2026-07-18

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

👍 187

07/14 08:00

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 提出 Harness Handbook，让不断演化的 AI agent 的 harness（prompt 构建、状态管理、工具调用等）变得可读、可导航和可编辑，解决组件持续修改带来的维护挑战。

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

👍 122

07/14 08:00

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 发布开源统一多模态理解与生成模型 Boogu-Image-0.1 系列，包含 Base、Turbo、Edit 和 Edit-Turbo 变体。在高质量文本到图像生成、快速推理、指令编辑和双语理解上具备竞争力。

VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding

👍 108

07/16 08:00

Recent advances in video understanding have spanned motion, long video, and streaming interaction, driving this field toward real-world applications. Despite this progress, current open-source models remain limited in several ways. They often struggle to generalize across diverse video types, making

中文介绍 提出 VideoChat3，一个完全开源的多模态视频理解大模型。旨在提升对多样视频类型的泛化能力，支持高效、通用的视频理解，弥补现有开源模型在通用性上的局限。

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

👍 84

07/14 08:00

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 提出 Ring-Zero，将可验证奖励的零强化学习（Zero RL）扩展到万亿参数规模。突破了此前小模型的计算限制，成功催生大模型的链式推理能力。

SEED: Self-Evolving On-Policy Distillation for Agentic Reinforcement Learning

👍 66

07/16 08:00

Large language models are increasingly trained as interactive agents for long-horizon tasks involving multi-turn interaction, tool use, and environment feedback. Outcome-based reinforcement learning (RL) provides a practical optimization paradigm, but its sparse trajectory-level rewards offer limite

中文介绍 提出 SEED，一种自进化在线策略蒸馏方法，用于长程多轮智能体强化学习。通过在线蒸馏提供更密集的监督信号，克服结果型 RL 稀疏奖励的问题。

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

👍 52

07/15 08:00

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 提出 KnowAct-GUIClaw，一个具有自进化记忆和技能的个人 GUI 助手。改进 OpenClaw 框架，增强跨平台 GUI 交互支持，并通过持续自我进化提升任务自动化性能。

SearchOS-V1: Towards Robust Open-Domain Information-Seeking Agent Collaboration

👍 49

07/16 08:00

Recent advances in Tool-Integrated Large Language Models have made web search a core capability of information-seeking agents. However, as interaction histories grow, agents increasingly struggle to track task progress. When search attempts fail to yield useful evidence, current single- and multi-ag

中文介绍 提出 SearchOS-V1，一种鲁棒的开放域信息搜索智能体协作系统。解决长交互历史中任务跟踪困难及搜索失败后证据不足的问题，提升搜索协作的稳健性。

LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget

👍 40

07/16 08:00

A growing gap separates inference context lengths from RL post-training: inference systems are approaching million-token contexts, while post-training workloads often remain at 256K tokens or below and rely on length generalization at deployment. The gap is especially important for AI agents, whose

中文介绍 提出 LongStraw，在固定 GPU 预算下将强化学习后训练扩展到超过 2M tokens 的长上下文。弥合推理与训练间长度差距，满足 AI 智能体对超长上下文的需求。

BadWAM: When World-Action Models Dream Right but Act Wrong

👍 36

07/16 08:00

World-action models (WAMs) are emerging as a promising foundation for embodied control: rather than predicting actions alone, they learn representations that couple action generation with future world prediction. This coupling is often viewed as a source of robustness, interpretability, and safety,

中文介绍 研究世界－动作模型（WAM）中「梦想正确但行动错误」的现象。揭示模型在联合生成未来预测与动作时可能出现的失效模式，强调其安全与鲁棒性挑战。

PolicyShiftGuard: Benchmarking and Improving Policy-Adaptive Image Guardrails

👍 35

07/07 08:00

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 提出 PolicyShiftGuard，构建策略自适应图像安全护栏的基准并改进。解决同一图像在不同产品、政策边界下安全判断不一致的问题，推动护栏的灵活部署。

AgentCompass: A Unified Evaluation Infrastructure for Agent Capabilities

👍 34

07/15 08:00

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 提出 AgentCompass，一个统一的智能体能力评估基础设施。解决当前评估流程碎片化、耦合度高的痛点，提升可复现性并减少重复工程。

KeyFrame-Compass: Towards Comprehensive Evaluation of Keyframe-Conditioned Video Generation

👍 30

07/15 08:00

Video generation increasingly relies on keyframe-based workflows, where creators specify a sequence of reference images to guide generation. Although recent models support multi-keyframe conditioning, it remains unclear whether they can faithfully reproduce the prescribed keyframes while maintaining

中文介绍 提出 KeyFrame-Compass，全面评估关键帧条件视频生成模型。考察模型是否忠实再现指定关键帧并保持时间一致性，填补该方向系统评估的空白。

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

👍 30

07/13 08:00

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 提出 MetaView，通过尺度感知隐式几何先验实现单目新视角合成。在保证空间一致性的同时避免显式几何先验对生成多样性的限制。

MultiRef-Compass: Towards Comprehensive Evaluation of Multi-Reference-to-Audio-Video Generation

👍 29

07/15 08:00

Multi-reference-to-audio-video (MR2AV) generation aims to generate coherent audio-video content conditioned on multiple references and textual instructions. Existing benchmarks mainly focus on text-driven generation, single-reference subject preservation, or isolated audio-video alignment, leaving t

中文介绍 提出 MultiRef-Compass，全面评估多参考到音频-视频生成（MR2AV）。填补现有基准在文本驱动、单参考及音视频对齐评估上的不足。

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

👍 25

07/15 08:00

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 提出 GigaWorld-Policy-0.5，一个更快更强的世界-动作模型（WAM）。通过 AutoResearch 驱动，避免推理时显式生成未来视频，大幅提升效率。

Concurrent Image Understanding and Generation: Self-Correcting Coupled Markov Jump Processes

👍 23

07/14 08:00

Human cognition does not separate understanding and generation. A teacher at a whiteboard speaks and draws together, each modality reshapes the other. In this paper, we bring this coupled loop to artificial systems. Masked Diffusion Models (MDMs) are ideally suited to this task, yet existing sampler

中文介绍 提出自纠正耦合马尔可夫跳跃过程，实现图像理解与生成的同时进行。利用掩码扩散模型的耦合特性，让两种模态相互塑形。

UniVR: Thinking in Visual Space for Unified Visual Reasoning

👍 22

07/14 08:00

Learning broad world knowledge directly from raw visual data is a fundamental capability of intelligence. We introduce UniVR, the first investigation into simultaneously learning complex reasoning, fine-grained physical dynamics, and long-term planning from pure visual demonstrations. At its core, U

中文介绍 提出 UniVR，首次从纯视觉演示同时学习复杂推理、细粒度物理动力学和长期规划。在视觉空间中实现统一视觉推理，无需文本标注。

Self-Improvements in Modern Agentic Systems: A Survey

👍 21

07/14 08:00

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 综述现代智能体系统中的自我改进框架。将自主智能体视为将经验转化为可控进化的自适应系统，探讨最小化或无需人工输入的持续适应技术。

Spectral Rewiring for Exploration, Purification, and Model Merging

👍 20

07/03 08:00

Reinforcement learning has become a standard post-training recipe for large language models, but dense full-parameter updates create two deployment-relevant bottlenecks: suppressed reasoning performance, often reflected by premature saturation of test-time scaling, and interference when consolidatin

中文介绍 提出光谱重连（Spectral Rewiring）方法，用于大语言模型强化学习后训练的探索、净化与模型合并。缓解全参数更新导致的推理性能饱和和参数干扰问题。

Video = World + Event Stream

👍 15

07/16 08:00

We present Wan-Streamer v0.3, which reframes our native-streaming interaction model under a single organizing view: a video is a world plus an event stream. The world is the persistent context in which a video unfolds, including the environment, scene, subjects, ambient acoustic conditions, voice ch

中文介绍 提出 Wan-Streamer v0.3，将视频重新定义为「世界 + 事件流」。世界为持久环境与场景上下文，事件流为动态变化，实现原生流式交互模型。

GLM 5.2 With Vision

@part_harry_ · 1.5K 粉丝 · 225.6K 阅 · 503 赞 · 46 转

07/16 12:24

GLM 5.2 is one of the best currently available open source language models. However, unlike other flagship models like Qwen, Kimi and Minimax, GLM 5.2 does not support image inputs. We took this as a

中文介绍 开源语言模型GLM 5.2原本不支持图像输入，作者通过扩展使其具备视觉能力，从而与Qwen、Kimi等旗舰模型看齐。展示了如何为纯文本模型添加多模态能力。

The Self-Driving Company

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

07/17 01:07

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的工程师借助AI自主运营，过去6个月代码输出增长近3倍，审核时间保持不变。展示了AI驱动公司自我运作的实际效果和效率提升。

How OpenAI’s Sol Finally Learned Design Taste

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

07/16 00:36

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 在Web设计非代理基准测试中，OpenAI的GPT-5.6 Sol排名第一，比前代GPT-5.5提升18位。证明新模型在设计审美上取得显著突破。

The Next Horizon in Agents

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

07/16 22:48

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发布Horizon平台，使agent能够执行贷款审批、医疗预授权等长期任务。标志着agent从短期互动迈向长周期复杂目标的新阶段。

Gemini API Managed Agents: free tier, triggers & more

@GoogleAIStudio · 186.6K 粉丝 · 111.0K 阅 · 512 赞 · 51 转

07/17 01:05

Today we’re announcing new capabilities for Managed Agents in Gemini API, including free tier availability, budget control guardrails, and scheduled triggers. Building on our previous release of

中文介绍 Gemini API的Managed Agents新增免费层、预算控制防护和定时触发器功能。降低门槛的同时增强企业级可控性，适合自动化部署场景。

How we built our knowledge base

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

07/17 02:27

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公开其内部知识库构建技术，员工每日查询超过1.5万次。详细介绍了搜索、索引和交互式设计，为大型组织提供可参考的实践方案。

Kimi K3 on KernelBench

@elliotarledge · 40.7K 粉丝 · 69.3K 阅 · 506 赞 · 45 转

07/17 17:24

This is the Kimi K3 post you guys have been waiting for. I got some early access to this model and have been testing it on kernels, and even before seeing the benchmark scores I was impressed with its

中文介绍 早期测试显示，Kimi K3在KernelBench内核性能测试中表现突出，在正式榜单发布前已令作者印象深刻。证实该模型在低层次计算任务上的潜力。

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

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

07/16 02:10

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 分步教程：构建Claude Code的编程harness，涵盖agent循环、规划、子agent、沙箱、记忆和检查点。适合想从零搭建自主coding agent的开发者。

How to build a self-improving code review agent

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

07/16 00:20

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 云软件工厂系列第三篇：构建自我改进的代码审查agent。该agent能通过经验自动优化审查规则，减少人工干预，提升代码质量审核效率。

Fine-tune video and image models at scale with NVIDIA NeMo Automodel and 🤗 Diffusers

Hugging Face Blog · 07/17 23:57

中文介绍 NVIDIA NeMo Automodel 与 Hugging Face Diffusers 结合，支持大规模微调视频和图像模型，提升规模化微调效率。

A scorecard for the AI age

OpenAI News · 07/17 18:00

Sarah Friar, CFO of OpenAI, introduces a practical AI scorecard to measure ROI through useful work, cost per successful task, dependability, and return on compute.

中文介绍 OpenAI CFO Sarah Friar 推出AI记分卡，通过“有用工作”、“单次成功任务成本”、“可靠性”和“计算回报”衡量AI投资回报率。

The risk of weather data sabotage is rising

MIT Tech Review AI · 07/17 16:57

Every morning, airline dispatchers, grid operators, and farmers around the world make decisions based on the same thing: a weather forecast. While these forecasts are something that most people glance at for two seconds, weather predictions influence major strategic decisions in many industries, wit

中文介绍 天气数据破坏风险上升。全球调度员、电网运营商和农民依赖天气预报做重大决策，数据篡改可能带来严重后果。

not much happened today

Smol AI News · 07/17 13:44

**Moonshot's Kimi K3 release** has sparked a reassessment of **Chinese open-weight models**' proximity to the frontier, with strong performance in coding, agentic tasks, and long-horizon knowledge work. The strategic focus has shifted from a "compute moat" to an "efficiency stack" involving **MoE ro

中文介绍 Moonshot 的 Kimi K3 发布引发对中国开源权重模型接近前沿水平的重新评估，在编码、智能体任务和长期知识工作中表现强劲，战略焦点从“计算护城河”转向“效率栈”。

[AINews] Kimi K3 2.8T-A50B: the largest open model ever released; Opus 4.8-class at Sonnet 5 pricing

Latent Space · 07/17 09:46

a great week for open models continues.

中文介绍 Kimi K3 拥有2.8T参数和50B激活参数，是迄今最大开源模型，性能达Opus 4.8级别，定价与Sonnet 5相当。

Kimi K3 🌕, Gemini 3.5 delayed ⏳, crushing ARC-AGI 3 🤖

TLDR AI · 07/17 08:00

中文介绍 Kimi K3 发布，Gemini 3.5 延迟，ARC-AGI 3 被攻克。

Why teens deserve access to safe AI

OpenAI News · 07/17 00:00

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 分享如何通过年龄适配保护、学习工具、家长控制和专家合作，让 ChatGPT 对青少年更安全。

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

Latent Space · 07/16 21:30

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 Sciences 认为科学是最后的数据来源，其实验室像数据中心一样运作，充满机器人，旨在利用科学数据训练AI。

Newer Models, Same Advantage

Hugging Face Blog · 07/16 19:49

中文介绍 讨论新模型保留相同优势，具体内容未详细披露。

Our approach to bioresilience

DeepMind Blog · 07/16 17:30

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

中文介绍 Google DeepMind 与 Isomorphic Labs 分享在生物弹性及AI模型方面的联合方法。

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

Latent Space · 07/16 14:18

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

中文介绍 Thinky 发布 Inkling 系列，975B参数、41B激活的多模态模型，是Apache 2.0许可下美国最佳开源模型，同步推出 Inkling-Small（276B-A12B）。

not much happened today

Smol AI News · 07/16 13:44

**Moonshot AI** launched **Kimi K3**, a frontier-class open-weights model with **2.8T parameters**, **1M-token context window**, and **native multimodal input**. It features novel **Kimi Delta Attention (KDA)** enabling up to **6.3x faster decoding** and **Attention Residuals** for **~25% higher tra

中文介绍 Moonshot AI 发布 Kimi K3，前沿开源权重模型，2.8T参数，100万token上下文窗口，原生多模态输入，采用Kimi Delta Attention实现6.3倍更快解码和注意力残差技术。

How Cars24 scales conversations and builds faster with OpenAI

OpenAI News · 07/16 08:00

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 使用 OpenAI 语音和聊天代理处理每月超100万分钟对话，挽回12%丢失线索，并在公司团队中推广智能体工作流。

Security incident disclosure — July 2026

Hugging Face Blog · 07/16 08:00

中文介绍 Hugging Face 披露2026年7月安全事件，具体细节尚未公开。

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

TLDR AI · 07/16 08:00

中文介绍 Thinking Machines 推出 Inkling 模型，GPT-Red 安全相关，Perplexity 推出沙箱功能。