Python · ★ 43,756 · 🍴 5,075 · 📈 3,274 stars today
VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.
中文介绍 VoiceStudio是一个开源的语音克隆、设计、配音、转录和有声书创作工具，支持646种语言，是ElevenLabs的替代品。
TypeScript · ★ 92,622 · 🍴 15,871 · 📈 3,185 stars today
The open-source app everyone uses to manage agents at work
中文介绍 Paperclip是一个开源应用，用于管理工作中的智能代理。
Python · ★ 40,862 · 🍴 5,523 · 📈 4,413 stars today
Hindsight: Agent Memory That Learns
中文介绍 Hindsight是一个智能代理记忆学习系统，用于增强智能代理的决策能力。
PLSQL · ★ 25,728 · 🍴 5,883 · 📈 145 stars today
Open-source, low-cost 10.5 GHz PLFM phased array RADAR system
中文介绍 PLFM_RADAR是一个开源、低成本的高频相控阵雷达系统，适用于10.5 GHz频段。
TeX · ★ 2,468 · 🍴 234 · 📈 316 stars today
Open Source Introductory Systems Programming Textbook for the University of Illinois
中文介绍 cs341-illinois/coursebook是一本开源的系统编程入门教科书，适用于伊利诺伊大学。
JavaScript · ★ 64,613 · 🍴 6,509 · 📈 310 stars today
An advanced guide which might benefit you a lot 🎉 . 韩先凯的人生进阶指南 人生进阶指南 离谱的人生 人生进阶 AI学习 AI指南 韩先凯的AI学习指南 英语学习指南/英语学习教程/英语学习/学英语
中文介绍 up是一份人生进阶指南，涵盖了AI学习、英语学习等内容，适合寻求个人成长和技能提升的学习者。
TypeScript · ★ 1,653 · 🍴 133 · 📈 781 stars today
Multi-agent harness that runs Claude Code and Codex together as one system
中文介绍 openrig是一个多智能体工具，可以将Claude Code和Codex结合运行，作为单一系统。
TypeScript · ★ 21,188 · 🍴 1,796 · 📈 1,105 stars today
The Office Harness for AI Agents — Spreadsheets, Docs, Slides, Canvas, Relational Tables, and PDF in one runtime.
中文介绍 univer是一个AI代理办公套件，支持电子表格、文档、幻灯片、画布、关系表和PDF等功能的运行环境。
👍 2
A main promise of looped language models is depth-adaptive inference. By looping a block of shared layers a variable number of times, the model can use less compute for "easy" tokens and more for "hard" ones. However, tokens with different numbers of loops cannot share a uniform forward pass and the
👍 3
Open rerankers trained for general web retrieval transfer imperfectly to e-commerce, where ranking decisions depend not only on topical relevance but also on user preferences, product constraints, and comparative product fit. These preference signals are difficult to supervise at scale: real search
👍 20
Scaling language models to long contexts is limited by the quadratic cost of self-attention. Block sparse attention offers an efficient alternative, but selecting the retained blocks remains a bottleneck. Conventional block selection requires scoring all query-block pairs and therefore remains quadr
👍 6
Existing multi-agent benchmarks primarily test in competitive settings, short-horizon interactions under 20 steps, or simply aggregate individual performance, failing to isolate and highlight genuine collaboration capabilities of LLM-based agents. We introduce AgentWorld, a benchmark of 100 human-an
👍 4
We introduce Kaggle Game Arena, an open and ever-expanding platform to evaluate large language models (LLMs) through competitive games. Different from static benchmarks, game arena enables models to play head-to-head matchups in structured environments where the gameplay strength naturally increases
👍 6
Long-horizon reasoning remains a central challenge for large language models (LLMs) under sparse-reward regimes. We argue that this brittleness arises from two biases induced by complex reasoning spaces: an exploration bias, where models are drawn toward locally plausible but structurally unstable b
👍 2
Public software repositories, like GitHub and Software Heritage Archive, store billions of files, yet extracting their implicit engineering knowledge ---i.e., the algorithms they implement, the paradigms they follow, the patterns they instantiate, and the application domains they serve--- remains ch
👍 4
Banking assistants must use account-specific information to answer requests and, in many cases, take actions through tools. Evaluating only the final response misses important errors. An assistant may ask for information it already has, rely on stale context, select the wrong account, or write an in
👍 8
Jev is a fast, low-cost decision model that answers natural-language questions with choices, binary judgments, and scores. As its public ecosystem grows rapidly, it remains unclear how Jev is used across applications and how public attention relates to project distribution. To answer these questions
👍 7
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 standard on-policy Reinforcement Learning (RL): algorithms like GRPO indiscriminately broadcast a h
👍 8
In this paper, we propose RGBD20K, a novel dataset for facilitating the development of more robust and general RGB-D semantic segmentation by encompassing abundant categories and high-quality annotations. RGBD20K possesses several attractive properties: (1) Expanded Semantic Space. In particular, it
👍 79
While Large Language Models (LLMs) rely on highly non-linear components, in this work we demonstrate that they exhibit fundamental linearity: when inputs from distinct text streams are linearly combined, the model outputs a superposition of the individual next-token distributions. We term this the S
👍 10
Detectors of alignment failures screen deployed language models and score alignment benchmarks. Most are generative judges that spend a decoding pass on every criterion, and classifiers that read token probabilities, such as Llama Guard, still score one fixed label per call. Jev, a model trained wit
👍 11
Task and motion planning (TAMP) problems remain difficult even with full observability and object-centric states because discrete decisions are tightly coupled to geometric, kinematic, and dynamic constraints. Generalized TAMP addresses this difficulty by exploiting regularities across problem insta
👍 21
Rufus-Air is an open and reproducible post-training recipe on GLM-4.5-Air-Base (106B-A12B), organized as a serial pipeline of eight stages: SFT, Reasoning RL, Coding RL, Instruction-Following RL, General Agent, Coding Agent, Search Agent, and RLHF. We document the data, reward design, infrastructure
👍 9
Deep search requires LLM agents to decompose complex queries, search for evidence, and synthesize grounded answers, yet existing ReAct-style agents suffer from two limitations: role coupling, where one policy must handle planning, evidence use, and synthesis; and context accumulation, where growing
👍 11
We introduce PUBG Ally, an embodied agent for PUBG: BATTLEGROUNDS that can reason, act autonomously, and play alongside players as a voice-enabled teammate. Building such a teammate requires combining two difficult capabilities: it must perceive and respond to a constantly changing game world under
👍 7
General-purpose vision-language models (VLMs) bring broad knowledge and spatial reasoning to robot manipulation, yet existing systems either use them indirectly, to predict constraints or write programs, or give them a view of the scene rather than a world in which to act. We present World Action Ag
👍 13
Recently, Large Language Models (LLMs) have been increasingly able to solve advanced mathematical problems, including many that have been open for decades. This opens the door to expansion of mathematical knowledge at unprecedented scale. Yet, while LLMs may be able to conjecture and prove more and
👍 5
World-action models (WAMs) transfer visual and motion priors from pretrained video generators to robot control by jointly modeling visual dynamics and actions. Existing WAMs, however, predict dense future frames during training, repeatedly modeling largely unchanged content and coupling action-condi
👍 209
Object permanence and solidity are hallmarks of human cognitive priors. Recent studies show that video generation models, a paradigmatic class of current world models, have begun to show emerged reasoning abilities, making them ideal candidates for building human-like physical intelligence. Do video
👍 42
Prefill and decode reward different approaches to quantization: low-precision arithmetic accelerates prompt processing, while compact weights reduce memory traffic during generation. We propose "disaggregated quantization" (DQ), which specializes computation formats, weights and storage placement to
👍 1
Latent world models predict the consequences of actions, but accurate prediction does not guarantee that latent distance reflects which candidate will execute successfully. We identify a decision-local prediction gap: among the few futures competing for execution, a candidate predicted closer to the
👍 2
Standard positional encodings represent position as a one-dimensional reading-order coordinate, but reading order alone does not determine hierarchical textual structure. We use a hierarchical rotary positional encoding (hRoPE) that represents paragraph, sentence, and token indices as separate chann
👍 4
Adapting large language models to individual users remains challenging due to the tension between fine-grained personalization and scalable deployment. We present CARD, a hierarchical framework that achieves effective personalization through progressive refinement. CARD first clusters users accordin
👍 4
Personalized large language models are often expected to follow explicit style instructions, yet we find that such instructions can undermine the user-specific characteristics that personalization methods aim to preserve. We call this failure mode personalization collapse: explicit style control can
👍 24
Humans can effortlessly localize the direction of a sound source and integrate it with visual cues for reasoning, yet this remains challenging for embodied agents. In particular, it is still unclear how to effectively evaluate and model spatial audio understanding in embodied settings. To address th
👍 8
Modern Transformer design and compression both reduce to allocating capacity under a budget. The standard scalars for these decisions, #Params and #FLOPs, capture size and compute but not architectural structure: two architectures with identical parameter budgets but different depth-width, head, or
👍 1
Merging low-rank adapters (LoRAs) promises to eliminate the overhead of swapping task-specific weights at inference time. However, existing merging methods assume every layer needs the same rank budget. Further, some methods assume that rank budget needs to be split equally among the tasks too. We s
👍 7
Modern AI agents routinely cross trust boundaries: they ingest untrusted content, combine it with privileged instructions, persist intermediate beliefs in long-term memory, and invoke privileged tools. This creates an attack surface in which malicious payloads can enter through model inputs and caus
@DhravyaShah · 63.2K 粉丝 · 112.4K 阅 · 608 赞 · 45 转
It's been just 2 weeks since we discontinued the company brain harness. We had to offboard tons of people to other products, but many of our customers, and others suggested us to open source our work
中文介绍 DhravyaShah分享公司脑力工具开源设计过程，回顾了停用工具后的转型挑战。
@saylor · 5.2M 粉丝 · 96.2K 阅 · 661 赞 · 98 转
Artificial intelligence will make it possible for individuals and companies to produce far more than they can today. That makes the freedom to create, finance, own, and exchange things more important.
中文介绍 saylor探讨数字时代人工智能对个体和公司生产力的影响，强调创造、融资、拥有和交换的自由。
@0xMovez · 36.4K 粉丝 · 78.9K 阅 · 737 赞 · 46 转
Most people who try motion design with Opus 5.5 end up with the same video: centered text on a gradient, everything fading in, a logo at the end. They don't give it a reference, don't give it a
中文介绍 0xMovez介绍如何使用Opus 5.5构建动态设计工作室，指出常见设计误区。
@joedaroo · 12.2K 粉丝 · 41.2K 阅 · 517 赞 · 76 转
A lot of the perspective on all the AI incidents has been shared from the outside in, and little has been said from the inside looking out, through the lens of a security person living through it.
中文介绍 joedaroo从安全人员的视角分析AI事件，强调内部视角的重要性。
@RaphaelAubryy · 393 粉丝 · 40.9K 阅 · 507 赞 · 42 转
Most people trying motion design with Opus 5.5 hit the same wall: ↳ the first render is "mid": centered text, a gradient, everything fading in ↳ the music and the picture live separate lives ↳ one
中文介绍 RaphaelAubryy开源使用Opus 5.5制作动态视频的整个工作流程，分享3天完成10+视频的经验。
@twoclipping · 7.5K 粉丝 · 40.5K 阅 · 510 赞 · 35 转
im open sourcing my entire workflow for making the motion designs that studios charge $5,000 to $15,000 for this uses 0 mcp, 0 external tools and im not selling you any subscription you dont need
中文介绍 twoclipping开源制作价值1万美元启动视频的工作流程，无额外工具和订阅费用。
@lilykeilani011 · 3.0K 粉丝 · 10.7K 阅 · 521 赞 · 71 转
The Problem Isn't Always the AI It's no secret that rejection of AI-generated art is highly visible on social media. Open any post related to AI art and you'll find hundreds of negative comments.
中文介绍 lilykeilani011讨论AI艺术的社会接受度，指出对AI艺术的反感往往是对自身喜好的否认。
This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. Last Wednesday, Anthropic announced that earlier this year it had launched a molecular biology lab, where Claude agents read and conjecture about hard biology pro
中文介绍 Anthropic宣布已成立分子生物学实验室，其Claude代理读取并推测科学发现。
中文介绍 Hugging Face推出Holo4通用计算机使用代理。
MIT Technology Review Explains: Let our writers untangle the complex, messy world of technology to help you understand what’s coming next. You can read more from the series here. Over the past few months, a cascade of cyberattacks by AI agents has stunned the world. In July, OpenAI disclosed that a
中文介绍 MIT Technology Review探讨AI代理失控时的责任归属。
OpenAI is expanding the Lenfest AI Collaborative and Fellowship Program with $5 million in funding and up to $5 million in software credits and engineering support.
中文介绍 OpenAI扩大Lenfest AI合作计划和奖学金项目，提供500万美元资金及软件和工程支持。
We’re collecting real stories of builders, tinkerers, researchers, and creators who are using Codex to do incredible things. If you want to be a part of the next chapter of the Codex Originals program, tell us more about your story and project below.
中文介绍 OpenAI征集使用Codex进行创新项目的真实故事。
GPT-6 Astra completed a 50-tab tax workbook twice as fast as GPT-5.6 Sol, and its stronger understanding of user intent gives Basis more confidence in real-world use.
中文介绍 GPT-6 Astra完成50页税务工作簿速度是GPT-5.6 Sol的两倍，并增强了对用户意图的理解。
中文介绍 TLDR AI报道Muse实时头像、DeepSeek 10亿美元年收入等AI新闻。
In 2023 most people doubted that there could be more than 1 or 2 frontier model labs. Now there are dozens.... and Stripe just bought the best known one for $7B.
中文介绍 Stripe以70亿美元收购了OpenRouter和AMP，标志着前沿模型实验室的发展。
With Codex, GPT-Live-1, and GPT-6 Astra, Proaction builds, operates, and sells modern fleet management faster.
中文介绍 Proaction利用Codex、GPT-Live-1和GPT-6 Astra提高销售效率并节省时间。
The US government wants to spend $30.3 million over the next five years on an improved form of lie detector, according to a Department of Defense budget request. The program, called Polygraph+ or Polygraph Next, will focus on scoring algorithms that use artificial intelligence and machine learning a
中文介绍 美国国防部计划投资30亿美元研发改进型测谎仪，使用人工智能进行评分算法。
A quiet day lets us discuss the work behind the scenes - now open for business!
中文介绍 Latent Space讨论其背后的工作，并宣布对外开放。
GWM Worlds 2 uses persistent context and timed actions to steer a world model generating video and audio in real time.
中文介绍 Runway的WorldPrompt利用持续上下文和定时动作生成实时视频和音频的世界模型。
中文介绍 TLDR AI报道ChatGPT Pro Max、Muse实时头像、DeepSeek 10亿美元年收入等AI新闻。
Guest Post: In science, thinking has gotten cheap but doing has not. This asymmetry is reshaping how research companies operate, largely inconspicuously.
中文介绍 科学研究中思考成本低而执行成本高，这种不对称性正在重塑研究公司的运营。
中文介绍 Hugging Face介绍LFM2.5-VL-DSpark加速视觉语言模型。
Two villages and a town near R.A.F. Fairford were thrust into the spotlight after the authorities said they were investigating a possible terrorist plot targeting the base.
中文摘要 英国当局调查针对美国空军基地的恐怖袭击阴谋，附近村庄受到关注。
Five British men who were detained near R.A.F. Fairford, a base used by American forces in the war against Iran, were being released on bail but remained under investigation.
中文摘要 五名在英国空军基地附近被捕的英国男子被释放，但仍在调查中。
Follow the day’s news live Get our breaking news email, free app or daily news podcast Chalmers doesn’t want Australia to ‘carry the can’ for decisions on other side of the world Chalmers said he doesn’t believe inflation is out of control, but he doesn’t want to see Australians “carry the can for d
中文摘要 澳大利亚财长Chalmers在现金利率决定前捍卫其记录。
Students are protesting over issues such as overcrowded classes, but the prime minister has condemned reports of violence.
中文摘要 法国总理警告学校抗议升级，164人被捕。
Tehran has been trying to step up attacks on targets in the west but its track record has been mixed Investigators may be some way from determining exactly what led five men with three vans to a village just outside RAF Fairford in the small hours of Sunday – but the fact they have all been released
中文摘要 关于伊朗与涉嫌英国空军基地炸弹袭击的联系仍存在疑问。
Gold hits seven-week low; silver follows suit and records a nearly 5 percent loss.
中文摘要 金价下跌，受油价上涨和美元走强影响。
Protesters have set up tent encampments in central Madrid to demand government action on soaring housing costs.
中文摘要 数百人在马德里设立帐篷营地，抗议高昂的住房成本。
Bayern Munich star Olise set off from inside his own half to score the winner in the 88th minute.
中文摘要 Olise在最后时刻进球，法国1-0击败比利时，赢得国家联赛胜利。
Two giant pandas from China have landed in the United States and are heading to Zoo Atlanta.
中文摘要 “友谊熊猫”抵达亚特兰大，标志着中美关系的一刻。
Mediapart's explosive report comes a day after France's National Rally party secures a record number of Senate seats.
中文摘要 法国极右翼领导人Bardella否认反犹主义指控。
Speaking at the UN General Assembly, South Sudan’s UN envoy Cecilia Adeng called for sanctions to be lifted
中文摘要 南苏丹在联合国大会上呼吁解除制裁。
In a faltering Chinese economy, working-class stories are striking a chord.
中文摘要 在放缓的中国经济中，工人阶级的故事引起共鸣。
Slovenia’s policy towards Palestine has changed sharply following a UNGA sideline meeting with Israeli PM Netanyahu.
中文摘要 斯洛文尼亚对以色列的政策发生转变。
At a youth hackathon in Gaza City, young inventors are developing robots.
中文摘要 加沙的年轻发明家利用现有资源制作机器人。
Turkish comedian Deniz Göktaş also faces a travel ban but has been released due to time served.
中文摘要 因侮辱埃尔多安被判罪的喜剧演员Deniz Göktaş获释。
Oil rose a second day as uncertainty over progress in US-Iran talks weighed against signs of strong demand and a resumption of flows through a key pipeline from top exporter Saudi Arabia.
中文摘要 油价连续第二天上涨，受美伊谈判进展不确定影响，抵消了强劲需求和沙特阿拉伯关键管道恢复流动的迹象。
World Labs was founded by Stanford University researcher to work on AI models that understand 3D environments
中文摘要 AMD 以 80 亿美元收购了斯坦福大学研究员创办的 AI 创业公司 World Labs。
Summit Therapeutics Inc. shares rose after it said AstraZeneca Plc would make a $2 billion strategic equity investment in the cancer drugmaker to jointly develop drugs.
中文摘要 Summit Therapeutics 股价上涨，因其宣布阿斯利康将投资 20 亿美元进行战略合作，共同开发抗癌药物。
Former Walt Disney CEO Bob Chapek says the firm needs a 'growth engine' while speaking during an interview on Bloomberg Businessweek Daily. Chapek also says that Disney employees have been discouraged from keeping in contact with him and that he hasn't heard from CEO Josh D'Amaro since 2022. Chapek
中文摘要 前迪士尼 CEO Bob Chapek 表示公司需要增长动力，同时透露与 CEO Josh D'Amaro 的沟通受限。
People Pleasing, Procrastination, Perfectionism, and Padding -- the four key barriers that author Fran Hauser says many women have trouble overcoming. She explains how she learned to reclaim her own time during her decades of executive experience in her new book 'Twenty Minutes Is the New Hour: The
中文摘要 作者 Fran Hauser 在新书《二十分钟：新一小时》中讨论了女性克服拖延、过度追求完美等四个关键障碍。
Policyholders of a struggling life insurer sued its private equity owner, accusing Golden Gate Capital of self-dealing and mismanaging PHL Variable Insurance Co. to the point that it now faces possible liquidation.
中文摘要 保单持有人起诉私募股权公司 Golden Gate Capital，指控其恶意经营 PHL Variable Insurance Co. 导致其资本短缺 22 亿美元。
Also in today’s newsletter: Nvidia’s $150bn share buyback and Tata family scion hits back with plan to keep holding company private
中文摘要 首尔指责乌克兰违反了关于朝鲜士兵的秘密协议。此外，英伟达宣布 1500 亿美元的股票回购计划，塔塔家族计划保持控股公司私有。
On today’s Big Take podcast, Bloomberg’s David Papadopoulos and Michael Regan join host Sarah Holder to discuss whether we’re on the brink of a great portfolio reallocation, two very different reasons why bond yields are rising, and why one prominent AI skeptic is embracing the trade he once shunned
中文摘要 Bloomberg 的 David Papadopoulos 和 Michael Regan 在 Big Take 播客中讨论了市场是否即将迎来大规模投资组合再平衡等问题。
$1.4bn sale of Hong Kong lender’s loan portfolio part of push to improve its capital efficiency
中文摘要 汇丰银行通过出售 14 亿美元贷款组合来改善其资本效率。
South Korean banks and insurers are hunting for overseas deals to lift growth, signaling a broader outward turn by an industry that largely stayed home while companies including Samsung Electronics Co. and Hyundai Motor Co. built global businesses.
中文摘要 韩国金融机构寻求海外并购以推动增长，表明该行业正逐步向外扩张。
Australia’s mining stocks are poised to extend their rally, analysts say, as surging copper demand for power infrastructure and artificial intelligence gives investors a new reason to buy the sector.
中文摘要 分析师表示，由于电力基础设施和人工智能对铜的需求激增，澳大利亚的矿业股票看涨。
Shares in aerospace group fall 6 per cent on Monday to $184.
中文摘要 软件故障将推迟波音 737 新型号的美国批准，导致航空航天集团股价周一下跌 6%，至 184 美元。
7 回复 · 程序员 节点
6 回复 · 程序员 节点
13 回复 · 程序员 节点
9 回复 · Apple 节点
22 回复 · Apple 节点
17 回复 · Apple 节点
29 回复 · Apple 节点
15 回复 · Apple 节点
14 回复 · Apple 节点
17 回复 · Apple 节点
该源今日无内容。