每日简报

2026-08-16

← 历史归档

cordiverse/cordis

TypeScript · ★ 4,048 · 🍴 198 · 📈 616 stars today

Meta-Framework of Spatiotemporal Composability

中文介绍 cordis 是一个时空可组合性的元框架,用于开发时空应用。

cathrynlavery/diagram-design

HTML · ★ 18,579 · 🍴 1,118 · 📈 1,619 stars today

29 editorial diagram types for Claude Code. Self-contained HTML + SVG. No shadows, no Mermaid-slop.

中文介绍 diagram-design 提供了29种编辑图表类型,支持自包含的HTML和SVG格式,无需阴影和Mermaid工具。

cursor/plugins

TypeScript · ★ 2,948 · 🍴 233 · 📈 152 stars today

Cursor plugin specification and official plugins

中文介绍 cursor/plugins 定义了Cursor插件的规范并提供官方插件。

cactus-compute/needle

Python · ★ 6,056 · 🍴 402 · 📈 551 stars today

14MB foundation model for tiny devices; phones, wearables, smart home, and robots.

中文介绍 needle 是为小型设备设计的14MB基础模型,适用于手机、可穿戴设备、智能家居和机器人。

unslothai/unsloth

Python · ★ 72,035 · 🍴 6,492 · 📈 435 stars today

Local UI to run and train LLMs and diffusion models, including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, FLUX and more.

中文介绍 unsloth 是一个本地用户界面,用于运行和训练LLM和扩散模型,包括Qwen3.8、Kimi K3等。

public-apis/public-apis

Python · ★ 460,140 · 🍴 50,852 · 📈 2,476 stars today

A collective list of free APIs

中文介绍 public-apis 收集了免费的API列表。

MakazhanAlpamys/Soup

Python · ★ 1,643 · 🍴 262 · 📈 303 stars today

Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU.

中文介绍 Soup 允许用户通过YAML文件微调LLM,支持在4GB笔记本电脑GPU上训练8B模型。

github/spec-kit

Python · ★ 129,179 · 🍴 11,548 · 📈 901 stars today

💫 Toolkit to help you get started with Spec-Driven Development

中文介绍 spec-kit 是一个帮助用户开始使用Spec-Driven Development的工具包。

megadose/holehe

Python · ★ 13,109 · 🍴 1,742 · 📈 389 stars today

holehe allows you to check if the mail is used on different sites like twitter, instagram and will retrieve information on sites with the forgotten password function.

中文介绍 holehe 可用于检查电子邮件在不同网站上的使用情况,并从具有忘记密码功能的网站上检索信息。

altic-dev/FluidVoice

Swift · ★ 10,313 · 🍴 698 · 📈 165 stars today

Fastest and only macOS Dictation app with on-device STT and custom trained AI enhancement model. A local Wispr Flow alternative. DM us on X for an easter egg 😉 - https://x.com/fluidvoiceapp

中文介绍 FluidVoice 是一个快速的macOS语音识别应用,带有设备上的STT和自定义训练的AI增强模型,可作为Wispr Flow的本地替代品。

ToolJet/ToolJet

JavaScript · ★ 39,524 · 🍴 5,294 · 📈 553 stars today

ToolJet is the open-source foundation of ToolJet AI - the enterprise app generation platform for building internal tools, dashboard, business applications, workflows and AI agents 🚀

中文介绍 ToolJet 是一个开源平台,用于构建企业级应用、仪表盘、业务应用程序、工作流程和AI代理。

HKUDS/CLI-Anything

Python · ★ 47,339 · 🍴 4,393 · 📈 100 stars today

"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/

中文介绍 CLI-Anything 是一个将所有软件转换为代理本地的工具,通过CLI-Hub提供统一的接口。

citrolabs/ego-lite

JavaScript · ★ 10,949 · 🍴 557 · 📈 546 stars today

The fastest browser for AI agents to run browser automation, built for sharing your logged-in browser state with your AI agents, like Codex or Claude Code, without disturbing you. Zero cost, zero config.

中文介绍 ego-lite 是一个专为AI代理运行浏览器自动化而设计的快速浏览器,允许将登录的浏览器状态与AI代理共享,无需打扰用户。

DreamX-Phi 1.0: Action-Conditioned Video World Model for Robotic Manipulation

👍 82

We present DreamX-Phi 1.0, an action-conditioned video world model for robotic manipulation that, given an observed frame, a language instruction, and a prescribed action sequence comprising end-effector poses and gripper states, predicts the resulting future observations. Yet realism alone does not

AutoDesign: Meta-Harness Optimization for Long-Horizon Agentic Design

👍 35

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 harness system should align with human design priors and accumulate reusable experience through empiric

OmniScientist: An Omni-Modal Omni-Discipline AI Scientist

👍 6

Recent advances in foundation models have enabled AI scientists to automate increasingly complete research workflows, from hypothesis generation and code execution to manuscript preparation. Yet workflow coverage alone does not provide access to the full evidence on which scientific discovery depend

Intern-S2-Preview: Scientific Agentic Foundation Model

👍 43

Scientific discovery increasingly requires AI systems that can reason over scientific evidence of heterogeneous modalities, interact with scientific tools and environments, and sustain progress across long task horizons. We present Intern-S2-Preview, a series of scientific agentic foundation models

Hybrid-Policy Self-Editing for Composable Unstructured Knowledge Editing

👍 4

Large language models (LLMs) achieve remarkable performance across natural language tasks, yet they are trained on static corpora and their knowledge quickly becomes outdated in a fast-changing world. This motivates knowledge editing (KE), which updates specific knowledge in an LLM without changing

Specification-first convergence with an AI coding agent: a case study of dismantling a core architectural invariant across 189 files in a 717k-line codebase with no test oracle and no human code review

👍 2

This paper reports a single, fully instrumented case study of a large-scale architectural refactoring by an AI coding agent under a specification-first protocol, with no human review of the generated code and no pre-existing oracle to validate the target behaviour. The task, dismantling a central in

AVA-Encoder: Towards Agent-Native Video Representation Learning

👍 6

Creative agents still lack an effective way to learn from high-quality human films, limiting their ability to produce cinematic-grade videos. A key challenge is the absence of a structured video representation that is both faithful to film content and directly usable for agentic reasoning and manipu

Thought-Level Beam Search for Reasoning

👍 6

Test-time compute scaling is a primary driver of performance in large reasoning models (LRMs), but extreme inefficiency bounds current approaches, shifting the critical question from how much compute to spend, to where to allocate it. We formalize test-time reasoning as a constrained compute allocat

Self-Evolving Embodied Agents via Skill-Harness Evolution

👍 11

Embodied agents are increasingly built as systems around foundation models, where performance depends not only on model weights but also on the skills, context, action interfaces, and execution harness surrounding the model. While supervised fine-tuning and reinforcement learning can adapt agents to

Parameter Exploration for RLVR via Variational Learning

👍 6

Exploration has been a focus of reinforcement learning research for a long time. Recently, there has been growing evidence that it is also an important ingredient in LLM reinforcement learning recipes that can significantly impact downstream performance. Many existing methods control exploration in

Mitigating Gender Bias in English to Romanian Machine Translation

👍 0

Machine translation (MT) systems often fail to correctly translate gender, especially when converting from a gender-neutral language like English to a gendered target language such as Romanian. This bias results in translations that default to masculine forms or reinforce gender stereotypes. We prop

Full-bandwidth transformer

👍 10

Autoregressive transformers compute along two axes: horizontally across generated tokens, and vertically through model depth. Dense attention gives each token broad horizontal access to the past, but the vertical feedback channel between decoding steps remains narrow: only the sampled token returns

An AI4AI Framework for Visual Token Pruning

👍 7

Visual-token pruning can substantially reduce the inference cost of multimodal large language models (MLLMs), yet existing methods largely rely on fixed, handcrafted heuristics and costly expert trial and error. As pruning objectives, budgets, and model architectures diversify, manually navigating t

Maglev: Sliding Recurrent Memory

👍 5

We introduce , a recurrent Transformer architecture with fixed-size memory that generalizes sliding-window attention while remaining parallelizable during training. consists of two coupled models: a prefiller Q, which leverages full attentionIn practice, we use interleaved full and sliding-window a

Knowing When to Quit: Diagnosing and Training LLMs to Abort Futile Reasoning

👍 3

Large language models generate computationally expensive yet semantically void reasoning on beyond-capability tasks, creating risks where plausible-sounding but incorrect derivations mislead users. We characterize this futile reasoning phenomenon through systematic analysis, revealing universal capa

DarwinX: Evolving Agent Harnesses Through Natural Selection

👍 66

An LLM agent's capability depends not only on model weights but on its harness: prompts, tools, skills, and control flow. Self-improvement loops already edit harnesses, yet single-lineage search is path-dependent and local wins often regress other tasks. We introduce DarwinX, which treats self-evolu

Poor Man's Agentic Modeling: Simulating Large LLM-Agent Societies on a Laptop

👍 4

Simulating societies of many large language model (LLM) agents is expensive, yet the questions asked of such simulations are usually macroscopic: phase behaviour, stylised facts, and scaling with the number of agents N, not the cognition of any single agent. We turn a statistical-physics observation

Cursor + SpaceXAI: the fastest iterating team wins

@sarahdingwang · 18.0K 粉丝 · 1.2M 阅 · 514 赞 · 73 转

By @BornsteinMatt, @martin_casado, and myself. AI coding could be the biggest addressable market of our lives. And, unless you think we’ve fully addressed it (which no one thinks; we are more like 1%

中文介绍 探讨AI编码市场潜力,强调目前只开发了1%,Cursor与SpaceXAI合作推进AI技术迭代。

Agent Plugins are the future of Agent Skills

@GoogleCloudTech · 1.3M 粉丝 · 56.7K 阅 · 505 赞 · 76 转

Agent Plugins is an open, vendor-neutral standard for packaging Agent Skills and the MCP servers they depend on into one portable folder that any compatible client can load. Google is joining the

中文介绍 介绍Agent Plugins作为开放标准,Google Cloud加入该标准,旨在简化Agent技能打包和部署。

The Elon–Jensen "Understanding"

@TeslaLarry · 53.1K 粉丝 · 56.5K 阅 · 507 赞 · 54 转

After the SpaceX earnings call on August 4, several pieces started lining up for me in a way that is hard to ignore. Elon didn’t just praise Nvidia hardware, he declared an exclusive commitment to

中文介绍 分析SpaceX财报会议,指出Elon Musk对Nvidia硬件的独家承诺,涉及AI技术合作。

Preparing GLM-5.3 for Open Release: A Responsible Path to Cyber Defense

@Zai_org · 136.2K 粉丝 · 42.7K 阅 · 583 赞 · 50 转

When GLM-5.2 helped Hugging Face investigate an incident in which an AI autonomously bypassed its own safeguards, it highlighted a broader shift. AI is becoming part of both cyber offense and cyber

中文介绍 GLM-5.2事件凸显AI在网络安全中的双重角色,Zai_org为GLM-5.3开源做准备。

The AI Engineering Skills Map

@AndrewYNg · 1.8M 粉丝 · 41.2K 阅 · 716 赞 · 89 转

I am delighted to present The AI Engineering Skills Map. AI allows us to build software very differently today than in 2022, and everyone with the skills to take advantage of this shift has numerous

中文介绍 发布AI工程技能图谱,探讨AI技术如何改变软件开发,并强调技能的重要性。

What's new in Gemini 3.7 Flash

@GoogleAIStudio · 192.9K 粉丝 · 31.5K 阅 · 535 赞 · 34 转

Gemini 3.7 Flash (gemini-3.7-flash) is generally available (GA) and ready for production use. It is our most intelligent workhorse model yet for software engineering, and agentic workflows. This guide

中文介绍 Gemini 3.7 Flash模型发布,适用于软件工程和代理工作流,强调其智能化和实用性。

Introducing Gemini 3.7 Flash

@GoogleAIStudio · 192.9K 粉丝 · 30.7K 阅 · 566 赞 · 60 转

Today, we’re building on the progress of our widely used Flash series by introducing Gemini 3.7 Flash, our most intelligent workhorse model yet for coding and agents. This release comes just three

中文介绍 介绍Gemini 3.7 Flash模型,作为最新智能模型,适用于编码和代理工作。

Qwen3.8-27B Practical Guide: Control Reasoning Depth and Extend Context to 1M Tokens

@QwenDevs · 8.0K 粉丝 · 25.6K 阅 · 559 赞 · 51 转

Qwen3.8-27B is here — a compact, deployment-friendly dense model built on the Qwen3.5 architecture, with native vision-language understanding and flexible thinking control. Two developer-facing

中文介绍 发布Qwen3.8-27B模型指南,强调控制推理深度和扩展上下文能力。

ORACLE: Confirmed AI Bots Actively Trading on Polymarket — Gains Reaching 200% Within 24 Hours

@OracleFoundtion · 19.4K 粉丝 · 18.8K 阅 · 3.1K 赞 · 615 转

In 2026, autonomous AI agents have become one of the most effective strategies on prediction markets. Over 30% of all activity on Polymarket now comes from algorithmic and AI-powered wallets. We

中文介绍 报告AI机器人活跃在Polymarket,24小时内收益达到200%,分析其在预测市场中的策略。

React for Agents: Astro Creator Brings Hooks to his Meta-Harness, Flue

Flue 2 takes its inspiration from React. Creator Fred Schott, of Astro fame, tells Latent Space why he added hooks and why agents are defined by their harnesses.

中文介绍 Astro Creator Fred Schott 向 Latent Space 解释了为何在 Flue 2 中加入 hooks,以及为何代理由其 harness 定义。

Flock is tightening its rules in response to a growing surveillance backlash

The police-tech giant Flock is announcing today that it will change officers’ access to its nationwide network of license plate readers, in an apparent effort to quell a growing backlash and win back contracts lost amid concerns about mass surveillance and police abuse. Several changes aim directly

中文介绍 Flock 为应对日益增长的监控反感和警方滥用问题,调整了全国车牌读取网络的访问权限。

The builder’s guide to GPT‑5.6

Learn how startups use GPT-5.6 to build faster, more cost-efficient AI agents with smarter model selection and new Responses API capabilities.

中文介绍 OpenAI 发布 GPT-5.6 构建者指南,介绍初创公司如何使用 GPT-5.6 构建更快、更高效的 AI 代理。

Previewing Ultrafast mode: GPT-5.6 Sol at up to 14X the speed

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.

中文介绍 OpenAI 预览 Ultrafast 模式,GPT-5.6 Sol 的速度提高至 14 倍。

OpenAI appoints Dali Rajic as Chief Revenue Officer

OpenAI appoints Dali Rajic as Chief Revenue Officer to lead its global revenue organization and help businesses realize the full value of AI.

中文介绍 OpenAI 指任 Dali Rajic 为首席收入官,领导全球收入组织。

How kids feel about AI, in their own words

When we set out to talk to kids about artificial intelligence, we thought we knew what we’d hear. We expected some to tell us they were using it to cheat a little, the way Millennials and Gen Xers opened up CliffsNotes or programmed formulas into their TI-82s, and others to share inspiring ways they

中文介绍 MIT Tech Review AI 发表关于孩子对 AI 的看法的文章。

not much happened today

**Google** rapidly released **Gemini 3.7 Flash** just three weeks after 3.6 Flash, targeting coding, web development, knowledge work, and agentic workflows with a 50% introductory price cut and improved benchmark scores like **DeepSWE 65.3%** and **Code Arena Elo 1588**. The update quickly integrate

中文介绍 Google 在 3.6 Flash 三周后迅速发布 Gemini 3.7 Flash,针对编码、网络开发、知识工作和代理工作流程,提供 50% 的入门价格折扣和改进的基准分数。

[AINews] SpaceXAI Grok 4.6 and Grok @Bot

AI teammate category just had its most significant new entrant yet

中文介绍 SpaceXAI 推出 Grok 4.6 和 Grok @Bot,成为 AI 同事类别中的新成员。

Australia news live: Pauline Hanson opposes major gun buyback agreed in wake of Bondi terror attack

Legislation for plan passed parliament earlier this year and is set to commence in NSW from November. Follow live updates Get our breaking news email, free app or daily news podcast ‘Additional arrangements’ being worked on for children removed from NDIS, minister says Asked what happens to a nine-y

中文摘要 澳大利亚新南威尔士州议会已通过大规模枪支回购计划,将于11月实施。

Why is the latest Ebola outbreak worsening?

Virus spreading in DRC faster than containment efforts.

中文摘要 刚果(金)埃博拉疫情扩散速度超过控制措施,疫情恶化。

Australian state to launch gun buyback after Bondi Beach attack

New South Wales, where the attack took place, will offer gun owners cash to give up their firearms.

中文摘要 新南威尔士州将在邦迪海滩袭击后启动枪支回购计划,向枪主提供现金以放弃枪支。

$10 Million in Gold Found Buried Under Former Brewery in Belgium

A construction worker’s drill struck gold during a renovation project in East Flanders, setting off a search for the owners of the stashed fortune.

中文摘要 在比利时一家前啤酒厂下发现价值1000万美元的金条和金币,一名建筑工人在翻新工程中意外发现。

Qatar denies it is detaining three Iranian bomber pilots

Iran says Qatar is holding three pilots who were shot down in March. Doha decries 'misleading statements'.

中文摘要 卡塔尔否认拘留了三名在3月份被击落的伊朗轰炸机飞行员。

Bond Traders Agonize Over AI Companies’ $70 Billion of Shadow Credit Backstops

Even before Nvidia Corp.’s splashy $500 billion financing partnership this week, investors were starting to fret over the roughly $70 billion in phantom liabilities that don’t appear on major AI companies’ balance sheets, but could materialize at the worst possible time.

中文摘要 投资者担忧约 700 亿美元的隐形负债可能出现在主要人工智能公司的资产负债表上。

Pointed! Bloomberg's Weekly News Quiz For Risk-Takers

Pointed offers a strategic twist to the news quiz format, testing not just players’ knowledge of the news but also their confidence in their answers. Join Bloomberg's Joe Mathieu, Carol Massar and Lisa Mateo as they play and check out the quiz for yourself at Bloomberg.com (Source: Bloomberg)

中文摘要 《点睛》系列新闻测验引入策略元素,测试参与者对新闻的理解和自信。

Colombia Quake Tests New President Days Into Term

Rescuers in western Colombia are still searching for survivors after a 7.4 magnitude earthquake killed at least 287 people. Bloomberg News Latin America Government Reporter Mie Dahl joins Bloomberg This Weekend from Bogotá and explains to hosts Joe Mathieu and Carol Massar that the disaster has also

中文摘要 哥伦比亚新总统上任数日即遭遇7.4级地震,救援人员正在搜救幸存者。

Drone Swarms Push AI Deeper Into Modern Warfare

Bloomberg News Cyber Security & Tech Reporter Katrina Manson got a firsthand look at autonomous drone-swarm technology that allows multiple aircrafts to coordinate and strike targets even when communications or GPS are disrupted. Joining Bloomberg This Weekend, Mason explains to Carol Massar that th

中文摘要 自主无人机群技术将人工智能进一步融入现代战争。

Lakers Deal Caps Billionaire Rush Into Sports

Bob Iger and Josh Kushner moved from pursuing a Las Vegas NBA expansion team to agreeing within days to buy the Los Angeles Lakers in a record $12.5 billion deal. Bloomberg News Host of Sports City Randall Williams is on Bloomberg This Weekend with hosts Joe Mathieu and Carol Massar and explains tha

中文摘要 伊格尔和库什纳以125亿美元的价格收购洛杉矶湖人队,标志着亿万富翁涌入体育产业的又一里程碑。

Strong Earnings Keep PNC’s Agati Bullish on Stocks

PNC's Chief Investment Officer Amanda Agati is on Bloomberg This Weekend with hosts Joe Mathieu and Carol Massar and says that strong earnings and economic fundamentals continue to support US stocks, even as elevated Treasury yields and midterm uncertainty threaten to bring more volatility. (Source:

中文摘要 PNC首席投资官Amanda Agati认为强劲的业绩和良好的经济基本面将继续支持美股。

US-Iran War Risks Repeated Cycles of Fighting

Shifting US objectives and neither side able to secure a decisive military victory make a lasting resolution difficult and Bloomberg Economics Defense Lead Becca Wasser says risks the the US-Iran conflict becoming a protracted war. Joining Joe Mathieu and Carol Massar on Bloomberg This Weekend, Wass

中文摘要 美伊冲突难以取得决定性胜利,可能导致冲突长期化。

By The Way: Headline You May Have Missed

Mystery vacations, college campuses for sale, buying sorority unwritten rules, and half off sushi: Bloomberg News' Lisa Mateo joins Joe Mathieu and Carol Massar on Bloomberg This Weekend to share headlines you may have missed from over the week. (Source: Bloomberg)

中文摘要 Bloomberg This Weekend分享一周中可能错过的新闻标题。

Trump's ‘Golden Age’ Pitch Meets Affordability Frustration

President Donald Trump is touting a “golden age” for the US economy, but Bloomberg News White House Correspondent Catherine Lucey is on Bloomberg This Weekend and says persistent concerns over gas, food and housing costs are complicating that message ahead of the midterms. Lucey explains to hosts Jo

中文摘要 特朗普宣称美国经济处于“黄金时代”,但高通胀和生活成本问题可能影响这一论调。

‘Outlier’ S&P 500 Earnings Strength Cheers Wall Street’s Bulls

Corporate America is blowing past Wall Street’s expectations for a stellar earnings season, adding to the rosy backdrop for a stock market trading near a record high.

中文摘要 企业美国业绩超出华尔街预期,为股市创纪录的交易提供有利背景。

Robinson Says Activists Need Leverage to Win

Author of the book 'From Presence to Power' Rashad Robinson is on Bloomberg This Weekend and explains to hosts David Gura and Christina Ruffini that visibility, media attention and public support do not amount to power unless activists build the leverage needed to change decisions and institutions.

中文摘要 Rashad Robinson认为,活动家需要建立杠杆才能取得胜利。

该源今日无内容。