2026-07-16

Harness Handbook: Making Evolving Agent Harnesses Readable,Navigable, and Editable

👍 163

07/13 20:00

The capability of a modern AI agent depends not only on its foundation model but also on its harness, which constructs prompts, manages state, invokes tools, and coordinates execution. As models, APIs, environments, and requirements evolve, the harness must be continually modified. Before such a cha

中文介绍 现代 AI 代理的能力不仅依赖于基础模型，还依赖于其马具（harness），该马具用于构建提示、管理状态、调用工具及协调执行。考虑到模型、API、环境和需求的演变，必须不断修改马具。本文提出了一种手册，以提高演化代理马具的可读性、可导航性和可编辑性，提升系统的适应性和效率。意义：为未来的 Agent 和系统自适应性提供支持。

Boogu-Image-0.1: Boosting Open-Source Unified Multimodal Understanding and Generation

👍 107

07/13 20:00

We introduce Boogu-Image-0.1, an open-source unified multimodal understanding and generation model family, comprising Base, Turbo, Edit, and Edit-Turbo variants. It delivers competitive performance in high-quality text-to-image generation, fast inference, instruction-based editing, and bilingual (Ch

中文介绍 Boogu-Image-0.1 是一个开源的统一多模态理解与生成模型家族，包括 Base、Turbo、Edit 和 Edit-Turbo 等变体。它在高质量的文本到图像生成、快速推理、基于指令的编辑及双语处理上表现出色，展示出竞争力的性能。这为多模态应用和生成任务提供了新的可能性。意义：推动多模态理解和生成领域的发展。

Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models

👍 90

07/13 20:00

Coding agents must integrate external tool returns into ongoing reasoning - a capability that standard left-to-right pretraining on code exposes only in its forward direction. We observe that the action-observation-continuation loop of a coding agent is structurally isomorphic to a function call sit

中文介绍 本文提出了一种基于中间填充的函数感知的训练方式，针对编程代理的基础模型进行改进。该方法能够有效整合外部工具返回的数据，并增强代理在代码推理中的能力。研究表明，这种结构与函数调用场景具有相似性，提高了模型的推理能力，推动编码代理的性能。意义：有助于提升 Agent 的推理和工具集成能力。

Ring-Zero: Scaling Zero RL to a Trillion Parameters for Emergent Reasoning

👍 79

07/13 20:00

Reinforcement learning with verifiable rewards without human-annotated data, often referred to as zero RL, has emerged as a powerful paradigm for eliciting chain-of-thought reasoning. However, due to computational constraints, existing studies are largely restricted to small models, leaving the trai

中文介绍 Ring-Zero 实现了将零强化学习扩展至万亿参数的能力，从而提升了链式推理的生成效果。该方法在无需人工标注数据的情况下，通过可验证的奖励机制，突破了现有小模型的局限性。实验结果表明，其在大规模模型训练中的有效性为未来的推理能力提升提供了新途径。意义：对 Agent 推理及自我学习的进展有重要影响。

Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation

👍 72

07/08 20:00

Visual generators excel at rendering, but they confidently fabricate what they do not know. User requests are unbounded, evolving, and deeply long-tailed: new characters, trending entities, post-cutoff events, and more. This world-knowledge bottleneck is structural: generators are trained on fixed c

中文介绍 视觉生成模型在渲染方面表现出色，但它们在面对未知信息时往往生成错误内容。本文探讨了用户请求的复杂性以及生成模型的知识瓶颈，提出了一种方法来突破这一限制，以增强生成内容的准确性和多样性。这为提升视觉生成系统的适应性提供了新思路。意义：推动视觉生成领域的知识边界扩展。

Read It Back: Pretrained MLLMs Are Zero-Shot Reward Models for Text-to-Image Generation

👍 70

07/12 20:00

In this paper, we propose SpectraReward, a training-free reward function that turns pretrained MLLMs into off-the-shelf reward models for image-generation reinforcement learning. Instead of asking the MLLM to judge a generated image or answer decomposed verification questions, SpectraReward measures

中文介绍 SpectraReward 提出了一个无训练的奖励函数，将预训练的大型语言模型（MLLMs）转化为图像生成强化学习的奖励模型。通过测量生成图像的特征，而不是直接判断图像质量，SpectraReward 改善了奖励信号的有效性，推动了图像生成的性能提升。意义：在图像生成和强化学习交叉领域具有重要应用潜力。

SynthDocBench: Controlled Benchmark for Long-Context Visual Document Understanding

👍 63

07/10 20:00

Vision language models (VLMs) have achieved strong performance on visual document understanding benchmarks such as DocVQA, ChartQA, and MMLongBench-Doc. However, real-world documents combine multiple factors such as length, layout complexity, modality, and question difficulty, which makes it difficu

中文介绍 SynthDocBench 是一个针对长上下文视觉文档理解的受控基准，旨在解决现有视觉语言模型在处理长度、布局复杂性和问题困难度等方面的挑战。通过多个维度评估文档理解能力，该基准推动了对真实文档的全面理解性能。意义：为文档理解和视觉处理技术的发展奠定了基础。

KnowAct-GUIClaw: Know Deeply, Act Perfectly, Personal GUI Assistant with Self-Evolving Memory and Skill

👍 44

07/14 20:00

OpenClaw has emerged as a leading agent framework for complex task automation, yet it faces insufficient cross-platform GUI interaction support and a well-built self-evolution mechanism. These flaws limit its adaptation to diverse device ecosystems and prevent performance improvements through contin

中文介绍 KnowAct-GUIClaw 是一个具有自我进化记忆和技能的个人化 GUI 助手，旨在克服 OpenClaw 框架在跨平台 GUI 交互支持不足和自我进化机制缺陷的问题。该模型通过持续学习来增强用户体验和任务实现能力，提升了在不同设备上的适应性。意义：对智能助手和交互系统的发展具有重要贡献。

PolicyShiftGuard: Benchmarking and Improving Policy-Adaptive Image Guardrails

👍 30

07/07 03:04

Image guardrails are typically trained and evaluated under a fixed safety policy, implicitly treating safety as an intrinsic property of an image. Real deployments are different: the same image may be allowed in one product, restricted in another, and newly disallowed when a policy boundary changes.

中文介绍 PolicyShiftGuard 提出了一个新的基准，可以评估和改进图像的安全策略适应性。该方法通过动态评估图像安全性，解决了当前静态策略的问题，使得相同图像在不同环境中可以灵活应用，提升了安全性和一致性。意义：为图像处理中的安全性和规范性提供了新的思路。

MetaView: Monocular Novel View Synthesis with Scale-Aware Implicit Geometry Priors

👍 22

07/12 20:00

Current visual generation models are capable of producing high-quality content, yet they lack a coherent perception of the spatial structure. Existing generative novel view synthesis methods typically introduce explicit geometry priors, which enforce spatial consistency but inherently restrict gener

中文介绍 MetaView 通过单目视角合成引入了尺度感知的隐式几何先验，以改善当前视觉生成模型在空间结构理解上的不足。该方法克服了显式几何先验的限制，提升了生成内容的一致性与准确性，为 3D 生成和视觉推理提供了新方法。意义：推动 3D 生成与空间感知技术的发展。

GigaWorld-Policy-0.5: A Faster and Stronger WAM Empowered by AutoResearch

👍 21

07/14 20:00

World Action Models (WAMs) improve robot policy learning by jointly modeling actions and future visual observations, using future scene evolution as dense supervision for physically grounded action generation. However, a common design in existing WAMs is to explicitly generate future videos at infer

中文介绍 GigaWorld-Policy-0.5 提出了利用世界行为模型（WAM）提升机器人政策学习的新方法，通过建模未来场景演变与视觉观察的关系，改善了物理基础的动作生成。此设计强调在推理时利用未来视频生成，推动了动作生成的质量。意义：在机器人自适应和学习的应用中具有重要影响。

Metacognition in LLMs: Foundations, Progress, and Opportunities

👍 19

07/13 13:58

Metacognition is a foundational component of intelligence critical to effective learning, problem solving, decision-making, communication, and more. In recent years, it has become increasingly recognized as a cornerstone of capable, transparent AI systems. Yet while LLMs have made significant progre

中文介绍 元认知是有效学习、问题解决和决策等过程中的基础组成部分。本文回顾了近期 LLM 领域中的元认知研究进展，并探讨了其在智能系统中的应用潜力，进一步强调了透明 AI 系统构建的重要性。意义：对智能系统的学习与决策能力发展具有深远影响。

Self-Improvements in Modern Agentic Systems: A Survey

👍 13

07/13 20:00

Self-improving autonomous agents are moving from research prototypes to deployed systems. The primary goal is controllable evolution, or adaptation, from experience with minimal or even no human input. This survey frames modern self-improving agents as adaptive systems that convert experience into a

中文介绍 自我改进的自主代理系统逐渐从研究原型转向实际应用，旨在实现经验驱动的可控演变。这篇综述阐述了现代自我改进代理的框架，为其适应性系统设计提供了新的视角。通过转换经验为知识，该方法有助于提升代理的自学习能力。意义：推动 Agent 自我学习和适应机制的研究进展。

Proxy Exploration and Reusable Guidance: A Modular LLM Post-Training Paradigm via Proxy-Guided Update Signals

👍 13

07/13 08:56

Post-training is essential for refining the domain-specific capabilities of large language models (LLMs), yet existing reward optimization and distribution matching methods tightly couple policy exploration with distribution alignment. This coupling forces expensive exploration directly on the polic

中文介绍 本文提出了一种新的模块化后培训范式，称为 Proxy-Guided Update Signals，旨在改进大型语言模型（LLMs）的领域特定能力。该方法通过解耦政策探索与分布对齐，减少了现有方法的代价，提高了探索的效率。意义：为后续的模型优化和自主学习策略奠定了基础。

Towards Autonomous and Auditable Medical Imaging Model Development

👍 12

07/11 20:00

Large language model (LLM) agents are beginning to automate machine learning engineering (MLE) by coupling planning, code execution, debugging, and empirical feedback. Translating this capability to medical imaging remains difficult because each task imposes modality-specific experimentation and str

中文介绍 大型语言模型代理开始通过规划、代码执行和调试等任务自动化机器学习工程，但在医学成像中的应用面临挑战。本文探讨了任务特性对模型开发的影响，并提出了适应性强的开发策略，从而加速医疗领域中的技术整合。意义：推动医疗成像和智能代理系统的融合与发展。

Hallo4D: Multi-Modal Hallucination Mitigation for Consistent Spatio-Temporal Generation

👍 11

07/14 20:00

While recent advances in 3D generation have enabled impressive visual synthesis, existing methods often rely on 2D diffusion supervision without explicit mechanisms for geometric consistency, leading to spatial hallucinations such as duplicated structures and misaligned geometry. These issues become

中文介绍 Hallo4D 针对多模态生成中的一致性问题，提出了一种新的方法来减少空间-时间生成中的幻觉现象。通过引入显式几何一致性机制，该模型有效改善了 3D 生成的质量，解决了结构重叠和几何失配等问题，为未来的多模态生成技术提供了新的方向。意义：对三维生成和视觉一致性具有重要的研究价值。

ShortOPD: Recovering Pruned LLMs with Short-to-Long On-Policy Distillation

👍 11

07/13 20:00

Structured pruning is a hardware-friendly way to compress LLMs, but it is mostly validated on multiple-choice recognition tasks, while the same compressed checkpoints can collapse on the free-form generation that deployment actually requires. Two observations trace this gap. First, greedy pass@1 nea

中文介绍 ShortOPD 提出了一种短到长策略的政策蒸馏方法，以恢复经过结构剪枝的 LLMs。这种方法解决了在自由表单生成任务中，压缩模型无法有效工作的挑战，优化了模型在多种生成任务中的表现。意义：为模型压缩与性能恢复提供了创新思路。

Multi-Agent LLMs Fail to Explore Each Other

👍 11

07/13 04:34

Exploration is essential for reliable autonomy in multi-agent systems, yet it remains unclear whether large language model (LLM) agents can explore effectively when interacting with one another. We show that modern LLM agents fail to do so, often exhibiting myopic and polarized interaction patterns

中文介绍 本文探讨了多智能体系统中的探索问题，研究表明现代大型语言模型（LLMs）在相互作用时未能有效探索，常常表现出短视和极端的互动模式。此发现为多智能体系统自主行为的提升提供了重要启示与研究方向。意义：对智能体之间的合作与互动研究具有重要影响。

AgentCompass: A Unified Evaluation Infrastructure for Agent Capabilities

👍 9

07/14 20:00

As Large Language Models (LLMs) evolve into autonomous agents, the need for unified evaluation infrastructure becomes critical. However, current evaluation pipelines remain highly fragmented and tightly coupled, hindering reproducibility and causing redundant engineering. To address this, we introdu

中文介绍 AgentCompass 针对大型语言模型（LLMs）逐步演变为自主智能体的趋势，提出了统一的评估基础设施方案。现有评估流程的分散性限制了其重复性和工程效率，本文通过系统化方法提升了评估标准，解决了当前的技术瓶颈。意义：为智能体能力评估提供了新的框架与方向。

Vinci2: Providing Proactive Assistance in Continuous Egocentric Videos

👍 9

07/13 09:09

When should an intelligent assistant speak up without being asked? Continuous egocentric video offers rich, evolving context that enables a new form of assistance: one that is proactive rather than merely reactive. Yet existing approaches either wait passively for user queries or treat every detecte

中文介绍 Vinci2 在连续自我中心视频中提供主动的辅助方式，改变了现有智能助手的被动响应模式。通过利用丰富和不断演变的上下文信息，该模型能够在合适的时机主动提供帮助，增强用户体验。意义：推动智能助手在人机交互中的主动性与适应性。

You just hired a million bad employees.

@gsivulka · 10.2K 粉丝 · 591.2K 阅 · 512 赞 · 99 转

07/14 12:41

AI was supposed to replace human labor. It did the opposite. For the first time in history, humans are cheaper than software. And AI is creating more jobs than it eliminates. Technology has

中文介绍 AI的出现反而使人类劳动变得更有竞争力，历史上首次人力成本低于软件成本。AI不仅未减少就业，还创造了更多工作机会，反映了技术变革对劳动力市场的影响。

The Self-Driving Company

@amasad · 472.6K 粉丝 · 223.7K 阅 · 560 赞 · 65 转

07/16 13:07

We are beginning to see what happens when a company learns to operate itself. In the past six months, engineers at Replit have nearly tripled code output. Review times held steady. Reversions and

中文介绍 Replit的工程师们在过去六个月内几乎将代码产出提升了三倍，评审时间保持不变。这表明，企业自我运营的能力正在不断提升，这是智能化运营的新趋势。

How OpenAI’s Sol Finally Learned Design Taste

@DesignArena · 17.6K 粉丝 · 191.5K 阅 · 509 赞 · 46 转

07/15 12:36

We benchmarked GPT-5.6 Sol on Design Arena’s Web Design (Non-Agentic) Arena, and we were surprised to find that it ranks 1st overall. This is 18 places higher than its predecessor GPT-5.5, and is the

中文介绍 GPT-5.6 Sol在设计方面的性能显著提升，相较于GPT-5.5排名提高了18位，成功展示了其更优秀的设计能力，标志着AI在创意领域的进一步突破。

The Next Horizon in Agents

@btaylor · 158.6K 粉丝 · 144.3K 阅 · 511 赞 · 40 转

07/16 10:48

Introducing Horizon Today, Sierra is announcing Horizon, a platform that enables agents to pursue long-horizon goals, like originating a loan or getting prior authorization for a healthcare procedure.

中文介绍 Sierra推出了Horizon，一个让智能体追求长远目标的平台，具体应用如低风险贷款或医疗程序的提前授权，揭示了智能体在复杂任务处理上的新发展。

A Framework for Frontier AI and the Dawning of a New Age

@demishassabis · 1.4M 粉丝 · 98.1K 阅 · 1.5K 赞 · 236 转

07/14 05:10

This is a pivotal moment in human history. Artificial General Intelligence (AGI), a system that exhibits all the cognitive capabilities the brain has, is probably only a few short years away. When we

中文介绍 人工通用智能(AGI)的到来迫在眉睫，可能将在几年内实现，预示着人类历史上一个重要的转折期。AGI将具备与人脑相同的认知能力，推动未来的发展。

How we built our knowledge base

@cerebras · 60.4K 粉丝 · 70.7K 阅 · 509 赞 · 38 转

07/16 14:27

Authors: @hi_im_isaac_, @learnwdaniel, @gaozenghao note: the interactive version of full technical blog available: https://www.cerebras.ai/blog/how-we-built-our-knowledge-base Employees ask our internal knowledge base more than 15,000

中文介绍 Cerebras介绍了其内部知识库的构建过程，员工平均每天咨询超过15000次，显示出对知识共享和高效信息管理的迫切需求。

I Brutally Modified My Front-End Design Skill ~ Now My UIs Don’t Look Like AI Crap

@JoePro · 949 粉丝 · 66.8K 阅 · 537 赞 · 29 转

07/13 23:51

To create a custom Claude skill, navigate to Customize > Skills in your Claude interface, click the + button, and select + Create skill

中文介绍 博主分享了如何创建定制的Claude技能，强调在Claude界面中的具体操作步骤，此内容实用性强，帮助用户提升前端设计能力。

Let's build Claude Code's harness (step-by-step)

@akshay_pachaar · 281.7K 粉丝 · 53.7K 阅 · 516 赞 · 74 转

07/15 14:10

We will cover everything that goes into building a coding harness, the agent loop, planning, subagents, sandboxing, memory, and checkpointing, built step by step. If you've ever tried building your

中文介绍 详细介绍了构建Claude Code编程工具的全过程，包括代理循环、规划和记忆等方面，适合有意探索智能代码构建的开发者参考。

How to build a self-improving code review agent

@zachlloydtweets · 12.1K 粉丝 · 42.7K 阅 · 527 赞 · 57 转

07/15 12:20

This is the third post in my series on how to build a cloud software factory: Build an issue triage agent Build a spec writing agent [This post] Build a self-improving code review agent I’ll describe

中文介绍 博主分享了如何构建自我完善的代码审查代理的系列文章，强调了在云软件工厂中的应用，展示了智能体在持续改进方面的潜力。

Your AI says it read your files. Did it really?

@conikeec · 7.5K 粉丝 · 1.1K 阅 · 1.0K 赞 · 121 转

07/14 10:39

AI always answers with confidence — even when it skipped half your documents. waggle shows you which files it really opened, so you know whether to trust the answer, or send it back. You hand an AI

中文介绍 介绍了waggle工具，帮助用户核实AI是否真实打开了指定文件，提升了对AI回答可信度的认知，强调了信息透明的重要性。

NVIDIA Nemotron 3 Embed Ranks #1 Overall on RTEB, Advancing Agentic Retrieval

Hugging Face Blog · 07/16 12:01

中文介绍 NVIDIA的Nemotron 3 Embed在RTEB评测中获得首位，显示了其在自主检索方面的优势。

Why teens deserve access to safe AI

OpenAI News · 07/16 12:00

Learn how OpenAI is making ChatGPT safer for teens with age-appropriate protections, learning tools, parental controls, and expert partnerships.

中文介绍 OpenAI致力于为青少年提供安全的AI访问，采用适龄保护措施、学习工具和家长控制，以提升安全性。

🔬 The Lab of the Future Should Feel Like a Data Center — Andy Beam & Rafa Gómez-Bombarelli, Lila Sciences

Latent Space · 07/16 09:30

Lila is betting that science, not the internet, is the last untapped source of training data. We went to find out what that actually looks like in a room full of robots.

中文介绍 Lila科学公司认为科学是未被开发的重要训练数据来源，展示了机器人实验室的前景。

Newer Models, Same Advantage

Hugging Face Blog · 07/16 07:49

中文介绍 新模型具备与旧模型相同的优势。

Our approach to bioresilience

DeepMind Blog · 07/16 05:30

Google DeepMind and Isomorphic Labs are sharing our joint approach to bioresilience and AI models.

中文介绍 谷歌DeepMind与Isomorphic Labs分享了其在生物韧性和AI模型的联合方法。

[AINews] Thinky's Inkling: 975B-A41B multimodal, new best American Apache 2.0 open model (with Inkling-Small, 276B-A12B)

Latent Space · 07/16 02:18

Thinky's first full LLM release is a banger and bonus: it's open weights!

中文介绍 Thinky首次发布的全功能LLM模型Inkling，参数为975B，为Apache 2.0开源模型，带有开放权重。

How Cars24 scales conversations and builds faster with OpenAI

OpenAI News · 07/15 20:00

Cars24 uses OpenAI-powered voice and chat agents to handle 1M+ monthly conversation minutes, recover 12% of lost leads, and bring agentic workflows to teams across the company.

中文介绍 Cars24利用OpenAI的语音和聊天代理每月处理超过100万分钟的对话，并恢复12%的失去线索。

Security incident disclosure — July 2026

Hugging Face Blog · 07/15 20:00

中文介绍 2026年7月安全事件的相关信息披露。

Thinking Machines Inkling 🧠, GPT-Red 🔒, Perplexity sandboxes 🛡️

TLDR AI · 07/15 20:00

中文介绍 关于思考机器Inkling、GPT-Red和Perplexity沙箱的资讯。

What building Shippy taught us about building agents

Hugging Face Blog · 07/15 13:29

中文介绍 在构建Shippy的过程中获得的经验教训。

Model Routing Is Simple. Until It Isn’t.

Hugging Face Blog · 07/15 13:27

中文介绍 关于模型路由的探讨，简单时各有优缺点。

Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer

MIT Tech Review AI · 07/15 13:09

OpenAI has built an LLM super-hacker called GPT-Red that it uses as a sparring partner to help its other models boost their defenses against cyberattacks. Last week the company released the latest version of its flagship LLM, GPT-5.6. OpenAI says that training it against GPT-Red made the model its m

中文介绍 OpenAI开发的LLM超级黑客GPT-Red用于提升其他模型的防御能力，最新版本为GPT-5.6。

The US is advancing AI safety through state and federal action

OpenAI News · 07/15 08:00

OpenAI outlines a “reverse federalism” approach to AI governance, where state laws help build a national framework for safe, democratic AI.

中文介绍 OpenAI概述了一种“逆联邦主义”的AI治理方法，州法律有助于构建安全和民主的全国框架。

GPT-Red: Unlocking Self-Improvement for Robustness

OpenAI News · 07/15 06:00

Explore GPT-Red, OpenAI’s automated red teaming system that uses self-play to improve AI safety, alignment, and prompt injection robustness.

中文介绍 探索GPT-Red，这是OpenAI的自动红队系统，通过自我对弈提高AI的安全性和鲁棒性。

not much happened today

Smol AI News · 07/15 01:44

**Thinking Machines Lab** launched **Inkling**, its first fully released open-weights foundation model family, featuring **975B parameters** with **41B active parameters** in a **Mixture-of-Experts** architecture. Inkling supports **multimodality** with text, image, and audio inputs and text output,

中文介绍 Thinking Machines实验室推出首个完全开源的基础模型家族Inkling，参数为975B，支持多模态输入。