026, Ricursive Intelligence co-founders Anna Goldie and Azalia Mirhoseini will take the Disrupt Stage to discuss closing the loop between AI and chip develop…
大厂造芯,正在从交付芯片,走向更广泛的开放共建阶段。
IT之家 9 月 25 日现场消息,在夏威夷举办的 2026 骁龙峰会第二日活动中,高通技术公司先进 AI 软件与平台执行副总裁克里斯 · 拉特纳(Chris Lattner)发表主题演讲, 称 AI 行业的竞争正逐渐从“拼芯片”转向“拼生态”。 在本次主题演讲中,拉特纳指出 AI 软件往往深度绑定特定硬件:换一家公司芯片,开发者可能就要重新适配、优化和调试…
AI 点评 · 拉特纳主导,高通借Mojo打通跨设备AI生态,开发者或告别硬件绑定。
Large vocabularies make output heads a substantial inference cost in small language models. We propose softmax reparameterization, a post-training method that selects a functionally equivalent output…
Dedicated decision models such as Jev map unstructured language to probability distributions over finite choices, allowing their outputs to directly route requests, select tools, and trigger actions.…
Pre-logit steering adapts a frozen language model to a test-time reward by adding vectors to its final hidden states. Unregularized reward optimization can substantially alter the output distribution…
The AWS WhisperX Deep Learning Container packages Whisper, wav2vec2 forced alignment, and speaker diarization into a GPU-ready image. Learn how to deploy it to Amazon SageMaker AI…

Google's experimental orbital data center will have four TPUs and only run for 15 minutes at a time.

IT之家 9 月 24 日消息,在夏威夷举办的 2026 骁龙峰会上,高通在宣布第六代骁龙 8 至尊版和第六代骁龙 8 超级至尊版芯片后, 今天发布第二代骁龙音频平台至尊版。 这也是高通首款融合优质音频、语音技术、终端侧智能与连接能力的音频平台,助力音频可穿戴设备开启智能体全新体验。 该平台整合骁龙畅听技术、终端侧 AI、安全云连接智能以及超低功耗 Wi‑F…
AI 点评 · 端侧AI与超低功耗融合,让耳机等可穿戴设备真正具备智能体能力,音频硬件正成为AI新入口。
We present LightMIS, a scalable family of ultra-lightweight convolutional networks for 2D binary medical image segmentation without a learned stage-wise decoder. LightMIS aligns the outputs of a five-…
Tool-calling agents produce heterogeneous outputs, interleaving structured tool invocations with user-facing natural language summaries. This output heterogeneity presents a structural failure mode in…
高通发布两款第六代骁龙 8 旗舰芯片,Claude Opus 5.5 与 GPT-6 Sol/Luna 发布等。 查看全文
AI 点评 · 芯片与模型同日密集更新,硬件与AI双线竞速,行业节奏明显加快。
Qualcomm said that its new top chip can run 30B mixture-of-expert model locally.
AI 点评 · 端侧跑通三百亿参数模型,手机AI算力竞赛进入新阶段。
Agents using the Model Context Protocol (MCP) rely on semantic matching to select tools from third-party servers, exposing a semantic supply-chain risk through attacker-controlled metadata and outputs…
Typed decision models are built for settings where model outputs are consumed directly by software. Instead of generating free-form text, they return a decision over a predefined set of options. By co…
CPU与GPU趋近1∶1

IT之家 9 月 22 日消息,在昨日的 2026 人工智能计算大会(AICC2026)上,浪潮信息宣布推出元脑 SD200 Ultra 超节点 AI 服务器与元脑 HC2000 多元算力机组。 据官方介绍,元脑 SD200 Ultra 基于国产 AI 芯片打造,单机可承载运行 2.8 万亿参数的 Kimi K3 大模型。 浪潮表示,SD200 Ultra…
AI 点评 · 国产芯片单机跑通万亿级大模型且时延压到毫秒级,推理效率的突破比参数本身更关键。
As agents are deployed with increased autonomy, even extremely rare events along their stochastic output trajectories can occur and prove catastrophic. Safe deployment therefore does not depend on whe…
AI 点评 · 算力从训练转向推理,智能体成新基建核心,增长近九成预示产业落地加速。
范式成为《中国AI算力管理平台技术能力评估,2026》综合评分位列第一的厂商。

Trump announces an "AI Force" modeled after the Space Force and plans to appoint an "AI czar" with a "high IQ." He claims AI could reach 25 percent of US economic output, rejects n…

IT之家 9 月 20 日消息,苹果 M6 芯片近日出现在 Geekbench 跑分数据库中,其 12 核 CPU 分别录得 4071 分、22783 分的单核 / 多核成绩,已经能够在多核方面追平 M3 Max。 如今,苹果 M6 的 GPU 跑分成绩已经出现在 Geekbench 平台。在 Metal 基准测试中, 该芯片的 12 核 GPU 录得 92…
AI 点评 · 苹果芯片迭代提速,GPU稳步升级,自研图形性能正持续缩小与独显差距。
AI 点评 · 算力瓶颈正从芯片转向编排调度,这决定智能体规模化的真实上限。
CANN跨过拐点,华为补上软件生态
让开放模型与可负担的算力成为数学研究的共享基石
A selective predictor acts as a safety gate: it returns an output only when the prediction appears sufficiently trustworthy. Deployments increasingly require this reliability to be certified at a targ…
Large language models can hold knowledge they do not report. A model may sandbag on a capability evaluation, or answer against what it internally knows, and its outputs alone cannot tell whether it is…
协同布局AI基础设施

Amazon SageMaker HyperPod Inference Gateway is a Kubernetes-native, GPU-aware routing add-on for Amazon EKS. It uses real-time GPU signals to send each inference request to the bes…
Mixture-of-Experts (MoE) scales capacity, but existing designs cannot set three quantities independently. For a single token, participation is how many experts contribute knowledge to its output, exec…
Large language models can hold knowledge they do not report. A model may sandbag on a capability evaluation, or answer against what it internally knows, and its outputs alone cannot tell whether it is…
Generative agents are increasingly used to select and narrate video highlights, but they typically operate over unstructured or frame-level representations. Their output is consequently difficult for…
Huawei is accelerating the launch of its next-generation Ascend 960DT AI chip as it pushes to compete with Nvidia and close China’s AI computing gap with the U.S.

IT之家 9 月 17 日消息,数据分析公司 Silicon Data 最新数据显示,英伟达 B200 GPU(Blackwell 架构)在出货约一年后,二手残值已涨至其原始上市价格的 158%。 换言之,英伟达 B200 当前二手交易价格较 MSRP 溢价约 58%,与消费电子硬件通常逐年贬值的情况形成鲜明对比。 Silicon Data 指出,英伟达更早…
AI 点评 · 算力供不应求颠覆硬件贬值规律,GPU正从消耗品变投资品,折射AI基建热潮的疯狂溢价。

Integrate NVIDIA Resiliency Extension (NVRx) into PyTorch FSDP training on Amazon EKS to overlap checkpoint I/O with training and recover from GPU faults in seconds. This post cove…
AI 点评 · 大模型训练最怕GPU故障中断,秒级恢复加检查点与训练重叠,直接拉高集群可用性。

“IT早报”时间,大家好,现在是 2026 年 9 月 16 日星期三,今天的重要科技资讯有: 1. 联发科天玑 9600 Pro 旗舰芯片发布:2nm 工艺、4.55GHz 双超大核、120 帧光追手游 联发科天玑 9600 Pro 旗舰芯片 9 月 15 日正式发布,定位旗舰 5G 智能体 AI 芯片,口号为“超冷劲,超智慧”。>> 查看详情 2. 苹果…
AI 点评 · 联发科抢先2nm,先进制程竞争白热化;存储涨价与问界主导权变动,折射产业链博弈加剧。
首发全场景倒车能力,体验再进一步
打破分子模拟“不可能三角”
Reliable confidence estimation is increasingly central to the trustworthy deployment of language models: a calibrated estimate of the probability that an output is correct decides what to ship, what t…

The “HardFlow” algorithm could help generative AI models produce high-quality outputs that obey strict requirements when “pretty close” doesn’t cut it.
Native 3D generators turn one image into a single mesh. TRELLIS.2 and its peers deliver high-fidelity non-watertight geometry with materials, but the output is one fused object, while downstream work…
Computer-use agents are increasingly evaluated in realistic desktop environments, but existing benchmarks provide limited coverage of professional engineering workflows whose outputs are persistent, s…
We study a practical question: can a small correction module fix errors in a frozen language model's outputs without degrading its base capabilities? We propose CRN v2, a lightweight logit-level corre…
This practical tutorial demonstrates how to build and accelerate machine learning workflows using NVIDIA cuML and RAPIDS. It covers GPU environment setup, zero-code scikit-learn ac…
AI 点评 · GPU加速机器学习全流程实战,零代码迁移scikit-learn是最大看点。

IT之家 9 月 13 日消息,在 CPU / GPU 跑分曝光之后,苹果 iPhone 18 Pro(对应机型 iPhone19,2)的 GeekBench AI 跑分昨日(9 月 12 日)现身,单精度得分 5,144 分。 IT之家发稿前查询 GeekBench AI 跑分平台,目前共收录了 2 条记录,分别是 Core ML 调用 GPU(通用图形计…
AI 点评 · 苹果自研NPU性能跃升超五成,端侧AI算力竞赛再提速,值得关注。
太初(杭州)集成电路有限公司新一代超智融合计算系统元碁Hypertintellix入选“算力中国·年度卓越成就”。
AI 点评 · 国产芯片以超智融合架构获国家级算力认可,展现自主算力突破新路径。
Few-shot in-context learning (ICL), the capacity of a model to infer abstract patterns from input-output examples provided in its prompt and apply them to new inputs, has been extensively studied in l…
Concurrent local LLM serving on unified-memory desktops must preserve memory headroom and output fidelity, which speed-only rankings overlook. We introduce SiliconBench, which evaluates nine Apple Sil…
Automated, production-grade CI/CD evaluation pipeline for Hugging Face chatbots using DeepEval and dynamic JSON test datasets.
9月9日,墨芯人工智能亮相以"共创AI新经济"为主题的2026 Inclusion·外滩大会。

IT之家 9 月 11 日消息,科技媒体 MacRumors 昨日(9 月 10 日)发布博文,报道称在 Apple Watch Series 12 和 Apple Watch Ultra 4 两款智能手表上, 苹果计划引入“实时回放”(Live Rewind)功能。 体验方面,该功能主要基于苹果 S11 芯片,用户双击数码表冠后,手表会显示此前 15 秒语…
AI 点评 · 苹果把隐私保护做进硬件,短时回放加自动清除,可穿戴转录更实用也更安全。
商务部回应美国 AI 蒸馏指控 谷歌声称为了合规将「降级」欧盟搜索服务 Apple 智能将有用量限制 OpenAI 将与三星合作开发芯片 小岛秀夫新游戏被索尼取消,转由微软发行 DeepSeek V4.1 发布 看看就行的简讯 少数派的近期动态 你可能错过的好文章 查看全文
AI 点评 · 美方拿蒸馏说事,实则剑指中国AI崛起,商务部回应关乎行业出海命运。
NVIDIA has detailed BioNeMo Inference Runtime (BioIR), a Python library that accelerates biomolecular structure-prediction models on NVIDIA GPUs while staying in plain PyTorch. In…
AI 点评 · 英伟达把生物分子预测推理性能再拉高,GPU加速正成为AI制药基础设施的关键竞争点。
Counterfactual regret minimization (CFR) is one of the few large numerical workloads that still runs faster on CPUs than on GPUs. Each iteration sweeps a game tree with up to billions of states in mil…
9月9日,以“JoyAI · 跃迁物理世界”为主题的JDDiscovery-2026京东全球科技探索者大会在北京举行

IT之家 9 月 10 日消息,小米 18 Fold 中折叠手机发布于 9 月 7 日,将于今日 10:00 正式开售, 售价 10999 元起 : 12GB+256GB:10999 元 12GB+512GB:11999 元 16GB+512GB:12999 元 16GB+1TB(LPDDR6):14999 元 16GB+1TB 陶瓷特别版(LPDDR6):…
AI 点评 · 自研芯片加持,折叠屏价格下探,小米能否搅动高端市场值得关注。

“IT早报”时间,大家好,现在是 2026 年 9 月 10 日星期四,今天的重要科技资讯有: 1. 苹果 iPhone 18 Pro 系列发布:新增可变光圈技术、首发 2 纳米制程工艺 A20 Pro 芯片,9999 元起 全新 iPhone 18 Pro 系列提供深、银、冰川蓝和勃艮第酒红色 4 种配色,外观与 iPhone 17 Pro 接近、首发 A…
AI 点评 · 苹果折叠屏与2纳米芯片双首发,叠加DeepSeek IPO动向,一文掌握AI与硬件双线变革。

Anthropic published an economic model with three scenarios for the US economy through 2030. In the extreme scenario, output doubles every 4.5 years and knowledge worker unemploymen…
AI 点评 · 经济模型自设极端假设,或为CEO悲观论调预留缓冲,看点在如何影响AI就业叙事。
Enterprises deploy systems, not checkpoints. Usable capability depends jointly on weights, serving route, precision, output contract, and harness, yet all 18 audited benchmarks score advertised model…

TorchServe is no longer maintained, leaving teams to own the entire GPU inference stack. The AWS Ray Serve Deep Learning Container is a supported, pre-tested container with the fra…

Kepler Computing claims a new approach to chip design—and a proprietary material—can help end the supply bottlenecks that have sent memory prices surging.

IT之家 9 月 9 日消息,在 9 月 7 日的 HarmonyOS 7 | HUAWEI Mate XT 2 及全场景新品发布会上, 华为 Mate XT 2 非凡大师全新展翼三折叠手机正式发布 ,定价 19,999 元起。 这款新机最大亮点之一就是搭载了 首款逻辑折叠 τ 芯片 —— 全新麒麟 9050 Pro 。 华为智能汽车解决方案 BU CEO…
AI 点评 · 六年磨一剑,国产芯片自主突破的标志性时刻。

IT之家 9 月 9 日消息,消息源亚伦 · 佩里斯(Aaron Perris)昨日(9 月 8 日)在 X 平台发布推文,通过挖掘 tvOS 27 Beta 更新代码,发现了 AppleTV18,1 硬件踪迹, 预估是苹果新款 Apple TV 4K 硬件。 芯片方面,根据代码细节,苹果新款 Apple TV 4K 硬件配备 A19 或者 A19 Pro…
AI 点评 · A19芯片加持,Apple TV或补齐AI短板,成家庭智能中枢关键棋子。
Existing bias auditing methods typically rely on model outputs, requiring costly benchmarks or judge models and potentially missing internal shifts that never appear in generated text. We propose a re…
NVIDIA has announced CUDA Rust, a push to make Rust a first-class language for GPU kernels. Two open-source NVlabs projects cover the two CUDA programming models: cuda-oxide compil…

Benchmark two 30B Mixture-of-Experts models, Qwen3-Coder-30B and NVIDIA Nemotron-3-Nano-30B, across G5, G6, G6e, and G7 GPU instances on Amazon SageMaker AI. Compare throughput, la…
近日,物理AI企业深度智控(DeepCtrls)完成新一轮B+轮数亿元融资。

IT之家 9 月 8 日消息,网友 pdfu 昨日(9 月 7 日)在 MacRumors 论坛发帖,通过挖掘 watchOS 27 Beta 8 更新,发现了关于 Apple Watch Series 12 智能手表的更多线索, 并预估苹果公司会带来 4 项升级。 硬件方面,基于 watchOS 27 Beta 8 代码信息,在关联 T8320 芯片的 N…
AI 点评 · 存储翻倍与训练功能升级,预示健康监测向专业力量训练深化。
AI 点评 · 跨国高管失踪案与工厂游热潮并存,折射AI产业全球化暗流与人才流动风险。
AI research agents can propose far more experiments than they can afford to run. Meta FAIR, Oxford and UCL introduce AI Research Preference Models — frozen LLM judges that rank 15…
A transformer can make an attribute linearly decodable in its residual stream at a depth where that attribute does not yet influence the output. This gap between where information is readable and wher…

Retrieval quality in an AI search product is bounded by two things: how good the embedding model is, and how cheaply you can run it across an index. This week, Perplexity Engineeri…
Visual generation is evolving from generative models used through a single invocation into agentic control processes that can plan, select tools, inspect intermediate synthesized outputs, revise failu…
Nous Research has collapsed local model setup into a single click in Hermes Desktop. The app reads your hardware, fit-checks the catalog against your GPU, picks the highest-quality…

IT之家 9 月 5 日消息,NotebookCheck 今日发现,三星电子正在对 Galaxy S27 系列进行关键芯片验证测试。 根据 NotebookCheck 提供的一份高通项目门户网站截图,三星正在测试的芯片型号为 SM8975,该芯片的正式名称为 Snapdragon 8 Elite Extreme Gen 6(第六代骁龙 8 超级至尊版)。 截…
AI 点评 · 芯片迭代提速,安卓旗舰性能竞争白热化,值得关注。
Large language models used for code editing can be trained and deployed in at least two output regimes: direct generation, where the model emits the entire modified file in one shot, and iterative dif…

Deepseek wants to put 160,000 Huawei Ascend-950DT chips into an Inner Mongolia data center for inference only, not training. It would be the largest known Huawei chip cluster. But…
国产卡跑出高品质 AI Token 生产级性能
WeatherNext 3 ingests live geostationary satellite mosaics, refreshes hourly, and outputs 5 km forecasts across Search, Gemini, Maps. The post Google DeepMind’s WeatherNext 3 Train…

IT之家 9 月 4 日消息,根据 NVIDIA(英伟达)官网更新,其面向 Windows on Arm PC 的 RTX Spark N1X 超级芯片处理器在笔记本电脑端提供 2 档配置,而 桌面主机上则仅有“满血”高配 。 可以看到,RTX Spark N1X 面向移动端提供了 5120 核 GPU + 18 核 CPU 的低配, GPU 规模较 614…
AI 点评 · 移动端低配曝光,凸显英伟达细分市场策略,Arm PC竞争再添变数。
Research and news coverage of language-model deception increasingly attributes human-like mental-state concepts to language models. Such claims can blur the distinction between behavior that looks dec…

The long-rumored deal will give the chip giant access to—and help it promote—a huge repository of open-source AI models and data sets.
Perplexity has open sourced Lily, the local inference engine behind Hybrid Compute in Perplexity Computer. Built in Rust with custom Metal kernels for one model on one chip family,…

IT之家 9 月 3 日消息,数码博主 @白饭炒白米饭 于 8 月 30 日发布微博, 通过焊点图,反推苹果 A20 Pro 芯片的规模以及内部结构 ,目前已被 NotebookCheck 等媒体、@VadimYuryev 等海外博主转发。 IT之家附上相关微博内容如下: 根据焊点图,来反推 A20 Pro 芯片的规模以及内部结构,顺便再来聊聊这颗芯片。根据…
AI 点评 · 焊点图反推芯片架构,曝光苹果未发布核心情报,技术含量极高。
Designing and authoring high-performance custom kernels for accelerators is a complex task that requires deep hardware-level expertise. Large Language Models (LLM) can be leveraged together with real-…
Release: llm 0.34 One new feature: llm logs --usage Markdown output now includes the response duration in milliseconds and as a human-readable duration. llm logs --short includes a…
LLMs are trained to generate natural language. However, various strands of evidence indicate that an LLM's externalized linguistic outputs and mechanistically-extracted linguistic features can be an u…

Queue the usual rush of model launches...
大模型推理案头手册(开源版):给定 GPU 算力,一个模型能扛多少 QPS?三层模型 × 三面墙 × 排队论 × 开环压测
Benchmarks for LLM-generated GPU kernels decide correctness with a few random inputs and a loose floating-point tolerance, and their verdicts now feed leaderboards and reinforcement-learning rewards.…
Nvidia is officially launching DLSS 5 this week, following a divisive announcement in March where we likened the AI upscaling tech to a "real-time generative AI filter for video ga…

IT之家 9 月 1 日消息,英伟达与联发科 8 月 31 日宣布加深长期合作,打造涵盖 AI 基础设施、本地 AI 计算和汽车领域的下一代 AI 计算平台。 作为扩展合作的一部分,联发科将采用 NVIDIA NVLink Fusion 平台,为超大规模企业、云服务提供商和前沿模型开发者提供预验证的路径, 开发定制 XPU 并将其引入 NVIDIA NVLi…
AI 点评 · 芯片巨头跨界结盟,AI、PC、汽车三线并进,生态版图再扩张。
Data-residency constraints force enterprises to self-host LLMs, but continuous adoption of newer models without decommissioning their predecessors expands the serving fleet, fragmenting a finite GPU p…
Nvidia invests $3.5 billion into Taiwanese chipmaker MediaTek. The deal shows how Nvidia plans to stay essential to AI infrastructure as Big Tech begins to build its own AI chips.
什么样的AI业务,英伟达GPU和谷歌TPU搞不定,非得用Mac

IT之家 8 月 30 日消息,博主 @熊猫很禿然 近日爆料,小米 18 Fold 折叠屏手机搭载 6000mAh 电,支持 67W 有线快充 +50W 无线快充,屏幕为 7.6 英寸左右。 IT之家注意到,在 8 月 24 日举行的小米玄戒芯片技术沟通会上,小米正式发布了玄戒 O3 芯片,同时宣布小米 18 Fold 折叠屏手机将首发玄戒 O3 芯片, 将…
AI 点评 · 电池与快充双升级,折叠屏续航痛点有望缓解,玄戒芯片首发更添看点。

IT之家 8 月 29 日消息,印度知名消费电子品牌 Mivi 近日官宣跨界进军智能手机市场,首款产品 9 月发布。 Mivi 官方在 X 平台透露, 公司将于 9 月在印度推出一款 5G 手机 。从官方发布的预热图片来看,这款手机采用绿色配色,背部为平面设计,边框采用平直造型,搭载骁龙芯片。 据科技媒体 91Mobiles 报道,这款智能手机的售价将低于…
AI 点评 · 新兴品牌跨界搅局,低价5G能否撼动印度市场格局,值得关注。
The new generation of data center systems is increasing efficiency with smarter traffic control instead of just more processor cycles.
AI 点评 · 芯片巨头转向系统级优化,智能调度成新战场,算力效率比拼升级。

IT之家 8 月 29 日消息,据央视新闻今日报道,AI 与铜的距离看起来很远。一边是发生在“云端”的前沿算法,一边是深埋地下的传统金属。可当 AI“跑”起来,它们就被紧紧连在了一起。服务器的高速运算,需要更强的供电、更快的数据传输和更高效的散热,在这些环节中,铜至关重要。据市场测算,2026 年 AI 服务器专用高端铜箔需求约 2.4 万吨, 同比增长约…
AI 点评 · 算力军备竞赛传导至资源端,铜箔缺口预示AI硬件供应链新投资逻辑。
We investigate whether automatic speech recognition (ASR) errors in user input can lead to unsafe outputs from Embodied AI (EAI) models. We find that ASR errors can lead to harmful instructions being…

IT之家 8 月 28 日消息,科技媒体 Linuxiac 昨日(8 月 27 日)发布博文,报道称开源操作系统项目 Asahi Linux 官方宣布, 针对苹果 M3 系列 Mac 电脑的原生 Linux 支持已步入最后冲刺阶段,核心驱动代码已被正式并入 Linux 7.2 内核主线周期。 该媒体指出苹果 M3 系列芯片采用了高度定制化的电源管理、显示控制…
AI 点评 · 底层驱动突破,M3跑Linux不再是梦,开源生态再进一步。
IT之家 8 月 28 日消息,据蓝鲸汽车报道,在理想汽车 2026 年 Q2 业绩电话会上,理想汽车 CEO 李想表示,今年下半年起, 理想品牌的电池将会陆续搭载到全系车型 。 需要强调的一点是, 自研并不代表供应商的产品不够好 。我们自研马赫芯片,并不影响英伟达仍然是全世界最好的芯片品牌。我们自研电池,也不影响宁德在内的品牌是全世界最好的电池品牌。 李想…
AI 点评 · 自研与供应商并行,理想电池策略透露行业新趋势。
国产算力跑通视频生成规模化应用
AI 点评 · 国产算力突破视频生成瓶颈,规模化落地价值凸显。
国产异构助力前沿智能进入普惠时代
AI 点评 · 国产算力与前沿模型强强联合,开创AI普惠新路径。

Serving automatic speech recognition (ASR) models at scale is costly when each request uses only a fraction of a GPU. Learn how NVIDIA CUDA Multi-Process Service (MPS) with NVIDIA…
Nvidia has reportedly agreed to buy Hugging Face, the popular open source AI hub, for $12.9 billion in a move that would let Nvidia both protect its chip empire and jump back into…

IT之家 8 月 27 日消息,今日网易有道将正式发布 OpenPods 有道 AI 耳机,即首款专为 iPhone 用户打造的 Agent 耳机,并同步开启预售,首发补贴后到手价 1499 元,将于 9 月 10 日正式发售,提供黑色、白色两款配色。 IT之家注意到,OpenPods 的核心差异化在于其“Agent 耳机”定位。产品采用苹果授权芯片,获得…
AI 点评 · 苹果生态专属AI耳机,1499元切入Agent硬件赛道,能否复制AirPods神话?
IT之家 8 月 27 日消息,金融媒体 marketwatch 今天(8 月 27 日)发布博文,报道称在英伟达最新财报电话会议上,公司 CEO 黄仁勋回应 OpenAI 首款自研推理芯片 Jalapeño, 表示英伟达正在打造的东西与 OpenAI 的自研芯片“截然不同”。 IT之家此前报道, OpenAI 公开首款定制 AI 推理芯片 Jalapeño…
AI 点评 · 芯片巨头正面回应客户自研,生态壁垒与定制化趋势的博弈看点十足。

IT之家 8 月 27 日消息,Perplexity AI 当地时间 25 日宣布推出 Portable Computer。这是其 Perplexity Computer 多模型编排智能体数字员工的本地化版本 ,可实现私密的工作流,仅在需要时才调用云端算力。 IT之家了解到,Portable Computer 搭载 Qwen 3.8 27B 或经 Perpl…
AI 点评 · 本地化智能体兼顾隐私与云端弹性,或开启企业级AI新范式。
Amazon is adding another 2 million Nvidia GPU chips to its data centers over the next two years. But this extended partnerships stretches beyond buying more chips.
AI 点评 · 云巨头加码英伟达芯片,凸显AI算力军备竞赛白热化,供应链格局或生变。
The fastest way to run Qwen3.8-Flash-Next on Strix Halo (gfx1151)
Generative vision-language models (VLMs) are increasingly used in human-centered settings, yet they can produce demographically biased outputs even when images differ only in controlled attributes suc…
Native visual reasoning treats visual generation as the medium of reasoning itself: visual states (i.e. images and videos) are not merely inputs to be understood or outputs to be rendered, but first-c…

OpenAI showed off "Jalapeño," its first in-house inference chip, with benchmarks at the Hot Chips conference. According to SemiAnalysis tests, the chip beats Nvidia's Blackwell and…
AI 点评 · 自研芯片性能反超英伟达,OpenAI硬件话语权大增,AI算力格局或将生变。
Tested on SemiAnalysis’ InferenceX benchmark, Jalapeño registered both more tokens per user and more throughput per kilowatt than the currently available state-of-the art.
AI 点评 · 推理性能与能效双突破,为规模化AI部署提供新标杆。
OpenAI says its new AI chip, Jalapeño, completes tasks more efficiently and returns responses faster than other AI systems, according to a blog post published on Tuesday. During a…
AI 点评 · 自研芯片直击算力瓶颈,推理速度成新战场,OpenAI硬件布局或重塑AI成本格局。

Nvidia is moving its Groq 3 LPX inference chip into full production and reports 3,400 tokens per second on Gemma 4 31B, four times faster than Cerebras. But the numbers don't tell…
AI 点评 · 数据亮眼但算法存疑,行业竞争焦点转向推理速度的真实性价比。
Jalapeño is a custom inference chip from OpenAI that delivers faster, more power-efficient AI inference, with higher throughput and lower latency for modern models.
A scalable serving system for building and running modular diffusion workflows across GPU clusters.
少数派的近期动态新一季少数派会员启航,更新权益,更多惊喜,还有实体纪念卡。点击了解能让AI助手通过自然语言指令直接与您的Quote/0摘录墨水屏交互的DotSkill已上线。点击了解你可能错过的文章快 ... 查看全文

IT之家 8 月 24 日消息,ACEMAGIC 现已推出新一代迷你主机 F2A,新品主打小巧机身,搭载英特尔 Panther Lake 平台。 IT之家注意到, 这款迷你主机采用英特尔酷睿 Ultra X7 358H 处理器 ,拥有 16 核心 16 线程,其 NPU 的 AI 算力最高可达 50 TOPS,配备 Arc B390 核显,整机 AI 性能最…
AI 点评 · 酷睿Ultra X7加持,50 TOPS算力让迷你主机迈入高性能AI时代。

IT之家 8 月 24 日消息,小米今天(24 日)发布了三款自研芯片,随后在当天晚间登上央视财经频道。小米集团副总裁朱丹在采访中介绍,小米的判断很明确,AI 时代算力是一切智能体验的根基,不掌握芯片能力,就无法从底层定义产品体验。所以这笔投入不是“烧钱”,是在 为下一个十年买“入场券” 。 报道援引业内人士的分析称,国内多家企业持续加码芯片自研,通过长期高…
AI 点评 · 自研芯片成科技巨头分水岭,小米押注底层算力为未来十年卡位。

To generate intelligence at scale, AI factories run continuously, and their economics are defined by delivered output: tokens per second, tokens per watt, cost per token, utilizati…
AI 点评 · 算力经济转向单位产出,XPU架构成AI工厂效率关键。
The next era of AI inference won’t be defined by a single breakthrough chip, network or system. It’ll be defined by how every layer of the AI factory works together. That’s why NVI…
AI 点评 · 英伟达量产新一代推理芯片,聚焦AI工厂全栈协同,标志智能体规模化部署进入新阶段。

Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe no…

Cerebras has introduced its CS-4 AI accelerator, which CEO Andrew Feldman calls the fastest system in the industry. The article Cerebras unveils CS-4 with double the performance on…
定义国产推理算力新范式 赋能万卡级AI推理算力规模化落地
The five largest GPU neoclouds now run on very different models. CoreWeave and Nebius report to the SEC; Lambda and Crusoe are private and heading toward IPOs; Groq rebuilt itself…

IT之家 8 月 23 日消息,据外媒 Interesting Engineering 今天报道,美国初创公司 Starcloud 已完成 2.5 亿美元 (IT之家注:现汇率约合 16.86 亿元人民币) A 轮扩展融资,估值达到 23 亿美元 (现汇率约合 155.09 亿元人民币) 。这笔新资金将用于研发配备高性能 GPU 的卫星,为 AI 提供算力。…
AI 点评 · 英伟达加持下,太空算力布局提速,或成AI基建新赛道风向标。
FreeToken splits MoE cache misses between PCIe fills and CPU execution using measured bandwidths, unlocking frontier models locally The post Meet FreeToken: An Edge-Native MoE Serv…
IT之家 8 月 22 日消息,据界面新闻今天报道,中诚华隆现已推出全新 HL200 推理芯片和超节点智算集群方案。 据报道,该芯片的单卡 FP4 算力为 4P、FP8 算力 2P,FP16/BF16 算力 0.5P, 能效比达 5.12 TFLOPS/W ,并原生支持 FP4、FP8 低精度推理。 而 HL200 超节点智算集群方案单机柜支持 64 张 G…
AI 点评 · 能效比突出,万卡集群扩展力强,国产推理芯片赛道再添有力竞争者。

IT之家 8 月 22 日消息,当贝现已在京东上架 RX10 投影仪,该产品主打灵动云台设计,采用 0.33" DMD 显示芯片,提供 1200CVIA 流明, 定价为 5499 元 。 外观方面,该机采用一体式灵动云台,支持上下 225° 的俯仰角度调节,无需额外垫高即可自由改变投射位置。同时支持 AI 实时校正与全向无感自动梯形校正功能,用户在大幅度转动…
AI 点评 · 云台调节与AI校正结合,或成中端家用投影新标杆。
The five largest GPU neoclouds now run on very different models. CoreWeave and Nebius report to the SEC; Lambda and Crusoe are private and heading toward IPOs; Groq rebuilt itself…

Waymo built its own chip for its robotaxis, cutting its reliance on Nvidia. The article Waymo builds its own chip for its robotaxis, cutting its reliance on Nvidia appeared first o…
Three ~300M drafters bring speculative decoding to LFM2.5, delivering up to 3.18x faster decoding with identical greedy output. The post Liquid AI Releases LFM2.5-DSpark Draft Mode…
AI 点评 · 算力成金融标的,AI基建资本化浪潮开启,市场规则重塑。

IT之家 8 月 20 日消息,据彭博社今日报道,Meta 正通过微软 Azure 云服务采购 AI 服务,年支出高达数亿美元,已成为微软规模最大的 AI 客户之一。知情人士称,Meta 每周通过 Azure 消耗的 AI 算力达到了数万亿词元(Token)。 目前,微软 Azure 正通过 Foundry 平台提供来自多家供应商的 AI 模型,包括 Ope…
https://www.seangoedecke.com/readers-cant-identify-watermark...
Frontier language models are compared, marketed, and benchmarked on capability -- what their best or average output can achieve. I argue this measures the wrong axis. The models have saturated accurac…
Hi HN! I’m Barnaby, founder of machine0 ( https://machine0.io ). I’m building a CLI for long horizon agent compute: `machine0 new mybox` gives your agent a persistent cloud VM, bil…
AI 点评 · 从命令行一键拉起持久化CPU/GPU虚拟机,为长时运行AI代理提供基础设施,直击开发者痛点。
We introduce PTXBench, a benchmark for evaluating and adapting large language models (LLMs) to use architecture-specific PTX for GPU kernel optimization. PTXBench measures functional correctness, whet…
Reinforcement Learning (RL) has been promising in single-turn LLM fine-tuning. However, long-horizon agentic reasoning introduces increasingly branching interactions and sparse rewards, exposing sever…
A language model's output does not by itself provide verifiable evidence about the internal computation that produced it. We study computational provenance: whether generated text can carry detectable…
Regulatory compliance monitoring in deployed language models is increasingly implemented as a legal and audit control, checking model outputs against written rules spanning data protection, healthcare…

Anthropic's text watermarking for Claude is supposed to make AI-generated content detectable. But critics doubt that word choice stays unaffected, and lawyers are facing new transp…
😭(放声大哭)成为最流行 emoji 美国要求合作方在中美 AI 竞赛中「选边站」 ChatGPT 上线使用历史记录功能 美国政府敦促苹果不要购买中国内存芯片 uBlock Origin 放弃维护 Facebook 广告规则 索尼加速向娱乐业务转型 看看就行的简讯 少数派的近期动态 你可能错过的好文章 查看全文

A representative survey by Epoch AI found that 20 percent of employed Americans hand off at least one task to AI that a human used to do. Generally, they accept AI output with litt…
Language-specific competency (LSC) is the phenomenon of a language model performing better or worse depending on the language of the prompt. In other words, a language model outputs different (and pot…

IT之家 8 月 15 日消息,据 Wccftech 今日报道,Reddit 用户“stlk0”通过自制开源工具,将其搭载英特尔酷睿 i9 处理器的 2019 款 MacBook Pro 在 Windows Boot Camp 下的空闲 GPU 功耗从 14W 降至 3W,降幅近 80%。同时,其搭载的 AMD Radeon Pro 5300M 显卡温度也从…
AI 点评 · 民间高手破解老款MacBook发热难题,48Hz刷新率让显卡功耗暴降八成,小技巧解决大痛点。
High-order multiple-input multiple-output (MIMO) detection requires efficient search over a large discrete symbol space while producing reliable soft information for channel decoding. This paper devel…
Vision-language models (VLMs) enable embodied agents to reason and act from visual observations and language instructions. Reinforcement learning (RL) post-training enhances these capabilities using t…

AI agents using Claude Opus 4.8 and GPT-5.6 Sol were given six days, $3,000 in API credits, and GPU access to independently write AI research papers. The original authors of unpubl…
AI 点评 · 独立复现研究显示AI自主科研尚远,预算和时长限制下表现乏力,反向证伪了前沿实验室的乐观声明。

IT之家 8 月 14 日消息,晚点 Auto 今天(14 日)晚间援引多个信源消息称,理想汽车正在探索自研云端推理芯片,将采用与智驾芯片相同的数据流架构,项目仍处早期阶段。 AI 数据中心普遍使用 GPU 完成模型训练和推理,也有数据中心部署云端推理芯片。一位芯片行业人士说,云端推理芯片可以 承接部分原本在 GPU 上运行的推理任务 ,这其中包括 智驾模型…

OpenAI is launching "Ultrafast," a new inference mode that delivers GPT-5.6 Sol at up to 750 output tokens per second, powered by Cerebras hardware from their $10 billion partnersh…
Multi-GPU acceleration for MiniMax H3 video generation on NVIDIA V100 (sm_70). Ulysses sequence parallelism as a drop-in ComfyUI custom node — ~19 min to ~7 min…
少花一半成本,部署效率提升80%!机器人研发可以不用重复造轮子了。
LLM coding agents issue Bash commands through interfaces that may serialize, wrap, and reparse model output. Matched execution scores alone cannot distinguish command-generation errors from failures i…
首颗AI芯片已进入量产
Preview Ultrafast, a new OpenAI API service tier that runs GPT-5.6 Sol up to 14× faster. Powered by Cerebras, it delivers up to 750 output tokens per second.
Transforming multimodal sources into condensed and structured media outputs can be fundamentally conceptualized as a long-horizon agentic process centered on a model-harness system. While an ideal har…
LLM coding agents issue Bash commands through interfaces that may serialize, wrap, and reparse model output. Matched execution scores alone cannot distinguish command-generation errors from failures i…

Researchers at IIT Bombay and Adobe Research have built an inverse language model that reconstructs the original prompt from an LLM's output with near-perfect accuracy. Their metho…
AI 点评 · 逆向还原提示词,直击大模型隐私与安全要害,或将重塑AI审计规则。
AI 点评 · 算力按能力分级配置,打破平均主义,直击AI人才成长痛点。

IT之家 8 月 12 日消息,据央视财经今天(12 日)报道,当前,一个万卡集群一天至少吞掉 20 万度电。据测算,到 2030 年全国算力用电量将达到 8000 亿千瓦时。人工智能的未来,不仅是“算”的比拼,更是“电”的竞赛。 报道指出,算力在狂奔,绿电在托底。截至 6 月底, 我国非化石能源发电装机容量 25.2 亿千瓦,占总装机容量比重的 62.4%…

Running large language model inference at scale forces a KV cache trade-off: oversized GPU instances or slow time-to-first-token. This post builds a tiered KV cache on Amazon SageM…
Hands-on, framework-free Colab notebooks for the AI Engineer / Forward Deployed Engineer (FDE) skill set — model APIs, structured output, tool calling, RAG, eva…
LLM-agent services repeatedly execute small deterministic transitions between model and tool calls: route an outcome, update state, and emit the next effect. We ask when this control path exposes enou…
The attention mechanism forms the foundation of many modern AI models such as the Transformer. In one subclass of problems where attention is used, inputs and outputs are bound to the probability simp…
We propose that a model's uncertainty about a token is reflected not only in the breadth of its output distribution but also in whether a confident prediction is \emph{fragile} under perturbation of i…
算力竞赛的天平,正在向电力倾斜
谷歌算力分配内耗严重

Anthropic will embed invisible watermarks in all Claude-generated text and sign files using the C2PA standard. New models shipping from August 2026 onward will have labeling built…
真就把GPU当理财产品了~

IT之家 8 月 11 日消息,据彭博社报道,知情人士透露,人工智能公司 Anthropic 已与比特币挖矿企业 Riot Platforms 达成一项价值 91 亿美元 (IT之家注:现汇率约合 614.74 亿元人民币) 的协议。Riot Platforms 近期开始将数据中心算力出售给人工智能企业。 这一交易凸显出 Anthropic 正在加紧争夺计算…
AI 点评 · 跨界算力交易折射AI军备竞赛白热化,矿业转型云服务成新趋势。
Meta's Muse Glimmer is a 30B open-weights agentic model under Apache 2.0. It fits 24 GB VRAM and decodes 3.1x faster with DFlash speculation. The post Meta AI Releases Muse Glimmer…
TokenFlow is an enterprise-grade, token-aware load balancer and AI gateway for vLLM. It prevents GPU OOM crashes using Redis-backed admission control, semantic…
TokenFlow is an enterprise-grade, token-aware load balancer and AI gateway for vLLM. It prevents GPU OOM crashes using Redis-backed admission control, semantic…
从云计算到云原生,产业真正的跃迁从来不是简单堆算力,而是找到更适合新世界的架构。
央视曝光酒店评级标识消费陷阱 《时代》网站展示仅 AI 可见广告 Edge 浏览器将禁用 Manifest V2 扩展 油价上涨致美国混动车需求激增 廉价手机壳存在毒性隐患 美国调查中国 AI 企业使用海外英伟达芯片 看看就行的小道消息 少数派的近期动态 你可能错过的好文章 查看全文
The AI-focused hedge fund is still making some big bets.
AI 点评 · 对冲基金逆势重注芯片初创,显示AI硬件投资热度不减。
Benchmarks for systems that are optimized against the evaluation signal measure something different from what they claim. We document this concretely in two GPU-kernel-optimization suites with held-ou…

IT之家 8 月 8 日消息,一款型号为 V2603A 的 vivo 新机近日现身 Geekbench 跑分数据库,该机已于近期通过 3C 认证。 Geekbench 数据显示,该机搭载联发科 MT6881 处理器,拥有 4 核 2.0GHz+4 核 2.6GHz CPU,GPU 为 Mali-G625 MC2, 对应天玑 7500 。 IT之家注意到,该机…
AI 点评 · 天玑7500中端定位明确,跑分曝光预示千元机市场竞争再添变数。

IT之家 8 月 8 日消息,据界面新闻今天报道,今年 5 月以来,腾讯 WorkBuddy 的广告从北京知春路铺到深圳高新园。打开手机,头部应用的信息流里也几乎都是 WorkBuddy 的身影。 据报道,这是腾讯近年来少见的饱和式产品投放。这种密集的攻势背后,是腾讯在渠道、算力、组织和生态资源上的集中倾斜。 据腾讯 CSIG 人士透露, WorkBuddy…
AI 点评 · 腾讯罕见倾全集团资源力推AI产品,战略信号强烈,或成其AI商业化关键落子。

IT之家 8 月 8 日消息,远景科技集团 8 月 6 日宣布在内蒙古乌兰察布建成 AI 算力超级单体,标志着“远景乌兰察布星河基地”正式投产。 该基地是远景“戈壁使命(Mission Gobi)”计划落地的首个旗舰项目,园区总规划容量达 2GW。 该超级单体采用超高比例绿电直连,建筑体量达 12 万平方米(约 20 个标准足球场)。基地具备百万卡并行能力与…
AI 点评 · 绿电直连百万卡集群,刷新算力基建天花板,能源与AI融合的新标杆。

AMD is buying Canadian startup Taalas, which hard-codes model weights directly into inference chips. That makes them extremely fast but locks each chip to a single model. A demo ch…
AI 点评 · 芯片直写模型权重,速度极致但灵活性受限,AMD此举或为定制化AI推理开辟新路。
While post-training improves the capabilities of large language models (LLMs), it generally lowers their output diversity and creativity, negatively impacting tasks that explicitly require creativity…
Under the standard split, Muon gets hidden matrices and AdamW embeddings/output head. Muon groks modular addition faster, but its solutions do not hold. All nine configurations on $(a+b) \bmod 113$ gr…

IT之家 8 月 7 日消息,据 Wccftech 今日报道,Anthropic 目前正在扩大内部芯片设计团队,并计划开发自研 ASIC 芯片。 与此同时,公司不同芯片相关岗位之间出现了明显的薪资差异:负责训练 AI 模型进行芯片设计的研究工程师岗位,年薪最高可达 85 万美元 (IT之家注:现汇率约合 574.9 万元人民币) ,而实际负责设计 Anthr…
AI 点评 · 高薪挖角透露AI巨头自研芯片战略,模型训练岗薪酬倒挂凸显行业人才争夺战。

IT之家 8 月 7 日消息,据《The Information》今日报道,亚马逊 AWS 最近对工程师提出一项颇为“朴素”的要求,让他们关闭不再使用的云计算机。 据报道,亚马逊此番要求主要针对闲置的 EC2 实例。随着 AI 正以越来越快的速度消耗计算资源,全球最大的云服务供应商也开始感受到算力紧缺的压力。 数据显示,AWS 的资源利用低下问题非常严重,大…
AI 点评 · 算力紧缺倒逼巨头精细化运营,AI扩张的隐忧终于浮出水面。

IT之家 8 月 7 日消息,海信官方今天在小红书平台公布 A10 墨水屏手机的更新参数, 新机将于 8 月 28 日正式发布 。 据介绍,这款手机采用高通 QCM6690 企业级平台 + 元太 T2000 硬件 TCON 芯片方案。其中 QCM6690 采用 4nm 八核心架构,并具备 6 TOPS 的端侧 AI 算力。 而元太 T2000 则是一款 TC…
算力、网络、内存与存储开始围绕每个Token协同
此后模型、算力、数据中心都在合作范围内
AI 点评 · 杭州双子星联手,AI与机器人产业协同将重塑生态,战略意义远超投资本身。

IT之家 8 月 7 日消息,AMD 今天(8 月 7 日)发布公告, 宣布计划收购 Taalas 芯片初创公司。 Taalas 面向 AI 推理设计加速器,针对单一 AI 模型定制芯片, 主要牺牲通用性,换取更低成本和更高输出速度。 在技术方面,Taalas 公司目前展示的 AI 推理芯片基于 Meta 公司的 Llama 3.1 模型,采用台积电成熟(较…
AI 点评 · 收购直击AI推理短板,定制化芯片策略或成AMD对抗英伟达的关键筹码。
Query-based mask transformers assemble segmentation outputs through pixel-wise competition among query predictions of the final layer, yet this inference process is not explicitly optimized during tra…

Ex-Google Deepmind CEO Demis Hassabis has reportedly stepped back from day-to-day operations for about a year, as he sees himself more as a scientist than a manager. Researchers ar…
AI 点评 · 人才流失折射AI竞争白热化,揭示大厂创新体制与芯片资源的核心博弈。

IT之家 8 月 6 日消息,据《DigiTimes》援引业内消息人士报道,2026 年全球折叠屏智能手机出货量预计将同比增长 20%,苹果首次进军折叠屏手机市场被认为是推动增长的重要因素之一。 报道称,整体智能手机市场目前正面临上游逻辑芯片供应趋紧以及存储芯片价格上涨等挑战,但折叠屏手机市场预计仍将保持增长势头。三星有望继续保持该领域的领先地位,而苹果、华…
AI 点评 · 苹果入场搅动折叠屏格局,2026年市场增长看点十足。

In this tutorial, we build a complete Bayesian marketing mix modeling workflow using Google Meridian. We begin by installing the required libraries, verifying GPU availability, and…
Near-term quantum hardware limits circuit depth and often imposes geometrically local connectivity for quantum generative models, restricting the output distributions accessible to shallow unitary Bor…

Mistral's new 3B Shieldstral model checks AI inputs and outputs for safety violations using natural language yes-or-no questions instead of fixed categories. It matches models seve…

Jeff Dean and other high-profile Google executives have founded Discovery Loop, a startup that will seek AI-powered breakthroughs in everything from drug discovery to chip design.

IT之家 8 月 5 日消息,阿里官方今日宣布,2026 云栖大会将于 9 月 22 日至 24 日在杭州举行。 本届大会主题为“智以致用”(Intelligence Goes Beyond),将以 Agentic AI 为核心,串联芯片、云基础设施、模型能力与模型服务、Agentic 应用的完整技术链路。 此次大会将设置三大主论坛。其中,“云栖主论坛”将聚…
Anthropic is building a team for designing its own custom AI chips. The Claude maker said it would co-design hardware and models to help its technology run faster and more efficien…
二季度AI收入暴涨247%
知情人士称,Anthropic PBC与成立仅数月的基础设施初创公司Volta Infra Holdings Ltd.达成一项价值100亿美元的算力协议,此举是这家Claude开发商为满足产品需求而采取的最新举措。知情人士表示,Anthropic已签署协议,将使用由英伟达支持的云基础设施初创公司Volta管理的数据中心。Volta周二早些时候宣布,已与一家未…
AI 点评 · 百亿算力大单折射AI军备赛白热化,新秀Volta获巨头押注,产业链格局生变。

IT之家 8 月 5 日消息,埃隆 · 马斯克(Elon Musk)旗下 SpaceX 公司今天(8 月 5 日)宣布携手英伟达公司,探索在地球之外部署 AI 算力,共同开发名为 Starmind AI1 的算力卫星。 根据 SpaceX 公布的宏伟蓝图,该公司设想构建一个轨道网络, 最终规模可达 100 万颗卫星 ,从而在太空打造一个分布式人工智能超级计算…
AI 点评 · 太空AI算力竞赛开启,星链与英伟达联手,万亿级新赛道浮现。

Google is working with Broadcom, Apollo, Blackstone, and Morgan Stanley on a multibillion-dollar financing structure that supplies Anthropic with AI chips and data centers while ke…
AI 点评 · 科技巨头联手资本运作,风险转移模式或成AI基建新范式。

IT之家 8 月 4 日消息,海外博主 ETA Prime 上周(7 月 31 日)在 YouTube 平台发布视频,实测 WoA(Windows on Arm)笔记本外接显卡能否玩游戏。 据悉,博主使用华硕 Zenbook A16 笔记本(IT之家注:骁龙 X2 Elite Extreme 芯片)和 AG03 显卡坞完成测试。英伟达已经在近日发布 RTX…
出海AI,正被“三重算力锁链"捆死
今日热点导览 OpenAI下一代模型Astra曝光,以2000美元算力成本破解十项数学难题 江小白起诉东方甄选结果公布:东方甄选构成商业诋毁判赔30万 iPhone被曝最高或涨价超千元,折叠屏iPhone至少2000美元起售 多平台牛蛙抽检抗生素超标,有样本检测出禁用兽药 王思聪卸任万达产业投资公司董事 TOP3大新闻 宗馥莉同父异母兄弟出任新公司董事长,实…
Don't be a meat proxy Niklas Gruhn coins an excellent new term - meat proxy - for people who blindly copy and paste the output of AI systems to their peers. By all means, prompt AI…
从全球经验来看,很多科技领军企业的崛起,都少不了创业投资的支持。今年上半年,高技术产业投资同比增长4.6%,这一数据背后,是创业投资的活跃度在持续回升。业内人士普遍认为,AI技术的发展,是这一轮创投市场回暖的核心驱动力,资金正加快涌向AI、量子科技、算力等领域。数据显示,2026年上半年中国AI创业公司融资总额突破3000亿元,已超过2025年全年。而资金的…
AI 点评 · 创投风向标转向硬科技,3000亿真金白银验证AI赛道热度。

IT之家 8 月 4 日消息,科技媒体 9to5Mac 昨日(8 月 3 日)发布博文,报道称基于苹果首席执行官蒂姆 · 库克(Tim Cook)在财报电话会议上释放的信号,苹果仍在评估 iOS 27 中 Siri AI(Siri 人工智能)的算力成本与收费方式。 在运行方式上,根据苹果此前公示的细节,通过拆分 AI 任务, 尽可能在设备端本地处理。这种方式…
AI 点评 · 苹果AI商业化探路,Siri收费模式成行业风向标。
Fairness evaluation concerns not only what a model produces, but also what its outputs ought to be compared against. When a model generates "a CEO in the United States," the prompt leaves demographic…

IT之家 8 月 3 日消息,据彭博社记者马克 · 古尔曼(Mark Gurman)报道,全球内存芯片短缺正在影响苹果 Mac 产品供应,其中受消费者欢迎的 MacBook Air 正面临库存压力。 古尔曼表示,此次内存短缺主要受到 AI 产业快速扩张影响。随着人工智能企业大量采购高性能存储芯片,挤压了消费电子市场供应,苹果旗下部分产品已经受到影响。 此前,…
AI 点评 · AI算力挤占内存产能,消费电子供应链联动效应显现,涨价与缺货风险值得警惕。
Optimization-based latent reasoning improves large language model outputs by optimizing instance-specific continuous states at test time while keeping model parameters frozen. Existing methods, howeve…
Weight-only quantization substantially reduces the storage of large language model (LLM) transformer blocks, but practical backends often retain the final language-modeling head (LM-head) in BF16 or F…
Scientific poster construction compresses a long multimodal paper into a readable, editable canvas. Existing systems hide request-level failures by scoring only completed outputs; direct image generat…

A Munich court ruled that AI music generator Suno violated copyrights through both training and output. The court found six songs reproducibly stored in Suno's models and rejected…
AI 点评 · 欧洲首例AI音乐版权判例,明确训练与输出双重侵权,或重塑全球AI合规标准。
A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.
Enterprise workflows increasingly rely on agents for schema-guided extraction: given a document and a user-defined schema, the agent faithfully follows the schema to produce the correct output with so…
Diff your AI agent's behavior between two runs. See exactly which tool calls, args, costs and outputs changed when you swap models or edit prompts.