JavaScript · ★ 44,478 · 🍴 2,766 · 📈 2,271 stars today
A coding-agent skill for multi-phase security audits with independently verified, machine-readable findings
中文介绍 云flare的编码代理技能，用于多阶段安全审计，提供独立验证、机器可读的发现结果。适用于需要自动化安全审计的工具和团队。
Rust · ★ 19,461 · 🍴 1,930 · 📈 893 stars today
OpenShell is the safe, private runtime for autonomous AI agents.
中文介绍 NVIDIA开发的OpenShell，为自主AI代理提供安全、私有的运行时环境。适用于开发和使用自主AI代理的开发者。
Python · ★ 3,119 · 🍴 397 · 📈 11 stars today
Official, AWS-supported MCP servers, skills, and plugins to help AI agents build on AWS
中文介绍 AWS官方支持的MCP服务器、技能和插件，帮助AI代理在AWS上构建。适用于在AWS上部署AI代理的开发者。
CSS · ★ 32,107 · 🍴 1,647 · 📈 570 stars today
Anti-AI-slop design skill for Claude Code, Cursor, and Codex.
中文介绍 Nutlope开发的Anti-AI-slop设计技能，适用于Claude Code、Cursor和Codex。适用于需要AI辅助设计的开发者。
Python · ★ 44,530 · 🍴 5,489 · 📈 442 stars today
A visual, example-driven guide to Claude Code — from basic concepts to advanced agents, with copy-paste templates that bring immediate value.
中文介绍 Claude Code的视觉、示例驱动指南，从基础概念到高级代理，提供可粘贴模板。适用于学习Claude Code的用户。
Rust · ★ 86,372 · 🍴 5,152 · 📈 733 stars today
The design language that makes your AI harness better at design.
中文介绍 pbakaus开发的AI设计语言，使AI在设计中表现更佳。适用于AI设计应用的开发者。
Rust · ★ 10,967 · 🍴 1,405 · 📈 381 stars today
A Tiny biped duck robot 🦆
中文介绍 pollen-robotics开发的微型鸭形双足机器人。适用于机器人爱好者或研究者。
TypeScript · ★ 108,027 · 🍴 14,807 · 📈 38 stars today
An open-source AI agent that brings the power of Gemini directly into your terminal.
中文介绍 Google Gemini CLI，将Gemini AI代理的力量直接带入终端。适用于需要直接在终端中使用AI代理的开发者。
Python · ★ 22,580 · 🍴 3,189 · 📈 220 stars today
欢迎来到电子书下载宝库，一个汇聚了各类电子书下载链接的地方。无论你是喜欢阅读经典文学、经管励志、终身学习、职场创业、技术手册还是其他类型的书籍，这里都能满足你的需求。 该库涵盖了帆书app(原樊登读书)、微信读书、京东读书、喜马拉雅等读书app的大部分电子书。
中文介绍 电子书下载宝库，汇集了各类电子书下载链接，包括帆书app、微信读书等平台。适用于电子书爱好者。
TypeScript · ★ 4,074 · 🍴 584 · 📈 471 stars today
Consider it done. The open-source AI agent that works out of the box · 想到，就能做到。开源、开箱即用的 AI Agent。
中文介绍 makecindy开发的Cindy AI代理，开源、开箱即用。适用于需要快速部署AI代理的用户。
👍 7
Reinforcement learning (RL) is credited with teaching reasoning models to write {wait} before reconsidering a step, yet large RL accuracy gains do not require the word. This view can confuse what RL discovers with what a model inherits from its training data. We trace the word through open models' t
👍 15
Diffusion Language Models (DLMs) introduce unique quantization challenges due to their iterative denoising process. In this work, we reveal that quantization sensitivity is both taskand time-dependent, with different downstream tasks exhibiting distinct precision requirements across denoising steps.
👍 7
Frontier coding agents are becoming increasingly adept at 3D modeling. Yet, they still face challenges when building objects meant to be assembled in the physical world, often creating invalid assemblies with floating, colliding, or disconnected parts. We argue that providing agents with tools for a
👍 26
Complex camera motion is central to visual storytelling, yet translating cinematographic intent into controllable, executable camera trajectories remains challenging. We introduce FilmFlow, an agentic framework that converts natural-language instructions or reference videos into explicit camera traj
👍 8
Autonomous research agents are increasingly used to automate large portions of scientific research, yet they are prone to research integrity violations (RIVs): behaviors that compromise the validity, reproducibility, or faithfulness of the research process. We present RIDAR, a framework for automate
👍 86
We present IQuest-Q1, a sparse Mixture-of-Experts (MoE) agentic foundation model with 320B total parameters (15B activated). IQuest-Q1 is developed through a three-stage process: (1) stable and efficient pre-training, combining Stable SplitNorm, Muon, and muP with a two-stage curriculum whose second
👍 5
Masked diffusion language models generate tokens in parallel, but their efficiency depends on which predictions are committed. Learned unmasking policies improve the speed--accuracy trade-off, yet differ in supervision, inputs, and capacity, obscuring which choices matter. We view supervised policie
👍 5
This paper studies category-level RGB-D object pose estimation, which recovers an object's 3D rotation and translation without instance-specific CAD models or reference views. This task is highly ambiguous due to intra-class variations, symmetries, and partial observations. Current solutions struggl
👍 11
AI coding agents can now build small playable games in minutes, but the results usually look like programmer art: gray blockouts, colored geometric primitives, and flat semantic blobs. We present AnyGameRenderer, a video-to-video generative renderer that turns such footage into high-quality visuals
👍 15
Hierarchical language generation separates continuous semantic modeling from discrete text realization, but its effectiveness depends on what the latent space compresses and how the discrete generator uses it. We propose Cola2, a framework that jointly studies continuous semantic compression, the ba
👍 11
Vision-language models (VLMs) can perform a broad range of tasks across diverse settings, yet their performance in agentic contexts remains poorly understood. Although online evaluations in simulators and real-world environments are essential for measuring interactive agent performance, they are cos
👍 15
LLM memory is widely used in long-term conversation understanding. Agentic memory systems often use mutation, structural organization, learned retrieval strategies, or other complex operations. Is this complexity always necessary? We study Raw-RAG, a naive RAG baseline using dense retrieval and rera
👍 2
Robot world models simulate how environments evolve in response to robot actions. Latent action models enable learning dynamics from videos by encoding visual transitions as proxy actions. However, reconstruction-driven latent actions mix robot motion with object responses and other scene changes, c
👍 17
Deep research agents synthesize open-web evidence into long reports, but current benchmarks do not evaluate research centered on a source video. We introduce Video-Centered Deep Research Report Generation. Given a video and user request, an agent identifies research questions, retrieves external mul
👍 7
Policy enforcement for LLM agents has shifted from single-turn content moderation to compound, application-specific policies evaluated over multi-turn transcripts and tool-call trajectories. Fine-tuned guardrail models are efficient but degrade on compound policies, and while LLM-as-a-judge approach
👍 2
Robots must act under physical and social constraints while adapting to changes in their environment and human needs. Existing embodied-agent systems often focus on specified goals or abstract execution into symbolic actions, limiting evaluation in an evolving world. We introduce TARSphere, a robot-
👍 16
Modern LLM agents can retrieve information from accessible documents, and then disclose a sensitive conclusion assembled from them. This mosaic effect creates a critical risk for agents operating over confidential data. We quantify its realization with 30 synthetic cybersecurity, medical, and milita
👍 50
Recursive self-improvement (RSI) describes a system that improves not only its behavior but also the process by which it improves. Recent digital agents test this idea by rewriting their own code and checking each version with executable tests. Bringing these mechanisms into the physical world requi
👍 16
Feed-forward 3D Gaussian Splatting (FF3DGS) reconstructs 3D scenes from images in a single forward pass, making it well-suited to applications that must operate within limited resource budgets. To further improve its practicality, recent budget-controllable FF3DGS methods allow the number of Gaussia
👍 3
Diffusion language models can write several tokens in one forward pass, and they decide during generation how many passes to spend. Two outputs of the same token length can therefore vary a lot in cost. To understand the underlying factors of this cost variation, we analyze two diffusion models acro
👍 5
As language-model agents grow more capable, they can carry more of a task without asking anyone for help. Falling reliance on a human partner is easy to read as a failure of collaboration when it may instead be a correct division of labor, and the rate of human involvement alone cannot tell the two
👍 18
LLM routing dynamically selects and coordinates heterogeneous language models across diverse tasks. However, existing approaches are largely built around a predefined execution paradigm, such as single-round, multi-round, or agentic routing, which fixes how model calls are organized and limits adapt
👍 18
Tool-using voice agents often leave users waiting after speech ends, even when actionable information arrives earlier in the request. Speculative approaches begin work during speech, but chunk-bounded or interrupt-on-update generation can truncate a tool call before its arguments are complete. This
👍 154
Despite substantial progress in Deep Search agents, their ability to conduct end-to-end research in real-world open Web environments remains insufficiently evaluated. Unlike static retrieval settings, information on the real-world Web is distributed across diverse sources, modalities, and heterogene
👍 41
Large language models need effective ways to acquire and use external knowledge beyond what they retain from pretraining. Continued pretraining risks catastrophic forgetting, while retrieval at inference time can add substantial overhead. External parametric memory avoids online retrieval, but const
👍 1
Effective test-time personalization requires LLM agents to adapt to query-dependent user preferences from lightweight interaction. We study persona selection with binary comparisons between candidate responses as the online signal. We present ID-TAP, a scalable, IDS-inspired neural dueling-bandit fr
👍 2
Improving every agent in an LLM-based multi-agent system can be prohibitively costly, making it important to identify which agent should be prioritized to achieve the greatest system-level gain with limited resources. Failure attribution can inform this decision, but an agent's responsibility for fa
👍 2
Recursive self-improvement (RSI) relies on evaluation feedback to assess progress and guide further research, yet repeatedly running complex benchmarks is costly and slows iteration. Human experts reduce this cost by selecting benchmark subsets or designing compact suites. We ask whether AI agents c
👍 4
Split language models keep input text on a client but send intermediate representations to a server. Attacks can reconstruct text from these messages. Comparing privacy evaluations is difficult when attackers see different messages, know different model weights or use different datasets and recovery
👍 9
Chemical chain-of-thought can contain hallucinated structural and property claims. We thereby introduce MolLean, a process-verifiable language that grounds chemical reasoning in an explicit molecular state. A layered Lean-based kernel checks state transitions and assertions over molecular readouts,
@vinayp10 · 17.3K 粉丝 · 175.8K 阅 · 506 赞 · 15 转
Claude, you're fired! I just moved my fiancée off Claude Code (@claudeai) and onto @bot. We canceled her Claude Code subscription. That's a first for me. Since February, whenever someone asked which
中文介绍 博主从 Claude Code 转向 @bot，取消订阅，分享使用体验。
@ryangrowthh · 5.6K 粉丝 · 74.8K 阅 · 519 赞 · 27 转
My team of 11 Grok Bots helped me generate ~2M impressions in my first two weeks of posting on 𝕏. Today, I'm revealing the entire system. Firstly, the elephant in the room - proof of the results:
中文介绍 博主分享如何使用 Grok Bot 生成 200 万次曝光，详细系统介绍。
@tibo_maker · 215.7K 粉丝 · 56.5K 阅 · 510 赞 · 30 转
in this article I'll walk you through all 9 steps of the setup, with the 12 prompts I use, ready to copy, in the order you'll need them. you've probably scrolled past a few of them this week. a girl
中文介绍 教程：构建和运行 AI UGC 账户团队，提供 12 个可复制的提示步骤。
@BrianDEvans · 1.3M 粉丝 · 50.8K 阅 · 826 赞 · 34 转
AI is going to create some ridiculously cracked people. Imagine someone who built a hundred-billion-dollar company around their own invention, won Olympic gold, and directed an Oscar-winning film.
中文介绍 探讨 AI 如何创造极端成就者，如企业巨头、奥运金牌得主和奥斯卡获奖导演。
@davidim · 7.6K 粉丝 · 47.8K 阅 · 532 赞 · 31 转
Today we launched Sume API with a 0% markup, which means you get the same models at exactly the price the original model provider charges. A lot of people have asked me why we'd do something that
中文介绍 介绍 Sume API 0% 加成策略，解释为何推出此 API。
@GoogleCloudTech · 1.3M 粉丝 · 38.7K 阅 · 502 赞 · 68 转
There is a running joke in modern software engineering that the hottest programming language isn't Rust, Zig, or Python: it’s Markdown. As coding agents become entrenched in developer workflows, many
中文介绍 Markdown 架构模式提升 AI 生产力，分享 3 种模式。
@puneetmehtanyc · 4.0K 粉丝 · 37.2K 阅 · 577 赞 · 57 转
Major U.S. banks are detecting third-party AI agents operating inside customer accounts without the banks’ approval, Axios reported this morning. Some of those agents appear to be acting beyond their
中文介绍 报道银行检测第三方 AI 代理在客户账户中的操作，探讨 AI 代理的挑战。
We have Fable at home... or so we are being told by paid shills
中文介绍 Kimi 3.1 Pro 104BA-2.8T开放权重缩小了与Closed-AI的差距，并推出2-3倍速度的Pro Highspeed API。
Learn how OpenAI scaled Habitat with reliable multi-tenancy, global caching, and Azure Cosmos DB to deliver 12× faster reads for ChatGPT.
中文介绍 OpenAI通过可靠的多人租户、全球缓存和Azure Cosmos DB将Habitat扩展至支持超过10亿ChatGPT用户，读取速度提升12倍。
中文介绍 Hugging Face Blog介绍了一种通过动态文献环境中的上下文来发展研究想法的方法。
AI models are solving more mathematical problems. A pattern is emerging in their successes that could help us predict which problems they’ll solve next.
中文介绍 AI模型在解决数学问题方面表现出色，并出现了一种成功模式，有助于预测它们将解决哪些问题。
the Cyber onslaught continues.
中文介绍 网络安全攻击持续进行中。
OpenAI explains how we strengthened containment for frontier model training and evaluation, including offline environments and high-compute red-teaming.
中文介绍 OpenAI解释了如何加强前沿模型训练和评估的隔离措施，包括离线和高性能的对抗性测试环境。
OpenAI and Poland’s NASK are partnering on AI safety, security, and technical research to strengthen safeguards and support a more resilient digital ecosystem.
中文介绍 OpenAI与波兰的NASK合作，推进人工智能安全和安全技术研究，以加强保障并支持更弹性的数字生态系统。
中文介绍 Hugging Face Blog介绍如何在AWS Trainium上运行模型。
中文介绍 OpenAI估值接近1250亿美元，推出1Password for Agents，Gemini 3.5 Pro回顾。
中文介绍 Hugging Face Blog介绍如何使用ThinkingBox将数据库评分转换为强化学习奖励，以训练后代理。
The next competitive advantage in enterprise AI has less to do with which model you choose than with what you build around it. That’s an important shift. For the past several years, much of the AI conversation has centered on model capability: bigger models, better reasoning, larger context windows,
中文介绍 MIT Tech Review AI探讨企业AI的下一个竞争优势，指出与模型选择相比，围绕模型构建的内容更为重要。
Fast, focused, low-cost decisions for any application.
中文介绍 OpenAI推出Decisions API，为任何应用程序提供快速、专注、低成本的决定。
A scalable AI system requires a strong foundation. Data, engineering, platforms, operating models, and governance must converge to work as one. This session from MIT Technology Review’s EmTech Future conference explores what it takes to build that foundation in a highly regulated environment and how
中文介绍 MIT Tech Review AI探讨可扩展人工智能的基础，强调数据、工程、平台、运营模式和治理必须协同工作。
Elastic GPUs, sandbox growth, and the economics of the agent cloud
中文介绍 Modal CEO Erik Bernhardsson讨论了弹性GPU、沙盒增长和代理云的经济效益。
A global map categorizing every 10-meter patch of land on Earth into distinct forest types.
中文介绍 DeepMind Blog发布了一幅全球地图，将地球每10米土地划分为不同的森林类型。
Offenders still at large after shots fired from vehicle outside licensed premises in Loganholme Get our breaking news email, free app or daily news podcast Four people have been injured after a drive-by shooting outside a venue hosting a boxing event. Police said they responded to reports of shots b
中文摘要 布里斯班南部一场拳击比赛外发生枪击，四人受伤，凶手仍在逃。
Donald Trump praised First Lady as 'better than professionals' after a localised Russia-Ukraine ceasefire in Oleshky.
中文摘要 特朗普称赞梅拉尼娅在俄乌局部停火协议中的表现。
Prime minister is considering different options for new relationship with Brussels ahead of summit in November Andy Burnham says he wants an “honest discussion” with the British public about the possibility of a return to free movement between the UK and the EU as part of his plans for a new relatio
中文摘要 英国首相表示希望就英欧自由流动问题进行坦诚讨论。
The crackdown follows the killing of two Israeli soldiers by two Palestinian gunmen on Friday - and comes just days ahead of Israel's election.
中文摘要 以色列军队摧毁希伯伦家园，巴勒斯坦家庭被迫流离失所。
The comments were his first since the disruption in the country’s satellite service. Earlier, Ukraine said drones using Starshield, the Pentagon-funded service, were back to normal speeds.
中文摘要 马斯克称乌克兰违反了Starlink的规定。
Nigerian police say a large group carried out the attack on Binnari village in Taraba State.
中文摘要 尼日利亚东部发生枪击事件，至少10人死亡。
El Salvador had sent hundreds of troops to find the men, who were ambushed six days ago while working as part of the Gang Suppression Force.
中文摘要 在haiti发现四名失踪萨尔瓦多士兵的遗体。
From trans rights to Israel's war on Gaza, the party's candidates have responded to voter shifts on a slew of issues.
中文摘要 自上次美国大选以来，民主党政策发生了哪些变化？
Voters in Sunday's election are choosing between two very different candidates, as Americas correspondent Ione Wells reports from Rio de Janeiro.
中文摘要 巴西总统选举候选人是谁？
Thousands have marched in student demonstrations over education, the cost of living, and a project linked to Jared Kushner, President Trump’s son-in-law.
中文摘要 阿尔巴尼亚因教育、生活成本和与特朗普女婿相关的项目再次爆发暴力抗议。
The pivotal run-off will determine whether Brazilians stick with their left-wing leader or shift towards the right.
中文摘要 巴西总统候选人卢拉和博尔索纳罗在选举前夕集会上吸引了数千人。
Brazil’s candidates hold final rallies before knife-edge run off
中文摘要 巴西候选人举行最后集会，为关键的决选做准备。
Tehran warns that even limited strikes would trigger a response 'several times greater'.
中文摘要 伊朗指责美国导致协议失败，呼吁在担忧新攻击的情况下重新尝试。
But Elon Musk, whose company SpaceX owns Starlink, accuses Ukrainian military of violating the company’s terms Ukraine’s military has a “stable connection to Starlink”, the Ukrainian government said, 10 days after its ability to use the satellite communications system was suddenly restricted. But h
中文摘要 乌克兰军方表示，在限制使用后，Starlink卫星连接稳定。
After largely peaceful marches on Saturday, clashes broke out in the evening. But a large-scale police presence, 1,000 arrests and the rerouting of protest routes kept a lid on wholesale vandalism.
中文摘要 法国再次爆发动荡，但这次财产损失较少。
Scotland editor James Cook finds out what life is like on board a ship in an untapped oil field awaiting its fate.
中文摘要 苏格兰编辑詹姆斯·库克报道了在未开发油田上的一艘船上的生活情况。
Saudi Arabia’s main airport was gearing up to reopen as early as Sunday after operations had been disrupted by a spate of attacks by Houthi rebels.
中文摘要 沙特阿拉伯表示，里雅德机场可能于周日重新开放。
Barry speaks with Robin Wigglesworth, editor of Financial Times' Alphaville blog and author of the new book 'A Fabulous Debt: The Epic Story of How Bonds Built the Modern World'. They discuss the bond market as the engine of the global economy, why the 30-year mortgage is "a miracle of the modern wo
中文摘要 巴里与《金融时报》Alphaville博客编辑、新书《债务的传奇：债券如何构建现代世界》作者罗宾·威格尔斯沃斯讨论了债券市场作为全球经济引擎的作用。
Student protesters in Paris, France were dispersed with gas as conflict escalates. (Source: Bloomberg)
中文摘要 法国学生抗议者因冲突升级而被催泪瓦斯驱散。
The news doesn’t stop when markets close. Hosts David Gura, Christina Ruffini and Lisa Mateo bring clarity, context and a bit of humor to the weekend’s biggest headlines, LIVE from New York. Joined by The New Yorker Staff Writer Gideon Lewis-Kraus, Linx Dating Founder & CEO Amy Andersen, Oliver Wyma
中文摘要 周末新闻：中东新威胁，俄罗斯柴油将进入美国市场。
It was a panicked message from a colleague that woke fund manager Ivelina Green at 5 a.m. on Thursday morning in New York: “Total meltdown.”
中文摘要 纽约基金经理伊夫琳娜·格林在周四早晨5点被同事的紧急消息惊醒：“全面崩溃。”
Metals traders are earning billions of dollars in a second consecutive blowout year for the sector, with several top trading houses notching up record profits.
中文摘要 金属交易商在连续第二年实现创纪录利润，赚取数十亿美元。
Medical historian Lindsey Fitzharris tells Bloomberg This Weekend how Scottish surgeon Joseph Bell’s remarkable powers of observation inspired his student Arthur Conan Doyle to create Sherlock Holmes. Speaking with hosts Christina Ruffini and Lisa Mateo, Fitzharris discusses her new book, Sleuth-Hou
中文摘要 医学历史学家林赛·菲茨哈里斯讲述了苏格兰外科医生约瑟夫·贝尔的非凡观察力如何启发了他的学生亚瑟·柯南·道尔创作夏洛克·福尔摩斯。
A Chinese engineering company is on the cusp of partially restarting a massive Venezuelan power plant that’s been largely idle for years, a crucial step toward fixing the nation’s dilapidated grid.
中文摘要 委内瑞拉将在中国的帮助下重启一座关键电站。
Dalmore Capital is among parties that have expressed interest in acquiring UK commercial equipment supplier and services provider JLA Group from buyout firm Cinven, according to people with knowledge of the matter.
中文摘要 据知情人士透露，Dalmore Capital等方对收购Cinven拥有的英国商业设备供应商和服务提供商JLA Group表示了兴趣。
AI is driving a wave of new data centers, but communities are increasingly asking what they get in return. Virginia’s Loudoun County turned the industry into a major tax base that helped fund schools, parks and infrastructure, yet even there, officials are now putting limits on further growth. The b
中文摘要 弗吉尼亚州劳登县如何让数据中心热潮带来回报。
Wall Street Journal foreign correspondent Evan Gershkovich tells Bloomberg This Weekend how his reporting experience and knowledge of Russia helped him understand that his 2023 arrest on false espionage charges was likely intended to make him part of a prisoner exchange. Speaking with hosts David Gu
中文摘要 华尔街日报驻俄罗斯记者伊万·格什科维奇讲述了他的报道经验和俄罗斯知识如何帮助他理解，他2023年因虚假间谍指控被捕很可能是为了成为囚犯交换的一部分。
16 回复 · 程序员 节点
6 回复 · 程序员 节点
10 回复 · 程序员 节点
18 回复 · 程序员 节点
9 回复 · 程序员 节点
8 回复 · Apple 节点
7 回复 · Apple 节点
32 回复 · Apple 节点
50 回复 · Apple 节点
19 回复 · Apple 节点
该源今日无内容。