2026-09-25

The Tasteful Agent: Measuring and Improving Taste in Long-Horizon Tasks

👍 119

09/22 08:00

LLM agents increasingly work on long-horizon tasks, and the decisions they make along the way, such as which hypothesis to test or which implementation to build on, determine the outcome of the whole run. Making these decisions well is becoming a key capability for both engineering and research agen

中文介绍 针对 LLM agent 在长时程任务中沿途决策（如选哪个假设验证、在哪个实现上继续）决定整体成败的问题，论文提出对 agent「品味」的度量与改进方法，分析其在工程与研究场景中的表现并给出提升策略。

SpeakerMem-R1: Speaker-Centered Dual-Track Memory for Multi-Party Dialogue

👍 79

09/22 08:00

Long-term conversational memory in multi-party settings requires more than retrieving relevant content from long-term conversations: it must distinguish who said what, whom each statement concerns, how individuals perceive one another, what information is shared by the group, and how states change o

中文介绍 多party对话的长期记忆不能只靠检索相关内容，还需区分谁说了什么、话语指向谁、成员彼此认知与群体共享信息。论文提出 SpeakerMem-R1，以说话人为中心构建双轨记忆，追踪个体与群体状态的动态变化。

Spatial-Interactor: Learning Spatial Reasoning through Interaction with the Observable Physical World

👍 44

09/19 08:00

Spatial reasoning is essential for vision-language models (VLMs) to understand and act in the physical world. Reasoning in dynamic environments requires VLMs to perceive local state transitions caused by object motion and viewpoint changes and integrate them over long trajectories to maintain an upd

中文介绍 为提升 VLM 的空间推理能力，论文提出 Spatial-Interactor，让模型与可观察物理世界交互，感知物体运动与视角变化带来的局部状态转移，并在长轨迹上整合以维持更新的世界状态，从而支持动态环境中的理解与行动。

HappyWorld-Bench

👍 39

09/21 08:00

Evaluating world models requires assessing both the quality of the worlds they generate and their consistency and responsiveness under exploration, interaction, and modification. We introduce HappyWorld-Bench, a comprehensive benchmark that evaluates whether generated worlds remain reliable as agent

中文介绍 论文提出 HappyWorld-Bench，从生成世界的质量，以及在探索、交互与修改下的一致性和响应性两方面评估世界模型，检验所生成世界作为 agent 环境是否持续可靠。

The Past Frames the Future: Memory for Autoregressive Video Generation

👍 36

09/23 08:00

Advances in generative models have improved video fidelity, enabling long-horizon generation, interactive world modeling, and evolving visual environments. Autoregressive (AR) video generation extends visual sequences through causal rollouts. However, a fundamental bottleneck emerges: as the generat

中文介绍 自回归视频生成通过因果 rollout 延长视觉序列，但随生成推进出现记忆瓶颈。论文提出面向自回归视频生成的记忆机制「The Past Frames the Future」，利用历史帧信息缓解长时程生成中的退化问题。

Ovis-Embedding: Pushing the Frontiers of Universal Omni-Modal Embeddings

👍 36

09/21 08:00

In this report, we introduce Ovis-Embedding, a state-of-the-art omni-modal embedding family built on native integration of text, image, video, and audio. Instead of assembling separate modality towers, Ovis-Embedding uses a shared multimodal backbone to encode different modalities in a common repres

中文介绍 Ovis-Embedding 是一个全模态嵌入模型家族，不采用分离的模态塔，而是用共享多模态 backbone 将文本、图像、视频与音频编码到统一表示空间，实现原生跨模态融合，刷新全模态嵌入的性能上限。

Just-in-Time Memory: Learning to Curate Task-Adaptive Memory for LLM Agents

👍 33

09/23 08:00

Agentic memory systems reuse past experience to improve future performance, yet most existing designs curate memory at write time: once a task is completed, its trajectory is distilled into a fixed artifact, such as a reflection, workflow, skill, or reasoning strategy, that is later retrieved by sim

中文介绍 现有 agent 记忆系统多在写入时把轨迹固化为反思、工作流或技能，检索时无法随任务调整。论文提出「即时记忆」（Just-in-Time Memory），让 agent 在任务执行中按需筛选与策展自适应记忆，提升经验复用效果。

Circuit Hypernetworks for Quantum-Augmented Diffusion Language Models

👍 27

09/21 08:00

Language models can be adapted by changing the computations applied to individual tokens. Quantum circuits offer one such approach, but evaluating wider circuits inside a large model can be computationally demanding. Here we introduce HyperQ, which adds token-conditioned quantum residual branches to

中文介绍 论文提出 HyperQ，在扩散语言模型中引入以 token 为条件的量子残差分支，通过超网络生成量子电路参数，避免在大模型内部直接评估宽量子电路的高昂开销，从而以量子增强方式适配 token 级计算。

Bellman Policy Optimization

👍 27

09/14 08:00

Reinforcement learning with verifiable rewards (RLVR) improves the reasoning capabilities of large language models (LLMs). We introduce Bellman Policy Optimization (BPO), a critic-free method derived from Policy Mirror Descent (PMD). For autoregressive generation with terminal rewards, BPO uses the

中文介绍 论文提出 Bellman Policy Optimization（BPO），一种从 Policy Mirror Descent 推导出的无 critic 方法。针对带终局奖励的自回归生成，BPO 利用 Bellman 式信号改进可验证奖励强化学习（RLVR）对 LLM 推理能力的提升。

JEV-as-a-Judge: Accept When Confident, Escalate When Unsure

👍 26

09/22 08:00

LLM-as-a-judge enables evaluation across diverse tasks, but inference cost and confidence reliability become critical at scale. We study whether a decision-only judge can provide an economical first pass and identify when stronger evaluation is needed. Comparing jev-as-a-judge with sixteen generativ

中文介绍 论文研究仅做判断的 judge 能否作为低成本初筛，并识别何时需要更强评估。作者提出 JEV-as-a-Judge「自信则接受、不确定则升级」的策略，与十六种生成式评估方法对比，在规模化 LLM-as-a-judge 场景中平衡成本与置信可靠性。

StableVQ: Practical Guidelines for Stable Vector-Quantized Tokenizer Training

👍 25

09/22 08:00

Vector Quantization (VQ) is fundamental to discrete visual tokenizers that power modern autoregressive and masked image generation models. While recent shared-projection codebook methods have substantially advanced codebook utilization, training stability remains a critical and underexplored challen

中文介绍 向量量化（VQ）是离散视觉 tokenizer 的基础，但训练稳定性长期缺乏系统研究。论文针对共享投影码本方法，系统分析不稳定来源，并给出使 VQ tokenizer 稳定训练、提升码本利用率的实用准则。

From Pattern Recognizers to Personalized Companions: A Survey of Large Language Models in Mental Health

👍 16

09/21 08:00

The rising global prevalence of mental health conditions, together with longstanding barriers in traditional healthcare, such as limited resources, high cost, stigma, and privacy concerns, has created an urgent need for accessible and scalable support. Large Language Models (LLMs) have emerged as a

中文介绍 综述全球心理疾病负担上升、而传统医疗受资源不足、成本高、污名化与隐私问题制约的背景，梳理 LLM 在心理健康领域的应用，呈现从「模式识别器」到「个性化陪伴者」的演进、能力边界与风险。

PACT: From Credit Assignment to Critic Alignment

👍 15

09/22 08:00

Reinforcement learning has become a central component of large language model (LLM) post-training, yet token-level credit lacks a generally accepted mathematical definition, leaving its relationship to commonly used training signals unclear. We formulate three regularity conditions, namely Completen

中文介绍 论文指出 LLM 后训练中 token 级信用分配缺乏公认的数学定义，提出完整性等三条正则条件，并据此推出 PACT，将信用分配问题转化为 critic 对齐，澄清其与常用训练信号之间的关系。

Schrödinger's Code Repository: Have LLMs Learned SWE-bench or Memorized It?

👍 14

08/21 08:00

Repository-level coding benchmarks have become the standard for evaluating coding agents, yet they inherently suffer from data leakage because they are built upon popular open-source repositories repeatedly used for training. Consequently, strong performance may reflect memorization of canonical rep

中文介绍 仓库级编码基准建立在被反复用于训练的开源仓库之上，存在数据泄漏，高分可能源于记忆而非真实编程能力。论文以「薛定谔的代码库」为题，系统检验 LLM 究竟是真正学会还是记住了 SWE-bench。

Agensh: Scaling Organizational Intelligence to 1,024 Agents

👍 14

09/22 08:00

A multi-agent system can reduce latency on complex tasks by executing work concurrently. Several pioneering harness frameworks support multi-agent systems. However, the scalability of current multi-agent harnesses is often constrained by a central orchestrator's capacity to allocate tasks and coordi

中文介绍 多 agent 系统可借并发降低复杂任务延迟，但现有 harness 的扩展性受中心编排器的任务分配与协调能力限制。Agensh 将组织智能扩展至 1024 个 agent，突破中枢瓶颈，实现大规模并行的多 agent 协作。

Lean Pool: An AI-Maintained Archive of Formalized Mathematics

👍 14

09/21 08:00

Lean Pool is a repository of formalized mathematics. It is grown, maintained and optimized by AI agents.

中文介绍 Lean Pool 是一个形式化数学仓库，其内容由 AI agent 自行生长、维护与优化，展示了用 agent 持续扩充和整理形式化证明库的可行路径。

PackLab: A Comprehensive Framework for Developing, Training, and Evaluating MLLMs in Robotic Bin Packing

👍 10

09/20 08:00

Robotic bin packing requires long-horizon sequential decision-making, as each object placement affects the available space for subsequent packing. Existing methods primarily rely on hand-crafted geometric heuristics that optimize predefined objectives or reinforcement learning policies learned throu

中文介绍 机器人装箱是长时程序列决策问题，每次摆放都会改变后续可用空间，现有方法多依赖手工几何启发式或强化学习策略。PackLab 提供统一框架，用于开发、训练与评测面向装箱任务的多模态大模型。

MemBodied: Recurrent Associative Memory for Vision-Language-Action Models

👍 10

09/23 08:00

Vision-Language-Action models provide a strong foundation for general-purpose robot control, yet a vast majority of policies do not preserve and leverage episode-level information beyond the current observation. This limitation is consequential in history-dependent manipulation tasks that depend on

中文介绍 现有 Vision-Language-Action 策略大多不保留当前观测之外的情节级信息，在依赖历史的操作任务中表现受限。论文提出 MemBodied，为 VLA 模型引入循环联想记忆，保存并利用历史信息以提升长时程操作性能。

Hunyuan-A13B Technical Report

👍 8

09/23 08:00

We present Hunyuan-A13B, an open-source large language model based on a Mixture-of-Experts architecture. It contains 80 billion total parameters but activates only 13 billion during inference, balancing model capability, computational efficiency, and deployment cost. The model is pretrained on a rig

中文介绍 Hunyuan-A13B 技术报告发布一个基于 MoE 架构的开源大模型，总参数 80B、推理仅激活 13B，在模型能力、计算效率与部署成本之间取得平衡，并公开了预训练与后训练细节。

WhatWorkedBench: Benchmarking Experimental Understanding in AI Agents

👍 8

09/23 08:00

AI research agents need reliable knowledge of how their experiments change outcomes. We introduce WhatWorkedBench to measure experimental understanding, the accuracy of predictions about component changes after budgeted experimentation. Agents inspect code, select measurements, and submit a response

中文介绍 AI 研究 agent 需要理解实验改动如何影响结果。论文提出 WhatWorkedBench，在有限实验预算下评测 agent 的实验理解能力：检查代码、选择度量并预测组件变更带来的效果，从而衡量其预测准确度。

How To Code Almost Forever for $20/Month With Codex + GPT-6 Luna

@sairahul1 · 147.5K 粉丝 · 785.8K 阅 · 506 赞 · 61 转

09/23 17:41

Yesterday OpenAI released GPT-6 Luna. And it might completely change the economics of AI coding. Here is what most people are missing. $20/month ChatGPT Plus plan with Codex now gives you access to

中文介绍 分享用每月 20 美元的 ChatGPT Plus 搭配 Codex 和刚发布的 GPT-6 Luna 长期写代码的思路，指出多数人忽略的订阅额度红利，讨论 AI 编程的成本结构正被改写。

How to automate SEO with Opus 5.5 (Full Course)

@EXM7777 · 146.5K 粉丝 · 316.1K 阅 · 552 赞 · 33 转

09/23 21:54

I'm going to walk you through the SEO & AEO setup i run with AI agents, from the lazy autopilot version to the full Claude Code build on Opus 5.5, step by step... because search changed shape a while

中文介绍 从懒人自动巡航到基于 Opus 5.5 的 Claude Code 完整搭建，逐步演示自己用 AI agent 跑 SEO 与 AEO 的全套配置，因搜索引擎形态已变，属于体系化教学课程。

I Think I’ve Been Using GPT-6 Astra Wrong

@BradGroux · 7.6K 粉丝 · 174.3K 阅 · 501 赞 · 5 转

09/23 01:49

I think I’ve been using GPT-6 Astra wrong this entire time. I kept giving it the same job I’ve given the strongest model available for the past several years: understand everything, plan everything,

中文介绍 反思自己一直把 GPT-6 Astra 当成最强模型来用，交给它理解一切、规划一切，认为这种用法有误，需要针对新模型特点重新设计任务分工与提示方式。

The $4 Trillion Revolution: How AI Agents Will Rewire Finance and Spark an Economic Big Bang

@joechalom · 18.9K 粉丝 · 105.3K 阅 · 505 赞 · 113 转

09/23 20:00

The convergence of new onchain financial primitives and unlimited agent attention promises unprecedented value creation for consumers. The battle over control has begun. Over $1 trillion of annual

中文介绍 认为链上金融原语与 AI agent 的无限注意力结合将带来空前价值创造，围绕超 1 万亿美元年收入的争夺已经开始，讨论 agent 如何重写金融与经济格局。

How to train your own Jev for $17

@nutlope · 100.4K 粉丝 · 89.8K 阅 · 518 赞 · 35 转

09/24 06:02

TLDR: We just launched our own Jev-like classifier, together/Tev1-4B-experimental, on top of Qwen3.5 4B. In this blog post we’ll show you how to fine-tune your own version! Jev has quickly become one

中文介绍 基于 Qwen3.5 4B 微调出 Jev 式分类器 Tev1-4B-experimental 并发布，博客公开完整流程，教你花约 17 美元训练出自己版本的分类模型。

Incubating The Horowitz Andreessen Academy

@eriktorenberg · 163.4K 粉丝 · 53.5K 阅 · 502 赞 · 55 转

09/22 22:18

The internet made it possible to know anything. AI makes it possible to do anything. We live in an age of infinite knowledge, and we are entering an age of infinite doing. One in which we are

中文介绍 提出孵化 Horowitz Andreessen Academy 的构想：互联网让人知晓一切，AI 让人做到一切，我们正从无限知识时代进入无限执行时代，需要新的教育形态。

Building a Custom Harness with Pi and Jev

@omarsar0 · 321.0K 粉丝 · 45.0K 阅 · 555 赞 · 46 转

09/23 22:09

An AI agent is a language model that works in a loop. It reads the task, uses a tool such as "read this file" or "delete that file", looks at the result, and keeps going until the job is done. Each

中文介绍 从「agent 就是语言模型在循环中调用工具」这一基础定义讲起，演示如何用 Pi 与 Jev 搭建自定义 harness，属于 agent 底层原理与工程实现的拆解。

Gemini 3.8 text-to-speech says hello

@GoogleAIStudio · 205.4K 粉丝 · 41.9K 阅 · 519 赞 · 50 转

09/24 00:20

Today, we’re introducing two new text-to-speech models to the Gemini family, transforming voice generation from static presets into a dynamic creative studio. These models enable creators, developers,

中文介绍 Gemini 家族新增两个文本转语音模型，将语音生成从静态预设升级为动态创作工作室，面向创作者与开发者开放，属官方发布的能力更新。

Introducing Gemini 3.8 Live with Live Avatar

DeepMind Blog · 09/25 00:20

中文介绍 DeepMind 发布 Gemini 3.8 Live，并推出实时虚拟形象（Live Avatar）功能，使模型可在实时交互中呈现动态虚拟人形象。官方博客未披露更多技术细节与开放范围。

Foundries vs Navigators: Lowering the Cost of Science

Latent Space · 09/24 23:03

Guest Post: In science, thinking has gotten cheap but doing has not. This asymmetry is reshaping how research companies operate, largely inconspicuously.

中文介绍 Latent Space 刊出客座文章，指出科学研究中「思考」的成本已大幅下降，而「动手做」的成本依然高昂；这一不对称正悄然重塑研究型公司的运作方式。

Accelerating vision-language models with LFM2.5-VL-DSpark

Hugging Face Blog · 09/24 22:08

中文介绍 Hugging Face 博客介绍 LiquidAI 的 LFM2.5-VL-DSpark，称其可加速视觉语言模型，具体方法与性能数据未在摘要中披露。

[AINews] Meta Connect 2026: Muse glasses, voice, video, and Charm

Latent Space · 09/24 16:12

Team Zuck is absolutely on fire.

中文介绍 Meta 举办 Connect 2026 大会，发布 Muse 眼镜，并展示语音、视频相关能力与 Charm 功能，呈现其在可穿戴设备与多模态交互上的最新进展。

Gemini TTS 🗣️, Claude’s novel enzyme 🧬, Google private memory 🔒

TLDR AI · 09/24 08:00

中文介绍 TLDR AI 2026 年 9 月 24 日简报涵盖三项内容：Gemini 文本转语音、Claude 相关的新型酶研究，以及 Google 推出的私密记忆功能。

How to Use NVIDIA Warp and MjWarp to Accelerate Robotics Simulation and Learning Workflows

Hugging Face Blog · 09/24 02:41

中文介绍 Hugging Face 博客发布使用指南，介绍如何借助 NVIDIA Warp 与 MjWarp 加速机器人仿真与学习工作流，为开发者提供实操方法。

Advancing Private AI Compute with secure, server-side memory

DeepMind Blog · 09/24 00:00

Introducing private, server-side memory to Private AI Compute for personal AI.

中文介绍 DeepMind 为 Private AI Compute 引入安全的服务端内存，以支持个人 AI 场景下的隐私计算，在服务器侧实现私密的内存处理。

Two years of OpenAI Academy

OpenAI News · 09/24 00:00

Marking two years of OpenAI Academy and bringing AI skills to even more communities.

中文介绍 OpenAI 纪念 OpenAI Academy 成立两周年，并表示将继续把 AI 技能培训带给更多社区。

Gemini 3.8 text-to-speech says hello

DeepMind Blog · 09/23 23:25

中文介绍 DeepMind 发布 Gemini 3.8 文本转语音功能，官方博客未披露具体参数、支持语言与开放范围。

🔬Bio-security is an AI Arms Race - Eric Nguyen (CEO, Radical Numerics)

Latent Space · 09/23 21:27

Radical Numerics is using biological chain-of-thought and multimodal perception to keep up with the bio-defense arms race, design new genomes and gain insights into biology itself.

中文介绍 Radical Numerics 首席执行官 Eric Nguyen 表示，生物安全是一场 AI 军备竞赛；该公司正利用生物链式思维与多模态感知，设计新基因组并深化对生物学本身的理解。

OpenAI extends cyber access to Ukraine for civilian defense

OpenAI News · 09/23 21:00

OpenAI is extending access to its Daybreak program to the Government of Ukraine to support the cyber defense of civilian infrastructure.

中文介绍 OpenAI 宣布将 Daybreak 项目的访问权限扩展至乌克兰政府，用于支持民用基础设施的网络防御。

Sam Altman’s remarks at the United Nations Security Council

OpenAI News · 09/23 20:00

OpenAI CEO Sam Altman discusses AI safety, human control, and international cooperation in remarks to the United Nations Security Council.

中文介绍 OpenAI 首席执行官 Sam Altman 在联合国安理会发言，谈及 AI 安全、人类对 AI 的控制以及国际合作等议题。

Harvey turns legal context into stronger drafts with GPT-6 Astra

OpenAI News · 09/23 20:00

GPT-6 Astra produces more structured, context-aware legal documents, freeing lawyers to focus on strategy.

中文介绍 Harvey 借助 GPT-6 Astra 将法律上下文转化为更完善的草稿，生成结构更清晰、更贴合背景的法律文书，使律师得以专注于策略。

How invideo improves color grading 3x with GPT‑6 Astra

OpenAI News · 09/23 20:00

With GPT‑6 Astra, invideo plans edits with greater precision, improves color correction and grading threefold, and produces 50 custom effects in one day.

中文介绍 invideo 借助 GPT-6 Astra 提升剪辑规划的精确度，调色与色彩校正效率提升三倍，并可在一天内产出 50 个自定义特效。

Ringg’s AI agents resolve up to 65% of customer calls with OpenAI

OpenAI News · 09/23 20:00

Using GPT-5.6, Ringg powers multilingual agents across voice, chat, WhatsApp, and web for 90% less cost vs. GPT-4.1.

中文介绍 Ringg 基于 GPT-5.6 构建多语言 AI 客服代理，覆盖语音、聊天、WhatsApp 与网页渠道，可解决最多 65% 的客户来电，成本较 GPT-4.1 降低 90%。