微信内可能无法直接打开本站。请点右上角 ··· → 在浏览器打开,或复制链接后用系统浏览器访问。
XMT
信流 · 上滑连读 · 来源可核
拆解 Claude 5.1:38 小时不睡觉的背后,Anthropic 正在终结「模型论」
状态校验、权限路由、异步调度,撑起一套可持续执行的运行时。作者丨郑佳美 编辑丨岑 峰 刚刚,Anthropic 更新了 Claude 5.1,一口气推出 Fable 5.1 和 Mythos 5.…
- 这两个名字看起来像两档模型,实际上底层用的是同一套模型。
- 区别在于 Anthropic 把能力开放范围单独拆了出来:Fable 负责通用编码、知识工作和 Agent 任务,Mythos 则面向经过验…
- Fable 和 Mythos 的分化,标志着 Anthropic 正式将‘智力’与‘权限’物理剥离。
RSS 官方收录 · 可信分层展示
IFA 2026 倒计时 | 魔法原子以“本体+大脑+场景”全栈能力叩开欧洲工业大门
2026年9月4日至8日,全球领先的具身智能公司魔法原子(MagicLab)将亮相IFA2026,完整呈现其"本体硬件、物理AI大脑、场景交付"三层全链路闭环能力,同时面向欧洲企业级市场战略聚焦柔性搬运与物流分拣两大高价值工业场景,展示其作为“端到端工业级具身智能解决方案提供商”的全球化竞争力。…
- 在现场,专业观众与行业买家将有机会近距离观摩MagicBot D1与HyperBot的多机协同调度,并亲身体验Magic-VLA K02大模…
- 此外,现场还将设置商务洽谈专区,供全球合作伙伴深入了解魔法原子的渠道政策与本地化售后支持体系。
- 新品亮相——三款新品登陆欧洲,本体矩阵覆盖全场景作业魔法原子此次将带来三款2026年全新发布的机器人产品,构成覆盖“技术摸高—工业主力—轻量…
RSS 官方收录 · 可信分层展示
实测飞猪AI:真正的突围才刚刚开始?
2026 年上半年,AI 行业的风向又变了。在经历一轮疯狂堆算力、不计成本砸研发的扩张周期之后,全球科技巨头都不得不直面同一个现实问题:“AI砸进去的钱,究竟该如何收回来?”AI Coding 或许是一个“标准答案”,过去很长一段时间里行业目光都集中于此。…
- 2025 年 5 月,Anthropic 正式推出 Claude Code,产品上线仅 6 个月,年化收入便突破 10 亿美元,把编程 Ag…
- 据悉,到 2026 年 2 月,这一数字进一步飙至 25 亿美元。
- 不过 AI Coding 的所谓范式,很大程度得益于封闭纯粹的数字环境:代码环境是确定的,结果可被客观校验,更容易形成 Agent 迭代闭环。
RSS 官方收录 · 可信分层展示
视频生成算力新选择:SmarCo GC3如何绕过“存储墙”这道硬门槛
大模型生成视频,正在成为这个算力饥渴时代最典型的应用场景之一。一段由AI生成的1080P视频,背后往往是数十亿参数神经网络的反复迭代,是海量张量运算在时空维度上的密集交织,也是无数像素数据在芯片与存储之间的辗转奔袭。…
- 对于大多数创作者而言,按下"生成"按钮之后等待渲染完成的焦虑,以及不断攀升的GPU集群账单,几乎已经成为一种常态。
- 这场无声的算力战争背后,一个结构性的困境正在浮出水面——传统芯片架构下,算力的大头消耗在了"搬运"而非"计算"上。
- 存储墙:视频生成绕不开的硬约束视频生成场景的算力特征与静态图像处理截然不同。
RSS 官方收录 · 可信分层展示
Agriculture relies on fossil fuels. It’s costing us.
If you’ve had to fill up your vehicle’s gas tank or buy a plane ticket lately, you’ve probably felt the effects of rising fossil-fuel prices.…
- But farmers buying fertilizer for their crops are especially aware of …
- Fertilizer prices have been on a roller coaster this year, kicked off …
- Let’s take a closer look at why conventional fertilizer prices are so …
RSS 官方收录 · 可信分层展示
NucleicBERT interprets RNA sequence space through self-supervised language modelling
Nature Machine Intelligence, Published online: 03 September 2026; doi:10.…
- 1038/s42256-026-01295-9RNA structure and function are hard to infer be…
- Upadhyay et al.
- trained a self-supervised model on large-scale RNA data that derives b…
RSS 官方收录 · 可信分层展示
Quantum neural operators with implicit quadratic frame and expressivity advantages
Nature Machine Intelligence, Published online: 03 September 2026; doi:10.1038/s42256-026-01289-7Wang et al.…
- introduce a hardware-efficient quantum neural operator that overcomes …
- Using an implicit quadratic frame, it offers accelerated expressivity …
RSS 官方收录 · 可信分层展示
Scaling agentic AI pilots across the enterprise
As agentic AI moves from experimentation toward enterprise deployment, the challenge is figuring out how agents can work together, connect to the systems and data they need, and operate safely across the workflows that run a business.…
- Although agentic AI has been adopted by some 80% of Fortune 500 compan…
- WATCH THE WEBCAST For Arun Chandra, chief operating officer at NiCE, t…
- “Everybody’s trying to figure out what can we do with this technology?
RSS 官方收录 · 可信分层展示
全场景智享硬核SUV 传祺越7开启预售 权益价17.18万-21.78万
9月2日,“都市自在 山海从容”传祺越7预售发布会于广州白云国际会议中心举办。广州这座活力之城,既是大湾区产业创新高地,亦是中国汽车产业的重要策源地。广汽传祺是广州首个乘用车自主品牌,更是广汽自主事业的核心根基,发布会选址在素有“岭南会客厅”之称的广州白云国际会议中心,既体现了传祺越7是广汽技术积淀的集大成之作,更代表着广汽全球化的布局和决心。…
- (传祺越7预售发布会合影)在本次发布会上,传祺越7正式开启预售,预售权益价为17.
- 为回馈广大用户的支持与厚爱,现下订传祺越7可享受至高价值28000元的四重好礼,丰富权益诚意满满。
- 更加惊喜的是,越7还送出一年免费使用权的超级福利。
RSS 官方收录 · 可信分层展示
「小」鸭子火了之后,端侧 AI 还有一个更「大」的问题
最近,一只只有 25 厘米高的机器鸭,意外把端侧 AI 的想象力拉到了聚光灯下。Hugging Face 旗下 Pollen Robotics 发布的 Microduck,是一款售价 399 美元的开源双足机器人。…
- 它约 800 克重,开箱即可行走、坐下、踢球、抓取,摔倒后还能自己爬起来。
- 更重要的是,它并不只是一个会动的玩具,而是一个面向开发者的开源机器人平台:用户可以在仿真环境中训练动作策略,再部署到真实机器人上。
- 这只“小鸭子”之所以引发关注,不只是因为它足够可爱、足够便宜,也不只是因为机器人终于开始进入普通开发者的桌面。
RSS 官方收录 · 可信分层展示
Codebook Agent: Amortized Topology Design for LLM Multi-Agent Systems
arXiv:2609.…
- 02264v1 Announce Type: new Abstract: Adapting the communication topolo…
- We argue that this formulation is misaligned with the problem.
- Empirically, topologies that survive a reward filter collapse to about…
RSS 官方收录 · 可信分层展示
APEx: Distillation of Agent Procedural Experience for Adaptive Deep Research Question Answering
arXiv:2609.…
- 02253v1 Announce Type: new Abstract: Deep research agents augment larg…
- Learning from prior experience is crucial for continual improvement, y…
- We propose APEx, a hierarchical experience utilization framework that …
RSS 官方收录 · 可信分层展示
LLM-as-a-Judge Is Not an Oracle: Why Self-Improving Agents Need Deterministic Guardrails
arXiv:2609.02246v1 Announce Type: new Abstract: Self-improving agent pipelines have a problem at their center.…
- An optimizer rewrites prompts to score higher, and the score comes fro…
- That judge has the last word on whether the system is getting better, …
- The judge should be demoted from oracle to advisor: its verdict become…
RSS 官方收录 · 可信分层展示
Task-Level Natural Language Priors as Learning Signals for Low-Resource LLM Training
arXiv:2609.02244v1 Announce Type: new Abstract: Large language models (LLMs) often struggle when low-resource training data are ambiguous or incomplete.…
- Task-level natural-language priors can provide useful guidance in such…
- We propose Prior-Guided Tuning (PGT), a training perspective that inco…
- Under this perspective, we introduce Contrastive Prior Steering (CPS),…
RSS 官方收录 · 可信分层展示
Propose to Learn, Learn to Propose: Evaluability-Aware Assistance under Bounded Rationality
arXiv:2609.02242v1 Announce Type: new Abstract: AI assistants often collaborate by proposing candidate edits, plans, or designs that users evaluate before adoption.…
- Existing assistance methods focus on proposal quality or user-goal inf…
- We study evaluability-aware proposal planning, where proposals serve b…
- We formalise this setting as ProSE, a hidden-parameter sequential assi…
RSS 官方收录 · 可信分层展示
PGPO: Potential-Guided Policy Optimization for Multi-Turn Agentic Tasks
arXiv:2609.…
- 02236v1 Announce Type: new Abstract: Group-based reinforcement learnin…
- To obtain more fine-grained credit assignment, recent work such as GiG…
- However, these step-level signals still rely on the final outcome of e…
RSS 官方收录 · 可信分层展示
PhoenixNest-Video: Evidence-Grounded Multimodal Agent Framework for Automated Video Interview Assessment
arXiv:2609.…
- 02231v1 Announce Type: new Abstract: Interview assessment requires per…
- We introduce PhoenixNest-Video, an evidence-grounded multimodal agent …
- It builds a semantic video graph as structured working memory, perform…
RSS 官方收录 · 可信分层展示