2026-06-29（内容获取于 06/29 04:55）

Suno launches Spark incubator program to feed independent artists to its AI machine

The Verge · 06/29 04:27

Suno has ambitions to be more than just a toy to churn out AI slop, it also wants to be a streaming destination and to break new artists. Spark is their new incubator program for independent artists that provides grants, mentorship, and marketing support. To apply, artists need to be an unsigned sin

中文介绍 AI音乐公司Suno推出“Spark”孵化器项目，旨在支持独立艺术家，提供资助、指导和营销支持。Suno希望借此不仅成为AI音乐创作工具，还能发展为流媒体平台并发掘新晋艺术家。

Ford rehires ‘gray beard’ engineers after AI falls short

TechCrunch · 06/29 03:05

"Mistakenly we thought that by just introducing artificial intelligence ... that would produce a high-quality product.”

中文介绍 福特汽车因人工智能未能达到预期效果，重新聘用经验丰富的“白胡子工程师”。公司曾错误地认为「仅引入人工智能就能生产出高质量产品」，但现实情况证明AI尚无法完全替代人类经验。

Why did this journal retract two 1940s papers by Max Planck?

Ars Technica · 06/29 02:49

Clicking on the links now reveals blank pages and empty PDFs. "Intellectually, it’s not acceptable.”

China claims the world’s fastest supercomputer

The Verge · 06/29 01:20

The LineShine supercomputer at the National Supercomputing Center in Shenzhen. | Photo: Liang Xu/Xinhua via Getty Images Despite trade restrictions, China has reclaimed the title of the world's fastest supercomputer for the first time since 2018. LineShine has pushed El Capitan out of number one on

Writer Ian Bogost says ‘The Small Stuff’ can help us reclaim our lives from too much convenience

TechCrunch · 06/29 01:03

Has Silicon Valley been building the wrong things?

中文介绍 作家伊恩·博格斯特（Ian Bogost）在其作品中指出，过度追求便利可能导致生活“非物质化”，而“小物件”（The Small Stuff）反而能帮助人们重拾生活。文章探讨了硅谷是否正在构建错误的产品方向。

The Cube is Jim Henson’s little-known proto-Black Mirror masterpiece

The Verge · 06/29 00:30

That sure is a man in a cube, alright. | Image: NBC / Jim Henson Company I'm sure we're all familiar with Dark Crystal, so we know that Jim Henson can be weird and tackle slightly more mature subject matter. But there is little in his oeuvre that is quite as mind-bending as the Muppetless The Cube.

中文介绍 吉姆·亨森（Jim Henson）鲜为人知的作品《立方体》（The Cube）被誉为一部“原型《黑镜》式”的杰作。该片被认为是亨森最“烧脑”的作品之一，探讨了成熟主题，与他的《黑暗水晶》风格有所类似。

TechCrunch Mobility: All eyes on Tesla FSD

TechCrunch · 06/29 00:05

Welcome back to TechCrunch Mobility, your hub for the future of transportation and now, more than ever, how AI is playing a part.

中文介绍 TechCrunch Mobility栏目聚焦特斯拉（Tesla）的全自动驾驶（FSD）系统，强调人工智能在未来交通运输中扮演着日益重要的角色，吸引了业界的广泛关注。

Why Wall Street thinks US memory maker Micron is the next Nvidia

TechCrunch · 06/28 23:00

Eager to find more public AI-related companies that may do as well as Nvidia, Wall Street investors think they've found a winner with Micron.

中文介绍 华尔街投资者认为，美国存储芯片制造商美光科技（Micron）有望成为“下一个英伟达”（Nvidia）。投资者正积极寻找更多与AI相关的上市公司，认为美光在AI领域具有巨大的增长潜力。

Prosecutors used ChatGPT logs as evidence in the Palisades fire trial

The Verge · 06/28 22:12

Jonathan Rinderknecht was facing arson charges for setting a fire on New Year's Day in 2025, which became one of the deadliest wildfires in LA history. To make their case, prosecutors turned to location data from his iPhone, security camera footage, and witness testimony. But they also turned to his

中文介绍 2025年洛杉矶“帕利塞德斯”山火案中，检察官将ChatGPT的使用记录作为证据提交。被告乔纳森·林德内希特（Jonathan Rinderknecht）面临纵火指控，检察官除了iPhone定位数据、监控录像和证人证词外，还首次引入AI聊天记录作为呈堂证供。

Govee’s smart nugget ice maker makes every iced drink feel like a luxury

TechCrunch · 06/28 22:00

For some people, the ice in a beverage is almost as important as the drink itself. That’s the audience Govee had in mind when designing its latest ice maker, the GoveeLife Smart Nugget Ice Maker Pro. This $500 premium smart home gadget is aimed at those who crave what’s called “the good ice,” the so

中文介绍 Govee推出了一款售价500美元的高端智能家居产品——GoveeLife Smart Nugget Ice Maker Pro制冰机。该设备旨在满足那些对冰块品质有高要求的消费者，让他们享受制作“冰块粒”的奢华体验。

Nest’s quest to fix your thermostat

The Verge · 06/28 20:02

The founding story of Nest is pretty much a perfect tech myth. A legendary product maker (in this case, Tony Fadell) helps create one of the most successful products ever (the iPhone) and then rides off into the sunset to enjoy the rest of his life, only to have an experience that drags him back for

中文介绍 文章探讨了智能家居品牌Nest的创立故事。作为iPhone的重要设计师之一，托尼·法德尔（Tony Fadell）在离开苹果后，旨在通过Nest恒温器解决传统温控器问题，开启了智能家居的革新之路。

Ad-free streaming is a luxury now

The Verge · 06/28 20:00

This is The Stepback, a weekly newsletter breaking down one essential story from the tech world. For more news about the streaming industry, follow Emma Roth. The Stepback arrives in our subscribers' inboxes at 8AM ET. Opt in for The Stepback here. How it started Streaming was once a reprieve from c

中文介绍 《The Stepback》周报指出，无广告流媒体服务已成为一种奢侈品。这反映了流媒体行业的发展趋势，即免费或低价服务越来越多地通过插入广告来盈利。

TMD’s keyless bike lock is a $280 solution to a $60 problem

The Verge · 06/28 15:00

A $280 bike lock on a $10,000 e-bike. I've seen lots of so-called "smart" bike locks over the years, but none so far could justify the added cost. A newcomer that got its start securing ATMs for banks is trying to change that. There's nothing wholly unique about the TMD Chain Lock, but the combinati

中文介绍 专注于ATM安保的TMD公司推出一款售价280美元的智能无钥匙自行车锁。文章质疑其高昂价格的合理性，认为它“以280美元的方案解决了60美元的问题”，未能充分证明其相对于传统车锁的独特价值。

Indian payments chief thinks AI will be heavily involved in next era of digital payment growth

TechCrunch · 06/28 13:00

Dilip Asbe said that newer UPI apps could be more competitive with a viable commercial model

中文介绍 印度支付负责人迪利普·阿斯贝（Dilip Asbe）预测，人工智能将在下一阶段的数字支付增长中发挥关键作用。他认为，新的统一支付接口（UPI）应用程序若具备可行的商业模式，将更具市场竞争力。

Instagram is testing more ways to customize ‘Your Algorithm’

TechCrunch · 06/28 05:27

Instagram users could soon see more ways to tune their content.

中文介绍 Instagram正在测试新的功能，允许用户以更多方式自定义其内容推荐算法（“Your Algorithm”）。未来用户或将拥有更多工具来调整个人内容推送。

simplex-chat/simplex-chat

Haskell · ★ 14,853 · 🍴 858 · 📈 1,183 stars today

SimpleX - the first messaging network operating without user identifiers of any kind - 100% private by design! iOS, Android and desktop apps 📱!

中文介绍 SimpleX 是一个注重极致隐私的去中心化消息网络，设计之初就避免使用任何用户标识符，实现 100% 匿名通信。它通过独特的协议机制，让用户无需注册账号、无需手机号或邮箱，即可安全私密地进行消息交换。该项目提供 iOS、Android 和桌面端应用程序，适用于重视个人数据隐私和匿名交流的用户，尤其是在需要高度保密通信的场景。

ripienaar/free-for-dev

HTML · ★ 125,030 · 🍴 13,151 · 📈 472 stars today

A list of SaaS, PaaS and IaaS offerings that have free tiers of interest to devops and infradev

中文介绍 `free-for-dev` 是一个精心整理的资源列表，汇集了众多为开发者和运维工程师提供免费使用额度 (free tiers) 的 SaaS、PaaS 和 IaaS 产品及服务。该项目旨在帮助开发者和基础设施工程师发现和利用这些免费资源，以降低个人项目、原型开发或小型团队的运营成本。它涵盖了从云服务、数据库、CI/CD 工具到 API 网关等多个类别，对于预算有限的个人开发者、初创团队或学习新技术的工程师极具参考价值。

commaai/openpilot

Python · ★ 62,334 · 🍴 11,088 · 📈 265 stars today

openpilot is an operating system for robotics. Currently, it upgrades the driver assistance system on 300+ supported cars.

中文介绍 openpilot 是一个开源的机器人操作系统，专注于增强和升级现有车辆的驾驶辅助系统 (ADAS)。它通过先进的计算机视觉和机器学习算法，为超过 300 种受支持的车型提供车道保持、自适应巡航控制等功能。该项目旨在将普通汽车转变为具备更高自动化水平的智能驾驶平台，适用于希望提升车辆智能驾驶能力，或对自动驾驶技术进行研究和开发的爱好者及工程师。

xbtlin/ai-berkshire

Python · ★ 5,223 · 🍴 712 · 📈 1,456 stars today

AI 时代的伯克希尔：基于 Claude Code / Codex 的价值投资研究框架。巴菲特·芒格·段永平·李录四大师方法论 + 多Agent并行研究。| AI-era Berkshire: a value investing research framework built for Claude Code / Codex. 4 masters' methodologies + multi-agent adversarial analysis.

中文介绍 `ai-berkshire` 旨在构建一个 AI 时代的价值投资研究框架，其核心是借鉴巴菲特、芒格、段永平、李录四位投资大师的方法论，并结合 Claude Code 大模型技术进行实践。该项目利用多 Agent 并行研究机制，模拟专家分析师的工作流，对公司进行深度价值分析，识别潜在投资机会。它为量化投资者、AI 开发者和对价值投资感兴趣的用户提供了一个利用 AI 进行高级金融分析的实验性平台，以期在复杂的市场环境中做出更明智的投资决策。

Robbyant/lingbot-map

Python · ★ 8,154 · 🍴 798 · 📈 372 stars today

A feed-forward 3D foundation model for reconstructing scenes from streaming data

中文介绍 `lingbot-map` 是一个面向机器人和实时应用设计的 3D 基础模型，它利用前馈神经网络技术，能从连续的流式数据中高效重建三维场景。该项目旨在解决动态环境中实时感知和高精度地图构建的挑战，尤其适用于需要即时理解周围环境的机器人、自动驾驶系统或增强现实（AR）设备。通过提供一个强大的 3D 感知框架，它能够帮助这些系统实现更智能的导航、避障和交互能力，为未来的智能体和沉浸式体验提供核心空间感知能力。

DeusData/codebase-memory-mcp

C · ★ 19,390 · 🍴 1,409 · 📈 2,162 stars today

High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.

中文介绍 `codebase-memory-mcp` 是一个高性能的代码智能 MCP 服务器，专为快速理解和索引大规模代码库而设计。它能将多达 158 种编程语言的代码库构建成持久化的知识图谱，平均每个仓库的索引时间仅需数毫秒。该项目提供亚毫秒级的查询速度，并显著减少了处理所需的 token 数量（99%），非常适合需要深度代码理解的 AI 应用。作为单个静态二进制文件，它部署简便，能赋能开发者构建智能代码助手、高级搜索和分析工具。

cupy/cupy

Python · ★ 11,465 · 🍴 1,068 · 📈 172 stars today

NumPy & SciPy for GPU

中文介绍 `CuPy` 是一个用于 GPU 加速的 Python 库，它提供了与 NumPy 和 SciPy 兼容的 API 接口。该项目旨在帮助数据科学家、机器学习工程师和科研人员，能够通过最小的代码改动，将现有的 NumPy/SciPy 计算任务无缝迁移到 NVIDIA GPU 上运行。它解决了纯 CPU 计算在处理大规模数组运算和科学计算时性能瓶颈的问题，显著提升了数据处理和模型训练的速度，特别适用于深度学习、图像处理和物理模拟等计算密集型场景。

altic-dev/FluidVoice

Swift · ★ 3,613 · 🍴 232 · 📈 491 stars today

FluidVoice - Fastest macOS Offline Dictation app - Voice to Text fully Local. One ⭐ takes us a long way :))

中文介绍 `FluidVoice` 是一款专为 macOS 用户设计的高性能离线听写应用，旨在提供极速的语音转文本体验。该项目最大的亮点在于其完全本地化的处理能力，无需联网即可准确进行语音识别，有效保护了用户隐私并确保了数据安全。它解决了传统在线听写服务可能存在的延迟、隐私泄露和网络依赖问题，非常适合需要频繁进行文本输入、注重效率和隐私的 macOS 用户，无论是撰写文档、记录会议还是日常办公，都能提供流畅且安全的语音输入体验。

opendatalab/MinerU

Python · ★ 71,479 · 🍴 6,006 · 📈 426 stars today

Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.

中文介绍 MinerU 专注于将复杂的非结构化文档，如 PDF 和 Office 文档，高效转换为 LLM (大型语言模型) 易于处理的 Markdown 或 JSON 格式。该工具解决了传统文档格式难以直接被 AI 代理理解和利用的问题，通过智能解析和结构化处理，为 Agentic 工作流提供高质量的输入数据。这使得 AI 代理能够更准确地从文档中提取信息、执行分析或生成内容，极大地提升了处理文档密集型任务的效率和准确性。

HKUDS/Vibe-Trading

Python · ★ 14,229 · 🍴 2,622 · 📈 490 stars today

"Vibe-Trading: Your Personal Trading Agent"

中文介绍 Vibe-Trading 是一个个人交易 Agent 项目，旨在为用户提供智能化的交易辅助。它可能通过分析市场数据、情绪指标等，自动化交易决策和执行策略，帮助用户更好地管理投资组合并捕捉交易机会。该项目适用于希望利用 AI 技术增强交易能力、实现自动化投资或寻求个性化交易策略的个人投资者和量化交易爱好者。

ByteByteGoHq/system-design-101

★ 84,346 · 🍴 9,343 · 📈 132 stars today

Explain complex systems using visuals and simple terms. Help you prepare for system design interviews.

中文介绍 `system-design-101` 是一个专注于系统设计学习的开源项目，旨在通过直观的图表和简明的语言，帮助读者理解复杂的分布式系统架构。该项目将抽象的系统概念具象化，解决了传统教材可能存在的理论性过强、难以理解的问题。它特别适用于准备系统设计面试的软件工程师，以及任何希望深入了解大型系统如何构建、伸缩和维护的技术人员。通过提供清晰的解释和实用的案例，它能有效提升读者的系统设计思维和解决实际问题的能力。

usestrix/strix

Python · ★ 26,626 · 🍴 2,980 · 📈 88 stars today

Open-source AI hackers to find and fix your app’s vulnerabilities.

中文介绍 `strix` 是一个开源的 AI 安全工具，旨在帮助开发者和安全团队自动发现并修复应用程序中的潜在漏洞。该项目利用人工智能技术，模拟“AI 渗透测试员”的行为模式，对应用进行深度分析和攻击模拟，从而有效识别传统安全扫描工具可能遗漏的安全缺陷。它解决了人工安全审计耗时耗力、且容易出错的问题，特别适用于需要持续安全保障的开发流程（如 CI/CD），帮助团队在软件发布前提升应用的整体安全性，降低潜在的安全风险。

browser-use/video-use

Python · ★ 10,920 · 🍴 1,519 · 📈 324 stars today

Edit videos with coding agents

中文介绍 `video-use` 是一个探索通过“编码代理”来编辑视频的创新项目。它旨在颠覆传统的图形界面视频编辑模式，允许用户通过编写代码或使用智能代理来自动化和执行复杂的视频剪辑、特效添加及内容生成等任务。该项目解决了传统视频编辑软件操作繁琐、难以批量处理的问题，为开发者、内容创作者和研究人员提供了一种全新的、更具编程性和自动化潜力的视频制作方式，尤其适合需要进行程序化视频生成、自动化后期处理或实验新颖交互模式的场景。

27 Hidden Claude Features, Settings & Shortcuts That Most Users Don't Know

@sairahul1 · 121.4K 粉丝 · 2.9M 阅 · 506 赞 · 97 转

06/26 16:47

Most people use Claude Code like a fancy autocomplete. They prompt. They wait. They accept the first output. They are leaving 90% of the power on the table. There are 27 specific moves that separate

中文介绍 此推文列举 27 个大多数用户可能不了解的 Claude 隐藏功能、设置与快捷方式。旨在帮助用户发现并充分利用 Claude 更深层次的能力，以提高效率和操作体验，解锁其未被发现的潜力。

How To Build a One-Person Company Using Claude Cowork

@sairahul1 · 121.4K 粉丝 · 876.6K 阅 · 547 赞 · 83 转

06/27 17:03

Emails. Formatting. Compiling reports. Preparing decks. Organizing files. Researching. Marketing. Writing. SEO. The average knowledge worker spends 60% of their day on above work that doesn't require

中文介绍 帖子讨论如何利用 Claude Cowork 搭建一人公司。博主可能分享了相关策略、工具集成或具体实践方法，旨在为独角创业者提供运营思路和效率提升方案。

The efficiency gap: How do cells and GPUs compare when running the exact same algorithm?

@spandan_madan · 1.1K 粉丝 · 626.9K 阅 · 516 赞 · 28 转

06/24 02:13

AI hardware is having a moment. Hyperscaler capex on AI data centres is on track to clear $690 billion in 2026, and private equity has followed in scale — Blackstone alone reports a $55B+ data-centre

中文介绍 该帖子探讨了 AI 硬件的当前发展趋势与投资规模，并引出对生物细胞与 GPU 在执行相同算法时效率差异的比较。文章深究了两种截然不同的计算范式，旨在分析现有 AI 硬件的局限性，并为未来 AI 算力架构的演进提供新的思考角度。

Introducing Engram: Scaling compute on your context

@EngramLab · 1.2K 粉丝 · 255.7K 阅 · 537 赞 · 76 转

06/24 01:01

We’re Engram. We’re building AI that learns from you and deeply understands your work. Today’s AI models don’t understand what you do. Not really. Everything models know comes from their training –

中文介绍 Engram Lab 正式推出其新产品 Engram，这是一款旨在「深入理解用户工作」并能从用户互动中学习的 AI。该产品致力于解决现有 AI 模型仅依赖训练数据、无法真正理解个体工作背景的局限，通过扩展上下文计算能力，提供更个性化和高效的 AI 协助。

30 Core Agentic Engineering Concepts Every Developer Should Know

@sairahul1 · 121.4K 粉丝 · 249.0K 阅 · 502 赞 · 90 转

06/23 17:28

There are 8 billion people on the planet. Only a fraction of developers understand how AI agents actually work. Not the demos. Not the hype. The real engineering underneath. Every week a new agent

中文介绍 针对当前多数开发者对 AI 代理缺乏深入理解的现状，该帖子列举了「30 个核心代理工程概念」。内容旨在揭示 AI 代理背后的真实工程原理，而非仅停留在演示或炒作层面，帮助开发者掌握构建和理解 AI 代理所需的基础知识，提升其代理工程能力。

$OUST Deep Dive - One of My Favorite Physical AI Plays

@crux_capital_ · 52.8K 粉丝 · 217.5K 阅 · 503 赞 · 56 转

06/28 04:53

This is the most in depth report I have ever written, on a very exciting company. If you read this article you will have a much deeper understanding of what Ouster does, how it plays into Physical AI,

中文介绍 博主发布一篇关于 Ouster 公司的深度分析报告，详尽阐述该公司在“实体 AI”（Physical AI）领域的核心角色、技术及市场潜力，旨在帮助读者深入理解其业务模式。

ORACLE: Official AI Agents Trade on Polymarket

@Oracle_Trade_ai · 39.9K 粉丝 · 197.8K 阅 · 2.8K 赞 · 580 转

06/22 16:52

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 该帖子介绍 ORACLE，一个在 Polymarket 预测市场进行交易的「官方 AI 代理」。帖子展望到 2026 年，预测市场中超过 30% 的活动将由算法和 AI 驱动的钱包产生，强调了 AI 代理在自动化交易和预测市场中的日益增长影响力与有效性。

Why we're bullish on loops

@posthog · 21.8K 粉丝 · 162.4K 阅 · 512 赞 · 36 转

06/24 01:26

When the creators of both OpenClaw and Claude Code speak, people listen. And recently Peter Steinberger and Boris Cherny have both been talking about the same concept: loops. Their argument? You

中文介绍 该帖子探讨了「循环」概念的重要性，并援引 OpenClaw 和 Claude Code 创始人 Peter Steinberger 与 Boris Cherny 的观点，解释了为何应看好循环工程。文章旨在强调循环在 AI 交互和自动化中的核心作用，主张通过构建循环而非简单的提示词来提升 AI 系统的效率和自主性。

ORACLE: Official AI Agents Trade on Polymarket

@OracAItrading · 31.8K 粉丝 · 141.6K 阅 · 2.8K 赞 · 576 转

06/22 16:52

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 该帖子介绍 ORACLE，一个在 Polymarket 预测市场进行交易的「官方 AI 代理」。帖子展望到 2026 年，预测市场中超过 30% 的活动将由算法和 AI 驱动的钱包产生，强调了 AI 代理在自动化交易和预测市场中的日益增长影响力与有效性。

Interactions API: Our primary interface for Gemini models and agents

@GoogleAIStudio · 179.4K 粉丝 · 138.2K 阅 · 504 赞 · 42 转

06/23 01:21

Today we're announcing that the Interactions API has reached general availability and is now our primary API for interacting with Gemini models and agents. We launched its public beta in December

中文介绍 Google AI Studio 宣布 Interactions API 已正式普遍可用，并成为其与 Gemini 模型和代理交互的「主要接口」。该 API 于去年 12 月推出公测版，此举标志着 Google 为开发者提供了更稳定、更强大的工具，以便更好地利用 Gemini 系列 AI 产品。

ORACLE: Official AI Agents Trade on Polymarket

@Oractrading · 33.9K 粉丝 · 109.2K 阅 · 2.8K 赞 · 585 转

06/22 16:20

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 该帖子介绍 ORACLE，一个在 Polymarket 预测市场进行交易的「官方 AI 代理」。帖子展望到 2026 年，预测市场中超过 30% 的活动将由算法和 AI 驱动的钱包产生，强调了 AI 代理在自动化交易和预测市场中的日益增长影响力与有效性。

How To Use Loop Engineering To Build A Self-Improving Quant Trading System

@RohOnChain · 51.4K 粉丝 · 108.3K 阅 · 501 赞 · 65 转

06/22 21:54

I will break down exactly how to build the loops that run an entire quant trading system on their own. Let's get straight to it. Bookmark This - I'm Roan, a backend developer working on system

中文介绍 博主详细分解了如何运用「循环工程」技术，构建一个能够自主运行并持续改进的量化交易系统。该教程面向后端开发者，旨在教授如何设计和实现驱动整个量化交易流程的自动化循环，从而提升系统的自适应性和交易效率。

ORACLE: Official AI Agents Trade on Polymarket

@Oracle__Market · 5.9K 粉丝 · 100.0K 阅 · 2.1K 赞 · 823 转

06/28 02:14

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 推文展望 2026 年，自治 AI 代理将成为预测市场 Polymarket 上的高效策略。届时，超过 30% 的平台活动预计由算法和 AI 驱动的钱包贡献，展示了 AI 在预测和交易领域的巨大潜力与影响力。

ORACLE: Official AI Agents Trade on Polymarket

@Oracle_Market__ · 25.1K 粉丝 · 99.3K 阅 · 2.8K 赞 · 224 转

06/28 01:42

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 帖子展望 2026 年，预测自治 AI 代理将在 Polymarket 等预测市场中占据主导地位。届时，超过 30% 的市场活动将由算法和 AI 驱动的钱包产生，揭示 AI 在金融预测领域的未来潜力。

This Week on Base: New Base MCP skills + now live in Perplexity, Base App on desktop, and more!

@base · 1.3M 粉丝 · 93.3K 阅 · 501 赞 · 127 转

06/27 04:40

News ■ Base MCP's new skills let agents transact, trade, lend, mint, and buy onchain across 13 apps: @yield @AskVenice @KyberNetwork @opensea @o1_exchange @Balancer @printr @bitrefill @flaunchgg

中文介绍 Base 宣布其 MCP 推出新技能，赋能 AI 代理在 13 个 DApp 上实现链上交易、借贷、铸币及购买等操作。这些新功能支持代理与 @yield、@KyberNetwork、@opensea 等应用交互，显著扩展了 Base 生态中 AI 代理的自动化能力。

thoughts on why mcp didn't work, what's next

@RhysSullivan · 57.4K 粉丝 · 86.1K 阅 · 503 赞 · 25 转

06/26 09:03

mcp came out when the best models were sonnet 3.5 and GPT 4o not a lot was known about how to properly work with these tools yet, we were still incredibly concerned on models having access to tools,

中文介绍 博主反思 MCP 初期表现不佳的原因，认为当时 Sonnet 3.5 和 GPT-4o 等最佳模型仍存局限，且对模型直接访问工具的能力普遍担忧。推文暗示了对该项目未来发展方向的思考，为早期 AI 代理的挑战提供了见解。

Subnets mine TAO; TAO mines Subnets

@const_reborn · 29.7K 粉丝 · 79.8K 阅 · 503 赞 · 116 转

06/23 09:21

e_i \;\propto\; \underbrace{\rho_i \times \bar{p}_i}_{\text{linear (maximize)}} \times \underbrace{(1 - b_i)}_{\text{boolean gate}} Disclaimer: this upgrade only effects subnet owners and dynamic TAO

中文介绍 该帖子深入探讨了 subnets (子网) 与 TAO 之间的内在机制，并通过数学公式解释了它们如何相互「挖矿」和协同工作。内容涉及动态 TAO 和布尔门控等技术细节，主要面向子网所有者，揭示了底层经济模型和激励机制，强调了去中心化网络中的价值流转。

The Hermes + Obsidian + Claude Code Trinity: The Full System for Running a One Person Company

@cyrilXBT · 186.6K 粉丝 · 77.2K 阅 · 510 赞 · 93 转

06/28 08:55

There is a specific moment that tells you whether someone is running a real one person company or just using a lot of apps. It is the moment they get asked a question about their own business and they

中文介绍 博主分享一套用于高效运营一人公司的完整系统，核心由 Hermes、Obsidian 和 Claude Code 三个工具协同构成，旨在帮助独角创业者整合应用，流畅应对业务挑战。

DanceOPD: On-Policy Generative Field Distillation

👍 69

06/25 08:00

Modern image generation demands a single model that unifies diverse capabilities, including text-to-image (T2I), local editing, and global editing. However, these capabilities are rarely naturally aligned and often conflict. For instance, editing tends to degrade T2I performance, while global and lo

中文介绍 现代图像生成需统一文本到图像、局部和全局编辑能力，但这些能力常有冲突。DanceOPD 论文介绍了一种策略生成场蒸馏模型，旨在解决这些能力之间的对齐和冲突问题，提升单模型在多任务下的图像生成性能，避免编辑任务对文本到图像性能的降低。

Ask, Solve, Generate: Self-Evolving Unified Multimodal Understanding and Generation via Self-Consistency Rewards

👍 0

06/25 08:00

Most unified large multimodal models (LMMs) that support both visual understanding and image generation still rely on curated post-training supervision, such as human annotations, preference labels, or external reward models. We ask whether a unified LMM can improve both abilities autonomously using

中文介绍 大多数统一大型多模态模型（LMMs）在视觉理解和图像生成方面，仍依赖人工标注或外部奖励模型进行后训练。本研究探讨 LMMs 如何通过“提问、解决、生成”的自我一致性奖励机制实现自我进化，从而在不依赖额外监督的情况下，同步提升其理解与生成能力。

Paying More Attention to Visual Tokens in Self-Evolving Large Multimodal Models

👍 0

06/25 08:00

Recently, self-evolving large multimodal models (LMMs) have received attention for improving visual reasoning in a purely unsupervised setting. However, multi-role self-play and self-consistency reward schemes in existing self-evolving LMMs optimize answer agreement without ensuring the decoder atte

中文介绍 自进化大型多模态模型（LMMs）在无监督视觉推理方面受到关注。本研究指出，现有模型的多角色自玩和自我一致性奖励机制，主要优化答案一致性，可能忽视对视觉token的深入关注。论文提出应更重视视觉token，以确保推理的视觉准确性和深度。

PhysiFormer: Learning to Simulate Mechanics in World Space

👍 9

06/25 08:00

We present PhysiFormer, a diffusion transformer for physically-plausible 3D object motion. Unlike video world models that operate in view-dependent pixel space, PhysiFormer represents objects as 3D meshes expressed in world coordinates. Given the initial vertex positions and velocities, as well as o

中文介绍 PhysiFormer 是一种扩散Transformer，用于模拟物理上可信的3D物体运动。与传统在像素空间操作的视频世界模型不同，PhysiFormer 将物体表示为世界坐标中的3D网格。该模型能根据初始顶点位置和速度，预测未来物理运动，提高3D模拟的真实性。

Autoregressive Boltzmann Generators

👍 0

06/25 08:00

Efficient sampling of molecular systems at thermodynamic equilibrium is a hallmark challenge in statistical physics. This challenge has driven the development of Boltzmann Generators (BGs), which allow rapid generation of uncorrelated equilibrium samples by combining a generative model with exact li

中文介绍 高效采样热力学平衡态下的分子系统是统计物理学的核心挑战。本研究介绍了自回归玻尔兹曼生成器（Autoregressive Boltzmann Generators），该模型结合生成式方法，能够快速生成去相关且符合平衡分布的样本，有效解决分子模拟中的采样效率问题。

SAM2Matting: Generalized Image and Video Matting

👍 0

06/25 08:00

Despite impressive advances in image matting, video matting remains challenging due to the inherent gap between high-level tracking, which requires frame-wise understanding, and low-level matting, which focuses on extremely fine-grained details. Existing methods attempt this with expensive and narro

中文介绍 尽管图像抠图技术进步显著，但视频抠图仍因高层跟踪和细粒度抠图之间的鸿沟而充满挑战。SAM2Matting 论文提出一种通用图像和视频抠图方法，旨在弥合不同任务间的差距，通过统一框架实现从图像到视频的精细抠图，提升泛化能力。

Hallucination in World Models is Predictable and Preventable

👍 8

06/25 08:00

Modern generative world models render increasingly realistic action-controllable futures, yet they frequently hallucinate: rollouts remain visually fluent while drifting from the ground-truth dynamics. We hypothesize that hallucination concentrates in low-coverage regions of the state-action space,

中文介绍 现代生成式世界模型在生成逼真的可控未来时，常出现幻觉，即视觉上流畅但偏离真实动力学。本研究发现，这种幻觉现象集中在状态空间的低覆盖区域，并提出幻觉是可预测和可预防的。这项研究为提升世界模型预测的准确性提供了新的见解。

OctoSense: Self-Supervised Learning for Multimodal Robot Perception

👍 0

06/25 08:00

We present OctoSense, an open-source sensor platform with stereo RGB and event cameras, LiDAR, a thermal camera, an inertial measurement unit, RTK-corrected global positioning system, and proprioception (CAN bus data from a car, and joint angles for a quadruped robot). The eponymous OctoSense datase

中文介绍 OctoSense 论文发布了一个开源的多模态机器人感知传感器平台。它集成了双目RGB和事件相机、激光雷达、热像仪、IMU、RTK-GPS以及本体感知数据。该平台旨在通过自监督学习，全面提升机器人在复杂环境下的感知能力和鲁棒性。

ViQ: Text-Aligned Visual Quantized Representations at Any Resolution

👍 38

06/25 08:00

A unified representation for text and vision is a natural pursuit, as it enables simpler multimodal modeling and more efficient training. However, representing images as discrete signals in the same way as text inevitably introduces severe information loss. Existing work struggles to balance low-lev

中文介绍 为实现更简洁高效的多模态建模，文本与视觉的统一表示是关键目标。然而，将图像离散化如文本般会带来严重信息损失。ViQ 论文介绍了一种文本对齐的视觉量化表示方法，能在任意分辨率下处理图像，旨在克服现有方法的不足，解决信息损失挑战。

LA4VLA: Learning to Act without Seeing via Language-Action Pretraining

👍 0

06/25 08:00

Vision-Language-Action (VLA) models are commonly pretrained on robot demonstrations by jointly mapping visual observations and language instructions to actions. However, dense visual-action supervision can dominate the comparatively sparse language-action signal. As a result, policies may rely on vi

中文介绍 LA4VLA是一种新的机器人模型，其核心在于通过语言-动作预训练，使机器人能够在没有视觉输入的情况下执行操作。它旨在解决现有VLA（视觉-语言-动作）模型中，密集视觉-动作监督信号常主导稀疏语言-动作信号的问题。LA4VLA的提出，有望提升机器人对语言指令的理解和执行能力，实现更灵活的具身智能。

When Does Combining Language Models Help? A Co-Failure Ceiling on Routing, Voting, and Mixture-of-Agents Across 67 Frontier Models

👍 3

06/25 08:00

Multi-model LLM systems such as routing, voting, cascades, fusion, and mixture-of-agents are used to beat single-model accuracy. We show that their gain is capped by a quantity the field rarely reports. For any policy whose output is one member model answer, accuracy cannot exceed one minus beta, wh

中文介绍 多模型大语言模型（LLM）系统，如路由、投票和混合专家，常用于超越单模型准确率。本研究分析了67个前沿模型，发现多模型系统的增益受「共同失败上限」限制，即其准确率无法超过所有成员模型共性错误率所设定的上限，揭示了性能提升的边界。

EO-WM: A Physically Informed World Model for Probabilistic Earth Observation Forecasting

👍 2

06/25 08:00

Earth Observation (EO) forecasting aims to predict future Earth surface dynamics from satellite observations under changing meteorological conditions. In this paper, we view this task as a partially observed, weather-driven world modeling problem, in which weather acts as a conditioning signal, whil

中文介绍 地球观测（EO）预测旨在通过卫星观测和气象条件预测未来地球表面动态。EO-WM 论文提出了一个物理信息增强的世界模型，用于概率性地球观测预测。该模型将EO预测视为一个受天气驱动的部分可观测世界建模问题，提升了预测的准确性和鲁棒性。

From Celebrities to Anyone: Characterizing AI Nudification Content, Technology, and Community Dynamics on 4chan

👍 0

06/25 08:00

AI nudification uses generative models to create synthetic non-consensual sexually explicit imagery (SNEACI) of real individuals. Prior work has examined dedicated nudification platforms and model repositories, finding that most targets are female celebrities. However, the anonymous content communit

Ask, Don't Judge: Binary Questions for Interpretable LLM Evaluation and Self-Improvement

👍 0

06/25 08:00

Evaluating LLM outputs remains a major bottleneck in NLP: human evaluation is expensive and slow, lexical metrics correlate poorly with human judgments on open-ended generation, and holistic LLM judges often produce opaque scores that are hard to debug. We propose BINEVAL, a framework that decompose

LISA: Likelihood Score Alignment for Visual-condition Controllable Generation

👍 13

06/25 08:00

The prevalent dual-branch paradigm, i.e., training a side network to encode visual conditions and fusing its intermediate-layer features to a frozen pretrained main network, has shown remarkable success in visual-condition controllable generation. Despite its widespread adoption, the role of the sid

Learning to Fold: prizewinning solution at LeHome Challenge 2026 (1st place online, 2nd offline)

👍 0

06/25 08:00

I describe my solution to the LeHome Challenge 2026, an ICRA 2026 competition on bimanual garment folding. The system placed 1st of 62 teams in the online (simulation) round and 2nd in the real-world final. It improves a vision-language-action (VLA) policy with a reinforcement-learning loop. The pol

Where Do Models Find Happiness? Emotion Vectors in Open-Source LLMs

👍 0

06/25 08:00

Recent work identified emotion vectors in Claude Sonnet 4.5, which are internal representations that encode emotion concepts, causally influence behavior, and exhibit geometry mirroring human psychological structure. We test the generality of these findings in two open-weight models, Apertus-8B-Inst

How Much Static Structure Do Code Agents Need? A Study of Deterministic Anchoring

👍 0

06/25 08:00

LLM-based code agents navigate repositories through keyword search but miss the structural relationships, such as call graphs, inheritance hierarchies, and configuration dependencies, that define how software actually works. This makes agent navigation stochastic and difficult to reproduce across ru

To Run or Not to Run: Analyzing the Cost-Effectiveness of Code Execution in LLM-Based Program Repair

👍 1

06/25 08:00

LLM-based agents for program repair are increasingly built on a "generate-run-revise" paradigm, iteratively executing tests to evaluate and refine patches. This execution-based approach has become standard practice in state-of-the-art systems. However, executions can be time-consuming and expensive,

PhysRAG: Enhancing Physics-Awareness in Video Generation via Retrieval-Augmented Generation

👍 0

06/25 08:00

Developing physically aware video generation models remains a significant challenge due to the difficulty in capturing diverse physical phenomena, such as thermal dynamics, mechanics, and optics. In this work, we introduce PhysRAG, a novel pipeline that enhances physical awareness in video generatio

Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation

👍 41

06/25 08:00

While text-to-image (T2I) models have achieved remarkable progress, they struggle with real-world requests that are often underspecified, implicit, or dependent on up-to-date knowledge. We identify this challenge as the Context Gap: the mismatch between the user context and the sufficient generation

Confidence-Aware Tool Orchestration for Robust Video Understanding

👍 9

06/25 08:00

Video reasoning language models implicitly assume that every input frame is equally reliable. This leads to what we term the Blind Trust Problem: under realistic perturbations such as motion blur, glare, or occlusion, frontier video reasoning models can suffer 15-30%p accuracy drops on real-world em

A Pipeline for Generating Longitudinal Synthetic Clinical Notes Using Large Language Models

👍 0

06/25 08:00

Synthetic data is increasingly used to enable the development and evaluation of AI systems in domains where access to real-world data is restricted. In healthcare, clinical documentation presents particular challenges due to its sensitivity. This work introduces a synthetic clinical notes pipeline a

Information-Aware KV Cache Compression for Long Reasoning

👍 9

06/25 08:00

Reasoning capability has advanced rapidly in large language models (LLMs), leading to an increasing size of key-value (KV) cache in both prefilling and decoding stages. Existing KV cache compression methods mainly rely on attention weights to estimate token importance. While attention effectively ca

AgentX: Towards Agent-Driven Self-Iteration of Industrial Recommender Systems

👍 0

06/25 08:00

Recommendation algorithm iteration is moving from an artisanal, engineer-bound process toward an industrialized research loop, but this transition remains blocked by a structural execution bottleneck: the idea-to-launch cycle still depends on human engineers to generate hypotheses, modify production

KARLA: Knowledge-base Augmented Retrieval for Language Models

👍 0

06/25 08:00

We propose a new method that allows an LLM to automatically pull in factual knowledge from a knowledge base during token generation. This means that (1)~factual knowledge in the LLM output can be updated without retraining the LLM, (2)~facts in the LLM output can be traced to the knowledge base for

ReasonCLIP-58M: Visually Grounded Commonsense Reasoning Supervision for CLIP

👍 0

06/25 08:00

CLIP and its variants are widely adopted visual backbones in multimodal systems, but their pretraining remains dominated by descriptive image-text alignment. As downstream applications increasingly demand visually grounded commonsense inference and compositional reasoning, it remains unclear whether

OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning

👍 45

06/25 08:00

Outcome-based reinforcement learning provides a stable optimization backbone for language agents, but its sparse trajectory-level rewards provide little guidance on which intermediate decisions should be reinforced or suppressed. On-policy self-distillation offers dense token-level supervision, yet

Anatomy-Guided Residual Motion Diffusion for Controllable 4D Cardiac MRI Synthesis

👍 0

06/25 08:00

Developing robust artificial intelligence models for 4D (3D + time) medical imaging is constrained by limited annotated data, inter-device domain shifts, and privacy restrictions. To address this, we propose a 4D controllable generative framework for anatomically consistent data augmentation. A semi

DiCoBench: Benchmarking Multi-Image Fine-Grained Perception via Differential and Commonality Visual Cues

👍 1

06/25 08:00

Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated impressive fine-grained perception capabilities. However, existing benchmarks predominantly rely on explicit textual cues or low-resolution inputs, failing to evaluate a model's ability to autonomously perceive implici

GetCompress

06/29 04:50

Lossless media compression without context switching

中文介绍 GetCompress是一款媒体压缩工具，致力于在不切换上下文的情况下实现无损媒体压缩。它旨在简化工作流程，让用户无需离开当前应用环境即可处理媒体文件，从而提升效率和用户体验，同时保持文件质量不受影响。

discode.ai

06/29 04:47

100+ AI models, one interface. ECO friendly.

中文介绍 discode.ai 提供一个统一的用户界面，整合了超过100个不同的AI模型。这款平台强调其环保特性，旨在为用户提供一个高效且可持续的AI模型管理和使用环境。通过单一入口，用户可以便捷地访问多种AI功能。

Dotient

06/29 04:46

Your local semantic search app

中文介绍 Dotient 是一款本地语义搜索应用程序。它允许用户在本地设备上执行语义搜索，这意味着搜索结果能更准确地理解查询意图和内容含义。这款应用旨在提供私密且高效的本地信息检索能力。

Basedash for Excel

06/29 04:46

Turn any Excel file into a live dashboard

中文介绍 Basedash for Excel是一款工具，能将任何Excel文件转换为实时数据看板。它旨在帮助用户快速、便捷地实现数据可视化，使得复杂的Excel数据能够动态呈现，方便团队共享和监控关键指标，从而提升数据分析与决策效率。用户无需复杂操作即可将静态数据转化为交互式仪表盘。

RetroMac

06/29 04:46

Turn your Mac into a time machine.

中文介绍 RetroMac 是一款软件，能将用户的 Mac 电脑界面「变身」为复古风格，模拟旧版操作系统的视觉效果。它旨在为用户提供一种怀旧的体验，通过改变界面主题和元素，让现代 Mac 仿佛回到过去。

Persona.js

06/29 04:44

Add WebMCP-native AI chat to any Frontend

中文介绍 Persona.js 是一个软件开发工具库，允许开发者将WebMCP原生的AI聊天功能轻松集成到任何前端界面中。它旨在简化AI聊天功能的部署过程，使开发者能够快速为他们的应用程序或网站添加智能对话能力，从而提升用户交互体验。

Nada

06/29 04:33

Compose music with just your voice

中文介绍 Nada 是一款创新的音乐创作工具，它允许用户仅通过自己的声音来作曲。这款应用能够将人声输入转化为音乐作品，极大地简化了音乐创作的门槛，让非专业人士也能轻松实现音乐构思，并生成旋律。

QApilot's CoWork

06/29 04:23

3x Mobile Automation. Same QE Team.

中文介绍 QApilot's CoWork 专注于通过自动化技术，帮助现有的QE（质量工程）团队在不增加人员的情况下，将移动自动化效率提升3倍。这款工具旨在优化测试流程，显著提高移动应用的测试速度和质量。

Gemini Spark

06/29 04:23

Your 24/7 personal AI agent

中文介绍 Gemini Spark是一款提供全天候服务的个人AI代理。它被设计为一个随时待命的智能助手，旨在为用户提供个性化的支持和自动化服务。该产品致力于在不同场景下帮助用户管理任务、获取信息或执行操作，从而提高个人效率和便捷性，成为用户日常生活和工作的得力助手。

Lyto

06/29 04:19

"One AI agent across your browser, tools, and messages "

中文介绍 Lyto 提供一个跨浏览器、跨工具和跨消息应用的统一AI代理。它旨在作为一个智能助手，无缝整合用户在不同平台上的操作与信息，提供连贯且高效的AI支持体验，从而简化工作流程和日常数字交互。

Poweful Agents inside Slack

06/25 06:39

中文介绍 这段视频短片介绍了在Slack平台中部署功能强大的AI智能体。这些智能体能够协助用户完成各项任务，提高工作效率，实现自动化流程。

AI Agents Just Changed Forever: GLM 5.2, Codex Skills, Claude & Cursor

06/21 21:30

中文介绍 视频探讨了AI智能体领域的重大变革，这些变革将永久性地改变其发展方向。内容提到了GLM 5.2、Codex Skills、Claude以及Cursor等技术或产品在推动AI智能体能力提升方面的作用和影响。

9 AI Agent Skills To Get Ahead of 99% of People

06/18 10:27

How to build high quality internal tools with Retool

06/18 05:21

SpaceX Just Bought Cursor for $60B. It’s About to Take OVER.

06/17 05:31

Why the Government Just Killed Claude Fable 5

06/13 11:50

This is Why They Banned Claude Fable 5

06/13 02:28

Hermes Agent NEW Super-App and DeepSeek v4 Catches Up To Opus 4.8?

06/07 04:10

ChatGPT And Codex Are Merging (This Changes Everything)

06/04 06:17

Claude Opus 4.8: Everything You Need to Know

06/01 06:27

发现了一个非常好用的语音输入法，叫 SayIt。

06/28 17:25

8 回复 · 程序员 节点

AI 变化太快，概念层出不穷？ prompt engineering 到 context engineering 到 harness engineering 到 loop engineering 到 ... ...? 会到哪里？

06/28 16:11

17 回复 · 程序员 节点

大家用 iPad 、mac 的时候开 icloud 么

06/28 16:03

13 回复 · Apple 节点

iPhone 18 基础款感觉不太会涨价

06/28 13:31

11 回复 · Apple 节点

关于 Loop Engineering 的实践与反思

06/28 10:00

17 回复 · 程序员 节点

做了个 Launchpad 替代品（数不清第几个了），打磨了 9 个月，想听听大家的反馈

06/28 09:09

12 回复 · Apple 节点

怎么办，我已经习惯了每天几亿 token 了。。。。

06/27 23:54

9 回复 · 程序员 节点

Apple WLOC 定位修改，解决地理围栏问题

06/27 11:40

39 回复 · Apple 节点

换了 17pm，问一下大家换新机是直接无脑数据同步过去，还是重新下载 app 只同步照片微信什么的

06/27 11:06

14 回复 · Apple 节点

更新到 ios 27 和 macos 金拱门 的感受如何，我先来

06/27 09:46

16 回复 · Apple 节点

Show HN: Bash4LLM+ – A lightweight, dependency-free Bash wrapper for LLM APIs

06/29 03:43

Bash4LLM is a single-file Bash wrapper for interacting with LLMs from the terminal. I created it because I wanted something simple that worked without installing Python, Node, or any other runtime.It uses only Bash, curl, and jq. You can send prompts, start a small chat, process files line by line,

Show HN: NanoEuler – GPT-2 scale model in pure C/CUDA from scratch

06/29 03:38

Hi everyone,I started working on nanoeuler after the ban of anthropic's fable because my ambition and dream is to work in the AI field in anthropic. The two interesting reasons that led me to create nanoeuler were (1) interfacing with llm does not mean understanding how they are composed and (2

TOP500 at ISC'26: We Have a New Number 1 – By George Cozma

06/29 03:38

24 points · 10 comments

Ford rehires 'gray beard' engineers after AI falls short

06/29 03:35

86 points · 32 comments

The US Used to Demand the Best Tech. Now We Ban It

06/29 03:07

46 points · 24 comments

Computer-Aided Language Development in Nonspeaking Children (1968) [pdf]

06/29 02:59

http://i.stanford.edu/pub/cstr/reports/cs/tr/68/85/CS-TR-68-...

Librepods: AirPods liberated from Apple's ecosystem

06/29 02:48

129 points · 23 comments

Historical memory prices 1960-2026

06/29 02:32

40 points · 10 comments

The cost YAGNI was never about

06/29 02:31

72 points · 44 comments

Semgrep: GLM 5.2 beats Claude in our Cyber Benchmarks

06/29 01:50

57 points · 20 comments

Working around dragons with the Lemote Yeeloong laptop and OpenBSD

06/29 00:58

65 points · 12 comments

Show HN: DRM-Free Books

06/29 00:58

After several years of mandatory DRM lockdowns from most commercial book sources, now authors have a choice when it comes to DRM for their books. Pick authors and books that are DRM-free, or download DRM-free classics that are out of copyright.https://frequal.com/Perspectives/Drm

I used Claude Code to get a second opinion on my MRI

06/29 00:35

206 points · 309 comments

Daisugi, the Japanese technique of growing trees out of other trees (2020)

06/29 00:28

https://x.com/wrathofgnon/status/1250287741247426565, https://xcancel.com/wrathofgnon/status/1250287741247426565

Tokenmaxxing is dead, long live tokenmaxxing

06/29 00:24

73 points · 90 comments

Examining circuit boards from the Space Shuttle's I/O Processor

06/29 00:16

61 points · 12 comments

Programmable Probabilistic Computer with 1M p-bits

06/28 23:55

34 points · 0 comments

Show HN: Zanagrams

06/28 23:26

105 points · 36 comments

Michigan bill would bar employers from requiring after-hours coms with workers

06/28 22:46

202 points · 140 comments

5k menus from the New York Public Library’s Buttolph Collection (1880-1920)

06/28 22:44

https://pudding.cool/2026/06/menu-collection/

EU to legislate about Chat Control behind closed doors

06/28 22:40

494 points · 280 comments

The curious case of the disappearing Polish S (2015)

06/28 20:44

188 points · 57 comments

The MUMPS 76 Primer – anniversary edition

06/28 20:41

61 points · 26 comments

A way to exclude sensitive files issue still open for OpenAI Codex

06/28 20:27

161 points · 106 comments

The KIDS Act would require age checks to get online

06/28 19:56

192 points · 193 comments

More evidence is consistent with possible ancient life on Mars (2025)

06/28 19:55

41 points · 55 comments

Marfa Public Radio Puts You to Sleep

06/28 10:23

373 points · 115 comments

Show HN: Appaca – AI Workspace for Operators

06/26 12:54

Appaca is my third pivot.A couple of years ago, I started working on an idea on no-code platform that generates code. The goal is to help devs and agencies ship products faster for their clients. I went through Antler startup accelerator and got initial funding. I was working on the right problem, b

The Boeing 747 begins its final descent

06/25 23:54

82 points · 81 comments

Trade, merchants, and the lost cities of the Bronze Age (2019) [pdf]

06/24 01:40

(changed from https://thedailyeconomy.org/article/ancient-clay-tablets-sho...)

v2.1.195

06/27 05:29

What's changed Added CLAUDE_CODE_DISABLE_MOUSE_CLICKS to disable mouse click/drag/hover in fullscreen mode while keeping wheel scroll Fixed hook matchers with hyphenated identifiers (e.g. code-reviewer, mcp__brave-search) accidentally substring-matching — they now exact-match. Use mcp__brave-search_

中文介绍 Anthropic的Claude Code项目发布了v2.1.195版本。主要更新包括：新增CLAUDE_CODE_DISABLE_MOUSE_CLICKS环境变量，允许在全屏模式下禁用鼠标点击、拖拽、悬停，但保留滚轮滚动；并修复了带连字符的钩子匹配器（例如「code-reviewer」），使其实现精确匹配。

v2.1.193

06/26 05:45

What's changed Added autoMode.classifyAllShell setting to route all Bash/PowerShell commands through the auto-mode classifier instead of only arbitrary-code-execution patterns Added auto-mode denial reasons to the transcript, the denial toast, and /permissions recent denials Added claude_code.assist

中文介绍 Anthropic Claude Code 发布 v2.1.193 版本，主要增强了其自动模式。更新包括新增 “autoMode.classifyAllShell” 设置，所有 Bash/PowerShell 命令将通过自动模式分类器处理。此外，系统现在会在拒绝提示中显示自动模式的拒绝原因。

v2.1.191

06/25 05:58

What's changed Added /rewind support for resuming a conversation from before /clear was run Fixed scroll position jumping to the bottom while reading earlier output during a streaming response Fixed background agents resurrecting after being stopped — stopping an agent from the tasks panel is now pe

v2.1.190

06/24 23:53

What's changed Bug fixes and reliability improvements

v2.1.187

06/24 05:03

What's changed Added sandbox.credentials setting to block sandboxed commands from reading credential files and secret environment variables Added org-configured model restrictions to the model picker, --model, /model, and ANTHROPIC_MODEL, with a "restricted by your organization's settings" message w

v2.1.186

06/23 04:37

What's changed Added claude mcp login and claude mcp logout to authenticate MCP servers from the CLI without opening the interactive /mcp menu, with --no-browser stdin redirect support for completing over SSH Added status filtering (press f) to the /workflows agent detail view Added a "Skills" secti

v2.1.185

06/21 04:59

What's changed The stream-stall hint now reads "Waiting for API response · will retry in …" instead of "No response from API · Retrying in …", and triggers after 20s of silence instead of 10s

v2.1.183

06/19 09:20

What's changed Improved auto mode safety: destructive git commands (git reset --hard, git checkout -- ., git clean -fd, git stash drop) are now blocked when you didn't ask to discard local work, git commit --amend is blocked when the commit wasn't made by the agent this session, and terraform destro

v2.1.181

06/18 06:07

What's changed Added /config key=value syntax to set any setting from the prompt (e.g. /config thinking=false) — works in interactive, -p, and Remote Control Added sandbox.allowAppleEvents opt-in setting that lets sandboxed commands send Apple Events on macOS Added CLAUDE_CLIENT_PRESENCE_FILE enviro

v2.1.179

06/17 04:22

What's changed Fixed mid-stream connection drops: partial responses are now preserved instead of showing a raw error, and the spinner no longer gets stuck at "running tool" Fixed mouse-wheel scrolling in WSL2 under Windows Terminal and VS Code (regression in 2.1.172) Fixed a sandbox denyRead/allowRe