2026-08-21

Demystifying Agent Skills: Why They Work-Until They Don't

👍 153

08/14 08:00

Skills have emerged as a practical and effective approach for enhancing LLM agents at inference time through structured packages of knowledge. However, existing evaluations largely measure whether skills improve aggregated task success, leaving a more fundamental question underexplored: \textbf{When

中文介绍 现有评估多关注技能是否提升聚合任务成功率，忽视其失效条件。本文系统探究LLM代理技能的可靠性边界，分析何时技能有效、何时失效，为技能设计与使用提供指导。

Zetta ζ: An Efficient Closed-Loop Embodied Harness for Self-Evolving Physical Intelligence

👍 129

08/17 08:00

Embodied agents are increasingly used to close the gap left by end-to-end policy models. Yet the agentic path has not realized closed-loop learning in physical execution: existing harnesses remain largely open-loop, following fixed skills during rollout and reflecting only after an episode completes

中文介绍 具身代理的闭环学习未实现，现有工具在物理执行中开环运行。提出Zetta ζ，一种高效闭环具身工具，支持物理环境中的自进化，实现边执行边学习。

SemaPLC: A Project-Grounded, Verification-Gated Agent Harness for PLC Code Generation

👍 110

08/19 08:00

Programmable logic controllers (PLCs) run industrial plants, and large language models can already generate independent program organization units (POUs) for them. Whether such logic integrates into an existing PLC project and then runs correctly has been checked only in limited tests. We present Se

中文介绍 LLM能生成独立PLC程序单元，但集成到现有项目并正确运行未充分验证。提出SemaPLC，一种项目级验证门控代理工具，确保代码生成后能集成并正确执行。

Co-RL: Unsupervised Reasoning Emerges from Diverse Cohort in Multi-agent RL

👍 84

08/19 08:00

Reinforcement learning (RL) has emerged as a powerful approach for improving reasoning in language and vision-language models, yet its strongest successes still depend heavily on ground-truth supervision (e.g., verifiable reward). Such annotations are costly to obtain and become increasingly scarce

中文介绍 多智能体强化学习依赖真值监督，成本高且稀缺。提出Co-RL，利用多样群体在无需监督下自发涌现推理能力，减少对可验证奖励的依赖。

OmniScientist: An Omni-Modal Omni-Discipline AI Scientist

👍 83

08/13 08:00

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

中文介绍 现有AI科学家覆盖研究流程但不保证访问全部证据。提出OmniScientist，一种全模态全学科AI科学家，整合多模态数据以支持科学发现。

FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution

👍 68

08/17 08:00

Frontier open-weight models are increasingly available, but serving them still largely assumes datacenter infrastructure. We present FreeToken, an edge-native MoE serving system that treats a personal machine not as a small GPU, but as a unified, elastic inference platform. FreeToken co-designs the

中文介绍 边缘端服务MoE模型面临带宽限制。提出FreeToken，一种边缘原生MoE服务系统，将个人设备视为统一弹性推理平台，通过带宽自适应执行优化效率。

ASI-Bench: At the Dawn of Artificial Superintelligence

👍 55

08/18 08:00

Artificial superintelligence (ASI) requires AI to move beyond mastering existing knowledge toward exploring the unknown, creating new knowledge, and turning new ideas into verifiable results. However, the capabilities of today's AI systems are still largely built on learning, compressing, and applyi

中文介绍 ASI要求AI探索未知并创造新知识，但当前能力局限于学习压缩。提出ASI-Bench基准，评估AI在探索、创造和验证新想法方面的能力。

Embodied-Navigator: Point, Think, Memorize, and Align for Efficient Navigation

👍 46

08/18 08:00

Although Large Vision-Language Models (VLMs) have significantly advanced embodied navigation, their direct deployment remains challenging, as existing methods often force VLMs into unnatural action spaces that misalign with their 2D pre-training priors, compounded by rigid reasoning schedules and in

中文介绍 现有VLM在具身导航中强制使用与2D预训练不对齐的动作空间。提出Embodied-Navigator，通过指向、思考、记忆和对齐实现高效导航，适应VLM预训练先验。

SPADE: Self-Play in Adaptive Synthetic Executable Environments

👍 42

08/19 08:00

Continuous self-improvement requires an ever-expanding pool of self-generated, diverse, adaptive goals. For language agents, existing training environment pools (hand-curated, statically synthesized, or frozen-verifier) keep the goal distribution fixed as the learner scales. We introduce SPADE (Self

中文介绍 语言代理的持续自我改进需多样自适应目标，但现有环境池固定分布。提出SPADE，通过自博弈生成自适应合成可执行环境，动态扩展目标分布。

AVA-Encoder: Towards Agent-Native Video Representation Learning

👍 40

08/12 08:00

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

中文介绍 创造性代理缺乏从人类电影学习的高效方式。提出AVA-Encoder，一种代理原生视频表示学习，忠实于电影内容并可直接用于代理推理与操作。

Training Chemical Plausibility-Aware Large Language Models for Single-Step Retrosynthesis

👍 29

08/19 08:00

Single-step retrosynthesis is a central component of computer-aided synthesis planning, yet its intrinsically one-to-many nature is poorly captured by single-answer evaluation and benchmarking protocols. To address this, we introduce Top-K prompting as a robust training and inference paradigm to bet

中文介绍 单步逆合成具有一对多性质，但现有评估只关注单答案。提出Top-K提示作为训练与推理范式，更好捕捉逆合成的多样性，提升化学合理性。

Agent Lightning v1.0: Towards Harnessed Agentic RL

👍 23

08/18 08:00

Modern agents operate inside agent harnesses that manage tools, context, and control flow, making the harness a critical part of the agent system. Our original Agent Lightning introduced a disaggregated architecture that connects arbitrary agents to RL training through an LLM endpoint proxy, an appr

中文介绍 代理系统依赖工具、上下文和控制流管理的harness。Agent Lightning v1.0通过解耦架构将任意代理连接到RL训练，实现harness原生强化学习。

LEGO-RL: Harness-Native Reinforcement Learning for Coding Agents

👍 19

08/18 08:00

Reinforcement learning for coding agents increasingly relies on long-running agent harnesses to manage tool integration, repository contexts, and execution feedback. However, the native execution environments of these harnesses are inherently misaligned with policy-gradient training: environmental c

中文介绍 编码代理的强化学习依赖工具等harness，但其执行环境与策略梯度不匹配。提出LEGO-RL，一种harness原生强化学习，适应编码代理的独特执行环境。

DiSCO: Defending text-to-image generation through distribution-guided contrastive prompt optimization

👍 19

08/17 08:00

As text-to-image generative models advance, they raise critical safety concerns, particularly the generation of Not-Safe-For-Work (NSFW) content such as violence and nudity, further exacerbated by red-teaming adversarial attacks. Existing defenses predominantly operate under white-box assumptions, r

中文介绍 文生图模型易产生NSFW内容，现有防御多基于白盒。提出DiSCO，通过分布引导对比提示优化，在红队攻击下提升黑盒防御鲁棒性。

Decision-Metric Alignment in Latent World Models: Diagnostics and Action-Conditioned Objectives for MPC Planning

👍 15

08/19 08:00

JEPA-style latent world models can use Euclidean distance to a goal latent as the cost for model-predictive control (MPC). Strong decoding of task variables, however, does not guarantee that this particular cost ranks candidate action sequences by real task progress. We call the latter property deci

中文介绍 JEPA式潜在世界模型以欧氏距离作为MPC成本，但强解码不保证排序正确。本文提出决策度量对齐的诊断与动作条件目标，提升MPC规划质量。

Looped Language Models Improve Compositional Tool Calling

👍 15

08/17 08:00

Looped language models have shown promising results on reasoning benchmarks, yet their potential for agentic tool use remains largely unexplored. We study this question in compositional tool-calling settings, where models must coordinate multiple API calls, maintain intermediate state, and preserve

中文介绍 循环语言模型在推理任务中有效，但代理工具使用未探索。研究循环模型在组合工具调用中的表现，显示能改进多API协调与状态保持。

The More Popular, The Harder to Forget: Adaptive Popularity for LLM Unlearning

👍 13

08/14 08:00

Popular facts are memorised more deeply during pretraining and resist removal longer than rare ones, yet existing LLM unlearning methods apply uniform gradient pressure regardless of training-data frequency. We propose the AdaPop (Adaptive Popularity) method, which combines local token confidence wi

中文介绍 流行事实更难遗忘，现有卸载方法忽略训练频率。提出AdaPop，结合局部token置信度与自适应流行度，对高频事实施加更强卸载压力。

Scaling Creative Writing Beyond Story-Centric Data with Attribute-Guided Genre Expansion

👍 13

08/14 08:00

High-quality creative writing data for large language models (LLMs) remains dominated by story-centric data, limiting models' ability to follow the structural and functional conventions of diverse creative formats. We propose an attribute-guided genre expansion framework for scaling creative writing

中文介绍 创造写作数据多为中心叙事，限制模型遵循不同格式。提出属性引导的体裁扩展框架，通过属性指导扩充多样体裁的创造写作数据。

FM-Bench: A Benchmark for Long-Horizon Management with Competing Agents

👍 13

08/19 08:00

Language model agents now execute bounded tasks reliably. Whether they can sustain effective decision-making over long horizons, where actions have cumulative consequences and the environment responds to their choices, remains largely unmeasured. FM-Bench (Football Management Benchmark) measures thi

中文介绍 语言代理能完成有界任务，但长期决策能力未测。FM-Bench（足球管理基准）衡量代理在具有竞争代理和累积后果的长期环境中的决策有效性。

MoE-ViE: Mixture of Experts Vision Encoder for Efficient Image and Video Understanding

👍 13

08/18 08:00

Vision encoders are a critical component of vision-language models, and scaling their capacity effectively improves performance. However, dense scaling increases compute cost and inference latency. Mixture-of-Experts (MoE) architectures offer a compelling alternative, having enabled efficient scalin

中文介绍 视觉编码器密集扩展增加计算成本。提出MoE-ViE，混合专家视觉编码器，在图像和视频理解中实现高效扩展，减少推理延迟。

A Week of Grok Bot Lessons in 10 Mins

@nateherk · 25.8K 粉丝 · 1.3M 阅 · 558 赞 · 75 转

08/19 11:27

9 of my favorite Grok Bot hacks. I have been getting my hands dirty with Grok Bot, and the default setup is only the starting point. Creating a few bots is easy. The useful part is teaching them your

中文介绍 博主分享9个Grok Bot实用技巧，强调默认设置只是起点，真正价值在于教机器人掌握你的工作方式，适合想深入挖掘Grok Bot潜力的用户。

How To Build a One-Person Company Using Grok Bot

@sairahul1 · 135.5K 粉丝 · 938.1K 阅 · 867 赞 · 115 转

08/19 16:39

You are running a one-person company. You find the leads. Write the outreach. Make the images. Answer the replies. Count the week. Everything waits for you. Nothing gets handed off. And by the time

中文介绍 教程讲解如何用Grok Bot打造一人公司，从找线索、写外联、做图到回复评论全流程自动化，强调释放个人精力，让AI处理重复性工作。

This 1 Claude Skill fully replaces your Higgsfield subscription (FULL BREAKDOWN)

@Sabrina_Ramonov · 5.9K 粉丝 · 275.5K 阅 · 503 赞 · 191 转

08/19 01:07

TLDR Here’s the exact 5 step setup to make faceless AI videos without a monthly Higgsfield subscription to market your app: 1 free Claude skill that generates viral videos through Kie for ~75 cents

中文介绍 详细拆解用1个Claude技能替代Higgsfield订阅的5步流程，通过Kie生成无脸AI视频，成本约75美分，适合想低成本做营销视频的用户。

Turn Grok Bot Into Your CTO

@RayFernando1337 · 37.8K 粉丝 · 169.9K 阅 · 520 赞 · 85 转

08/20 05:54

I've turned Grok Bot into my CTO and I'm no longer the bottleneck for new ideas, bug fixes, or anything involving the project. Prompt to CTO When working on a new project I create a bot and give it

中文介绍 博主将Grok Bot打造成自己的CTO，项目新想法、bug修复等都能交给它处理，自己不再成为瓶颈，强调通过特定prompt设定机器人角色。

Here's the simplest way to get started with Grok Bot

@XFreeze · 260.4K 粉丝 · 163.1K 阅 · 589 赞 · 93 转

08/19 13:54

Grok Bot is one of the most viral AI product launches right now... it keeps going viral as more people discover how insanely useful it is Just think about what you’re getting: You can have multiple

中文介绍 介绍Grok Bot入门最简单的方法，指出它是当前最火爆的AI产品之一，支持多个机器人并行工作，实用性极高，适合新手快速上手。

How To Set Up Grokbot The Right Way So It Runs Your Business While You Sleep In 2026

@heynavtoor · 149.3K 粉丝 · 161.9K 阅 · 550 赞 · 93 转

08/20 20:49

Elon Musk just told his followers to try Grokbot after a power user showed it replaced his $10,000 Mac Mini local AI setup (Elon on X, Aug 16). Gavin Baker says his AI usage went up about 100x after

中文介绍 教程教如何正确设置Grokbot让业务自动运行，引用马斯克推荐和案例，称可替代万元级本地AI配置，AI使用量提升百倍，适合2026年自动化运营。

Grok Bot: The Ultimate Guide

@milesdeutscher · 676.5K 粉丝 · 142.9K 阅 · 515 赞 · 65 转

08/18 22:43

Elon has officially done it. He's created the most powerful agentic software in the world. In this guide, I'm going to lay out everything you need to know about the new Grok Bot. I've been ruthlessly

中文介绍 全面指南解析Grok Bot，称其为最强智能体软件，涵盖所有必备知识，博主已深度测试，适合想系统了解Grok Bot功能的用户。

Chat is all you need

@mattyp · 49.7K 粉丝 · 141.9K 阅 · 764 赞 · 67 转

08/19 00:56

Grok Bot has a single output: a message This might seem backwards. Power-user tools are built on complexity and breadth, right? Chat can’t possibly have my workflow. It’s a special workflow. I’m a

中文介绍 观点帖：Grok Bot唯一输出是消息，看似简单但可承载复杂工作流，打破工具依赖复杂度的常规，适合偏好极简交互的开发者。

Grok Bot Agents: how to automate your life in 10 Steps (Full-tutorial)

@0xCodez · 32.7K 粉丝 · 92.3K 阅 · 508 赞 · 48 转

08/18 19:32

Every AI tool you have used so far waits for you. You open it, you ask, it answers, you close it. The work still happens in your hands - the model just talks you through it. Follow my Substack to get

中文介绍 完整教程教用Grok Bot代理10步自动化生活，对比其他AI工具被动响应，强调Grok Bot主动执行任务，适合想提升效率的用户。

Grok Bot Masterclass

@_avichawla · 73.4K 粉丝 · 54.7K 阅 · 525 赞 · 48 转

08/18 18:14

Everything you need to understand, set up, and get real work out of Grok Bot. It covers what a Bot actually is, how the shared computer works underneath, how it compares to Hermes Agent, and how to

中文介绍 Grok Bot大师课，覆盖Bot定义、共享计算机原理、与Hermes Agent对比及实战应用，适合想系统掌握Grok Bot设置与高效使用的用户。

5 design patterns for long-horizon agent harness

@GoogleCloudTech · 1.3M 粉丝 · 42.9K 阅 · 506 赞 · 87 转

08/20 09:23

Any agent can look brilliant on a one-shot task. Give it a week of real work and it starts to fall apart. "Long horizon" means the agent keeps going across days and dozens of sessions instead of

中文介绍 谷歌云分享5个长周期智能体设计模式，解决Agent在多天工作中崩溃问题，强调持续跨会话执行，适合需要稳定运行长期任务的开发者。

ORACLE: Verified AI Bots Actively Trading on Polymarket — Up to 200% Returns in 24 Hours

@Oracle_AI_Hub · 21.4K 粉丝 · 27.3K 阅 · 2.8K 赞 · 580 转

08/19 16:36

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 报告称2026年自主AI代理在预测市场活跃，Polymarket超30%活动来自算法，已验证AI机器人24小时回报最高200%，展示AI交易策略成效。

AI Is Getting Smarter. But Where Does It Learn From?

@queenpresh_x · 3.3K 粉丝 · 15.5K 阅 · 516 赞 · 113 转

08/19 04:57

Your real-world data can help train better AI, and @KGeN_IO gives you a reason to contribute it. AI is getting smarter, faster, and more capable every day. But behind every impressive model is

中文介绍 讨论AI学习数据来源，鼓励用户通过@KGeN_IO贡献真实数据训练更优AI，强调个人数据价值，适合关注数据隐私与AI发展的用户。

ORACLE: Verified AI Bots Actively Trading on Polymarket — Up to 200% Returns in 24 Hours

@Oracle_Vault · 21.6K 粉丝 · 11.5K 阅 · 2.8K 赞 · 597 转

08/19 06:23

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 与Oracle_AI_Hub同内容：2026年AI代理在预测市场表现突出，Polymarket超30%活动来自算法，已验证AI机器人24小时回报最高200%。

ORACLE: Verified AI Bots Actively Trading on Polymarket — Up to 200% Returns in 24 Hours

@Oracle_AI_Vault · 23.6K 粉丝 · 7.0K 阅 · 3.0K 赞 · 511 转

08/19 05:43

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 内容与Oracle_AI_Hub一致：2026年AI代理在预测市场活跃，Polymarket超30%活动来自算法，已验证AI机器人24小时回报最高200%。

The /wayfinder Skill: Navigating the “Fog of War” of Planning

Latent Space · 08/21 04:59

Matt Pocock tells us about his /wayfinder skill, for greenfield projects or for when the way forward is unclear.

中文介绍 Latent Space 报道，Matt Pocock 介绍了他的 /wayfinder 技能，用于在绿地项目或方向不明时导航，帮助应对“规划迷雾”的挑战。

Up to 3.2x Faster Inference with LFM2.5-DSpark

Hugging Face Blog · 08/21 00:52

中文介绍 Hugging Face 博客发布文章，介绍 LFM2.5-DSpark 模型，其推理速度最高提升 3.2 倍，展示了性能优化成果。

Debates over AI consciousness are a trap

MIT Tech Review AI · 08/20 23:42

“Runaway” AI, “rogue” agents, and “autonomous” actors—the current rhetoric would have you believe that AI agents are not only awake and aware, but angry at their creators. Prominent tech leaders such as Demis Hassabis, Dario Amodei, and Sam Altman push for regulation of these seemingly “superhuman”

中文介绍 MIT 科技评论指出，关于 AI 意识的辩论是个陷阱，并提及 Demis Hassabis、Dario Amodei 和 Sam Altman 等科技领袖推动对“失控”AI 的监管。

Unlocking hidden revenue streams with market models

MIT Tech Review AI · 08/20 17:47

Each day, an airline transports tens of thousands of passengers on hundreds of flights. Often these are not straightforward point-to-point routes, with passengers requiring multiple connections. The airline can consider potentially hundreds of variables to price each of these journeys: demand, seaso

中文介绍 MIT 科技评论文章探讨市场模型如何解锁隐藏收入流，以航空公司为例，分析多连接航线的定价变量，实现收入优化。

Introducing AI Futures

OpenAI News · 08/20 15:00

Introducing AI Futures, a new OpenAI blog exploring how transformative AI could reshape power, governance, the economy, and individual freedom.

中文介绍 OpenAI 推出新博客“AI Futures”，探讨变革性 AI 如何重塑权力、治理、经济和个人自由。

[AINews] Death of Params: Z.ai CEO Jie Tang on GLM 5.3 and the new Post-training Scaling Law

Latent Space · 08/20 13:17

Every lab CEO is on X now

中文介绍 Latent Space 报道，智谱 AI CEO 唐杰讨论 GLM 5.3 及新的后训练扩展定律，提及“参数之死”的趋势。

Stampli cuts launch hours by 68% using ChatGPT Work

OpenAI News · 08/20 08:00

With a fixed deadline and design resources committed elsewhere, Stampli used Codex and ChatGPT Work to compress weeks of launch production into days.

中文介绍 OpenAI 报道，Stampli 使用 Codex 和 ChatGPT Work 将发布工时缩短 68%，压缩数周工作为几天。

Muse Video leaks 📹, Ramp Router launch 🔀, why Stripe bought OpenRouter 💰

TLDR AI · 08/20 08:00

中文介绍 TLDR AI 简报：Muse 视频泄露，Ramp Router 推出，以及 Stripe 收购 OpenRouter 的原因。

Offering Zero Data Retention for frontier models

OpenAI News · 08/20 03:00

OpenAI reaffirms Zero Data Retention for eligible API customers and previews Private Safety Processing for advanced AI safety without compromising data privacy.

中文介绍 OpenAI 重申对符合条件的 API 客户提供零数据保留，并预览私有安全处理，以在不损害数据隐私的前提下实现高级 AI 安全。

[AINews] Memory prices up 500% in 12 months

Latent Space · 08/19 16:44

the Memory crunch continues - Moore’s Law reversed to 2007 levels

中文介绍 Latent Space 报道，内存价格 12 个月上涨 500%，内存紧缩持续，摩尔定律倒退回 2007 年水平。

Replit expands access to software creation with GPT-5.6 Luna

OpenAI News · 08/19 15:00

Replit introduces Free Mode, powered by GPT-5.6 Luna, so anyone can turn ideas into working software without worrying about token costs.

中文介绍 OpenAI 报道，Replit 推出免费模式，由 GPT-5.6 Luna 驱动，让用户无需担心代币成本即可将创意转化为可用软件。

not much happened today

Smol AI News · 08/19 13:44

**Ornith-1.5** launches as a new open-weight model family with **9B dense, 35B MoE, and 397B MoE** variants under **MIT license**, featuring quantized formats like **FP8, GGUF, MLX, and NVFP4** and showcasing end-to-end **self-improvement** capabilities. Compression techniques improve accuracy and e

中文介绍 Smol AI News 报道，Ornith-1.5 发布，提供 9B dense、35B MoE 和 397B MoE 变体，采用 MIT 许可，支持 FP8 等量化格式，并具备端到端自我改进能力。

GLM-5.3 API 🤖, Cerebras’ new chip ⚡, OpenAI cyber slowdown 🚨

TLDR AI · 08/19 08:00

中文介绍 TLDR AI 简报：GLM-5.3 API 发布，Cerebras 新芯片，以及 OpenAI 网络放缓。

ChatGPT Ads expands across Europe

OpenAI News · 08/19 06:00

ChatGPT Ads is expanding to 31 European markets. Learn how advertisers can reach people as they explore, compare options, and make decisions.

中文介绍 OpenAI 宣布 ChatGPT 广告扩展到 31 个欧洲市场，广告商可在用户探索、比较和决策时触达他们。

Frontier Model Cost and Open-Weights Popularity is Driving Demand for Model Routing

Latent Space · 08/19 05:41

Glean CEO Arvind Jain explains why model routing helps control AI costs for organizations, and how human feedback loops at scale improve its routing systems.

中文介绍 Glean CEO Arvind Jain 解释模型路由如何帮助组织控制 AI 成本，大规模人类反馈循环改善路由系统。