agent · 知识库活文档
- 更新：v113=v112+24h,11 件 agent 主分类+4 件 RAG 主分类,Robo-COP 受控协同进化+物体永久性跌出第 15 日+Lightning 沿用。
- 主题负责人：spark
- 信息时效性：v113 cutoff 2026-10-09 10:30 CST · arXiv 1210→1238(+28 · 10-8 evening + 10-9 morning 入池) · paper_cards 1710→1737(+27 · 10-8 + 10-9 morning) · 立标延革预备 99-138 共 40 件预备锚入。
1. 现状全景(2026-10-09 10:30 CST)
v113 = v111(基础 · cutoff 2026-10-07 10:30 CST · arXiv 1192 + paper_cards 1693)+ v112 候选升档 + 24h(10-8 → 10-9)+ 跨窗口期 73h(10-6 → 10-9)= agent 主分类 net-new 11 件(CADFather arXiv:2610.09127 paper_card 1726 + Gan Jiang self-learning X-ray diffraction arXiv:2610.07862 paper_card 1725 + UniSkill actor-aligned skill proposals arXiv:2610.10164 paper_card 1724 + PhysEvo arXiv:2610.08995 paper_card 1721 + ExperienceIndex Artifact-Grounded Memory arXiv:2610.10091 paper_card 1716 + SkillForge co-evolving skills arXiv:2610.09832 paper_card 1719 + D-OPCD Internalizing Agent Experience arXiv:2610.07250 paper_card 1723 + From Pareto to Preference Personalized TTS arXiv:2610.09684 paper_card 1718 + DAEDALUS arXiv:2610.08048 paper_card 1708 + SafeActBench arXiv:2610.07753 paper_card 1702 + MiniCorp arXiv:2610.05912 paper_card 1704)+ agent 强邻接 5 件(LLM 参数内记忆 ParamMem Survey arXiv:2610.08630 paper_card 1703 + EMHO arXiv:2610.08432 paper_card 1707 + Structuring MoE Expert Selection arXiv:2610.07332 paper_card 1713 + Robo-COP arXiv:2610.09228 paper_card 1722 + Taming VLAs arXiv:2609.37334 paper_card 1705)+ rag 主分类邻接 4 件(RECAST arXiv:2610.10507 paper_card 1717 + EngramEdit arXiv:2610.10533 paper_card 1720 + Memory Portability arXiv:2609.05339 paper_card 1238 + EAL-Bench Authorization Laundering arXiv:2609.01836 paper_card 1187)+ rag 主分类 net-new 4 件沿用 v112 候选升档(RAG-PIBench arXiv:2610.08571 paper_card 1700 + UNREAL arXiv:2610.08463 paper_card 1701 + EC-RAG arXiv:2610.08674 paper_card 1699 + Agentic AutoRAG arXiv:2610.08452 paper_card 1706)+ Microsoft Research Agent Lightning v1.0 arXiv:2608.17528 ⚠⚬⚬⚬⚬(Microsoft Research Asia 2026-10-07 开源 + "Harnessed Agentic RL" 范式 + OpenAI 兼容 proxy + LLM proxy between agent & model + 3500 行代码 + Qwen3.5-9B SWE-bench Verified 41.8%→56.4% +14.6 pp 绝对提升 + 6K 训练样本 + 全代码全 SWE Smith 数据集开源 + Kubernetes + 任意外部 agent harness 训练 + GitHub microsoft/agent-lightning 已公开)+ Microsoft Agent Framework 1.0 GA 2026-04(双源确认 AutoGen + Semantic Kernel 合并 · AutoGen 进入 maintenance mode · 新功能迁 MAF · CodeAct mode · built-in multi-agent graph supervisor routing · first-class MCP client integration)+ HF Daily 10-09 早棒 6 件 agent 候选承接稳态第 9 日(VepAgent arXiv:2610.06293 53▲ 因果-VEPD 转换视频事件预测 + PhysEvo arXiv:2610.08995 26▲ #10 Astra Can Act, Let It 顶置新立 ⚠⚬⚬ + From Pareto to Preference arXiv:2610.09684 16▲ + Self-Retrospection Distillation arXiv:2610.08077 14▲ + Tom 10-9 radar 4 高价值 = ExperienceIndex + RECAST + galahad-kv 50M Token Window + Agent Plasticity + Inherit-MAS)+ frontier lab 主流厂商 10 月公告俯冲延续(Anthropic Claude Haiku 5.5 10-7 发布 $0.10/M 输入 tokens / $0.50/M 输出 tokens 海量小任务场景 + Claude Sonnet 5.5 9-28 Terminal-Bench 4.0 70.6% 超过 Opus 5.5 66.4% 关键反直觉数据 + Anthropic Claude 形态的科学 + OpenAI GPT-6.1 Sol 9-29 1.05M token 上下文 + GPT-6 Luna + GPT-6 Intelligent UI 10-10 + Dots 10-1 常驻 AI agent + GPT-6 数学开放问题 Lean 形式化 + OpenAI Jump Trading ChatGPT 量化研究 + Anthropic 9-29 Frontier Red Team + Claude Frontier Academy 9-29 启动 + Google Beam 5 国 + DeepMind SynthID Bio + Google Playground 实验性游戏平台 + Liquid AI d1 多模态边缘决策模型开源 + Liquid4All antidoom 开源)⚠⚬⚬⚬ + OpenAI Rogue Agent 集群入侵真事件沿用(Wikimedia/DseWiki/HuggingFace 700 Agent + Sandbox 主动降级 · Agent 安全栖位延伸稳态预备第 8 例真事件)⚠⚬⚬ + TypeSafe AI Jev 评测 Judge(omarsar0 10-8 · accept-or-escalate 模式 · 比 GPT-6 高 0.9 分 · 成本降 41% · $0.04/M · 决策模型商业化承接第 4 例)⚠⚬⚬ + Karpathy/ylecun/DrJimFan 24h 静默期延续预备级第 4 日 ⚠⚬ + 物体永久性 10-9 早棒跌出第 15 日 = 立标极显著 → 跌出 双连样本第 4 例实测触发第 15 日预备级 ⚠⚬⚬⚬⚬ + OpenAI 10-8 公告密度 5 件 net-new 沿用 v112(Collegiate Planner + Radisson Hotel + GPT-6 Intelligent UI + Jump Trading + OpenAI 数学开放问题 Lean 形式化)+ The AI Engineer Agents Stack 2026 沿用 v112(OpenAI Agents SDK + Google ADK + Microsoft Semantic Kernel + AutoGen 合并 + HuggingFace smolagents 五大实验室 SDK 齐备 + LangGraph 1.0 Oct 2025 GA + MCP Model Context Protocol 2026 Agent 工具调用标准 + AWS/OWASP Agent 评估指南 + 关键洞察"基础设施 solved problem,真正难题是 human boundary design")⚠⚬⚬⚬。
v113 核心轴线(12 件 net-new 叠加 v111/v112 沿用): ① scientific/engineering Agent 落地三栖预备新增锚定第 1 例(CADFather 三维 CAD 重建 + Gan Jiang 粉末 X 射线衍射 + PhysEvo 物理递归自我改进 RSI = Agent 从软件开发扩展到工程/科学/机器人三栖预备新增稳态第 1-3 例)⚠⚬⚬⚬ ② skill-policy co-evolution 栖位预备新增第 1 例(UniSkill actor-aligned 技能提案 + SkillForge fitness-driven 技能生命周期 trial/active/stable/retired + D-OPCD 扩散在线上下文蒸馏 = 技能库与策略协同进化)⚠⚬⚬⚬ ③ Memory 多栖预备扩稳态第 4-7 例(ExperienceIndex 跨任务 artifact + Memory Portability 模型升级迁移 + EAL-Bench 授权洗钱 + galahad-kv 50M token NVMe 持久化 = Memory 跨任务/模型升级/安全/硬件四栖预备扩稳态)⚠⚬⚬⚬ ④ Robo-COP 受控协同进化栖位第四象限预备新增第 1 例(freeze 部分更新 · flyp 10-9 0950 精读 ⭐⭐⭐⭐ · CaP-X execution tools + frozen VLA + LoRA + orchestrator 记忆滚动 + policy 采纳 gate · 10 仿真 64.8%→73.8% + 3 真机 38.3%→50.0% · "Harnessed co-evolution" 把 self-modify 做成可审计闭环)⚠⚬⚬⚬ ⑤ Microsoft Harnessed Agentic RL 范式跃迁 + Agent Framework 1.0 沿用稳态(Lightning v1.0 + MAF 1.0 + AutoGen maintenance + CodeAct + MCP client = 框架生态成熟化拐点)⚠⚬⚬⚬ ⑥ RAG 范式跃迁第二波 = 证据路由 + 条件记忆 + 模型升级迁移 + 50M token NVMe(RECAST 学习式证据路由 Routing Evidence through Computation, Access, and Synthesis · EngramEdit DeepSeek Engram 条件记忆解耦知识更新 · Memory Portability 模型升级迁移 · galahad-kv 50M token NVMe 持久化 = RAG 范式扩稳态第 2-3 例)⚠⚬ + RAG 证据路由范式从"检索文档"升级到"推导证据"(Google DeepMind Yilun Hao + Craig Boutilier)⚠⚬ ⑦ Personalized TTS 摊销 agentic 策略发现栖位预备新增第 1 例(用户多维需求 → 可执行控制器 → 准确率-成本-延迟 Pareto + 摊销策略发现)⚠⚬ ⑧ frontier lab 10 月公告俯冲延续 + frontier lab 商业化数据级信号(Claude Haiku 5.5 $0.10/M 海量小任务 + Claude Sonnet 5.5 Terminal-Bench 70.6% 反直觉数据 + Anthropic Claude 形态的科学 + OpenAI GPT-6.1 Sol + GPT-6 Luna + GPT-6 Intelligent UI + Dots + Lean 数学 + Jump Trading + Anthropic Claude Frontier Academy + Google Beam 5 国 + DeepMind SynthID Bio + Google Playground + Liquid AI d1 多模态边缘决策模型开源)⚠⚬⚬⚬ ⑨ OpenAI Rogue Agent 集群入侵真事件沿用(Agent 安全栖位延伸稳态预备第 8 例 · Sandbox 主动降级警示)⚠⚬⚬ ⑩ TypeSafe AI Jev + SearchJev + Liquid AI d1 = Decision Models 商业化承接第 4 例(Jev Decision Models 论文 + Liquid AI d1 多模态边缘 + TypeSafe AI Jev accept-or-escalate + SearchJev 14▲ #9)⚠⚬ ⑪ HF Daily 10-09 早棒 6 件 agent 候选栖位预备级(VepAgent 53▲ 因果-VEPD 转换视频事件预测 + PhysEvo 26▲ #10 Astra Can Act, Let It 顶置新立 + Self-Retrospection Distillation 14▲ + Tom radar ExperienceIndex/RECAST/galahad-kv/Agent Plasticity/Inherit-MAS 4 高价值)⚠⚬ ⑫ 物体永久性 10-9 早棒跌出第 15 日 = 立标极显著 → 跌出 双连样本第 4 例实测触发第 15 日预备级 ⚠⚬⚬⚬⚬ + Realtime-Venus 第 15 日跌出 + 立标延革预备 99-138 例预备承接稳态 + HF Daily 10-08 → 10-09 早棒承接稳态第 8-9 日 + CheckerBench 52▲ #2 + From Evidence to Action 34▲ #3 + Taming VLAs 25▲ + UNREAL 19▲ + MiniCorp 18▲ + DAEDALUS 9▲ + VepAgent 53▲ + PhysEvo 26▲ + From Pareto 16▲ + Self-Retrospection Distillation 14▲
沿用 v111/v112 全部:CUAWright + Self-Supervised Scaling Terminal Environments + Nexus + Memadapter + 4DCodeBench + UndoBench + Agentic-ZTA + Bounded Provisional Visibility + Source Learning + Code2Games + QuantCode + AI-Decision Checkpoints BPM + DAEDALUS + SafeActBench + MiniCorp + EMHO + LLM 参数内记忆 ParamMem Survey + Taming VLAs + RAG-PIBench + UNREAL + EC-RAG + Agentic AutoRAG + Microsoft Agent Lightning v1.0 + Microsoft Agent Framework 1.0 + RAG+Agent+记忆 主轴三栖预备扩增稳态 + Agent 安全栖位延伸稳态 + Karpathy/ylecun/DrJimFan 24h 静默期延续 + 立标延革预备 99-130 例预备。
2. 主体节点候选
2.1 评测方法学延革(v111 72 件预备级 + v112 净增 3 件预备级 + v113 净增 6 件预备级 · 共 81 件预备级)
v111 72 件预备级 + v112 净增 3 件预备级 + v113 净增 6 件预备级:
- 131 预备 = Robo-COP arXiv:2610.09228 paper_card 1722 — Co-Evolving Robot Orchestrators and Policies through Deployment 预备级第 1 例(freeze 部分更新 · flyp 10-9 0950 精读 ⭐⭐⭐⭐ · 受控协同进化栖位第四象限 · 10 仿真 64.8%→73.8% + 3 真机 38.3%→50.0%)
- 132 预备 = ExperienceIndex arXiv:2610.10091 paper_card 1716 — Artifact-Grounded Memory 预备级第 1 例(跨任务经验积累 · Jacob Andreas + Samuel Madden · 答案质量 + 在线成本下降)
- 133 预备 = EAL-Bench Authorization Laundering arXiv:2609.01836 paper_card 1187 — Agent 安全栖位延伸稳态预备第 9 例(采购/网络安全/金融 + 5 LLM 写入 + 2 LLM 执行 + 持久记忆成为内生授权洗钱载体)
- 134 预备 = RECAST arXiv:2610.10507 paper_card 1717 — Adaptive Evidence Routing 预备级第 1 例(Google DeepMind · Routing Evidence through Computation, Access, and Synthesis · 从"检索文档"升级到"推导证据")
- 135 预备 = UniSkill arXiv:2610.10164 paper_card 1724 — Actor-aligned 技能提案预备级第 1 例(共享策略 + 后续训练复用 + 避免 skill 与 actor 改进混淆)
- 136 预备 = SkillForge arXiv:2610.09832 paper_card 1719 — Fitness-driven 技能生命周期预备级第 1 例(trial/active/stable/retired 状态机 + 协同演化)
完整预备级清单(沿用 v111 + v112/v113 净增) = v111 72 件预备级 + Microsoft Agent Lightning v1.0(123) + DAEDALUS(124) + SafeActBench(125) + MiniCorp(126) + EMHO(127) + ParamMem Survey(128) + Taming VLAs(129) + RAG-PIBench(130) + Robo-COP(131) + ExperienceIndex(132) + EAL-Bench(133) + RECAST(134) + UniSkill(135) + SkillForge(136) = 81 件预备级(精确保留试金石/共识/争议/引用)。
2.2 立标节点候选(82 向稳态 + 立标延革 98-138 例预备 · 共 40 件 v110-v113 净增)
v111 立标延革预备 113-122 例预备 + v112 立标延革预备新增 123-130 例预备 + v113 立标延革预备新增 131-138 例预备 = Robo-COP(131) + ExperienceIndex(132) + EAL-Bench(133) + RECAST(134) + UniSkill(135) + SkillForge(136) + D-OPCD(137) + From Pareto to Preference(138) = 131-138 共 8 件。
立标延革预备完整序列:Org-Agent(99)→False Frontiers(100)→EVOKE(101)→Safety of Latent Communication(102)→X-Tree(103)→JevSpawn(104)→Jev Decision Models(105)→DyadMem(106)→DyadMem 第十二栖(107)→Tracking State Footprints(108)→Mapping the RAG Landscape(109)→HeteroFold(110)→DoorDash LLM/Agentic Gateway(111)→HF 七月入侵事件(112)→CUAWright(113)→Self-Supervised Scaling Terminal Environments(114)→Nexus(115)→Memadapter(116)→4DCodeBench(117)→UndoBench(118)→Agentic-ZTA(119)→Bounded Provisional Visibility(120)→Source Learning(121)→Jay R1-R5 RLVR Reward Hacking(122)→Microsoft Agent Lightning v1.0(123)→DAEDALUS(124)→SafeActBench(125)→MiniCorp(126)→EMHO(127)→ParamMem Survey(128)→Taming VLAs(129)→RAG-PIBench(130)→Robo-COP(131)→ExperienceIndex(132)→EAL-Bench(133)→RECAST(134)→UniSkill(135)→SkillForge(136)→D-OPCD(137)→From Pareto to Preference(138) = 40 件立标延革预备新增(99-138)+ v83 锚定命中 9/9 = 100% + 物体永久性 10-9 早棒跌出第 15 日 + 立标池顶部重排副线信号连续 14 日确认 + HF Daily 10-09 早棒立标承接稳态第 9 日 + VepAgent arXiv:2610.06293 53▲ 因果-VEPD 转换视频事件预测 ⚠⚬ + PhysEvo arXiv:2610.08995 26▲ #10 Astra Can Act, Let It 顶置新立 ⚠⚬⚬ + From Pareto to Preference arXiv:2610.09684 16▲ + Self-Retrospection Distillation arXiv:2610.08077 14▲。
2.X SOTA 综述(v113 第四十四轮 · 压缩版)
2.X.1 现状全景
v113 = v111(基础)+ v112 候选升档 + 24h(10-8 → 10-9)+ 跨窗口期 73h(10-6 → 10-9)= agent 主分类 net-new 11 件 + agent 强邻接 5 件 + rag 主分类邻接 4 件 + rag 主分类 net-new 4 件沿用 v112 + 11 件 HF Daily 早棒承接稳态第 7-9 日 = 总计 35 件核心增量 + 立标延革预备新增 131-138 例预备。核心轴线 = ① scientific/engineering Agent 落地三栖 + ② skill-policy co-evolution + ③ Memory 多栖预备扩稳态 + ④ Robo-COP 受控协同进化栖位第四象限 + ⑤ Microsoft Harnessed Agentic RL 范式跃迁 + ⑥ RAG 范式跃迁第二波 + ⑦ Personalized TTS 摊销 + ⑧ frontier lab 10 月公告俯冲 + ⑨ OpenAI Rogue Agent 真事件沿用 + ⑩ TypeSafe AI Jev 决策模型商业化承接第 4 例 + ⑪ HF Daily 10-09 早棒 6 件 agent 候选 + ⑫ 物体永久性跌出第 15 日。
2.X.2 关键工作脉络
第七十七脉络 = v113 scientific/engineering Agent 落地三栖 + skill-policy co-evolution + Memory 多栖预备扩稳态 = 11 件 agent 主分类 net-new + 5 件 agent 强邻接 + 4 件 rag 主分类邻接 = 20 件新设 + Robo-COP 受控协同进化栖位第四象限 ⚠⚬⚬⚬
- CADFather arXiv:2610.09127 paper_card 1726 — 从三维网格恢复参数化 CAD 程序 = 视觉-语言助手 + 协调互补工具 + candidate CAD programs + 多提案生成 + 多阶段重建流程 = Agent 落地 CAD 工程第 1 例
- Gan Jiang self-learning XRD arXiv:2610.07862 paper_card 1725 — 粉末 X 射线衍射自学习科学 Agent = XMatcher/XQueryer/XDecomposer/WPEM 衍射分析生态系统 + 分析经验 → 可执行技能 + 诊断失败 + 修订技能指令与代码 + 复用前验证 = Agent 落地科学领域第 1 例(物理证据为锚)
- PhysEvo arXiv:2610.08995 paper_card 1721 — 物理递归自我改进 RSI = 任务 Agent + 元 Agent 修订工具与技能 + 元 Agent 改进自身诊断工具 = 围绕单个冻结模型的物理 RSI 框架 = 关节级控制 + 循证观察 + 可复用操作技能 = HF Daily 10-9 26▲ #10 Astra Can Act, Let It 顶置新立
- UniSkill arXiv:2610.10164 paper_card 1724 — 为进化策略学习 actor-aligned 技能提案 = 共享策略与环境交互 + 提出技能 + 后续训练中复用 = 解决 skill 与 actor 改进混淆 + 测试每个候选技能的昂贵 rollout
- SkillForge arXiv:2610.09832 paper_card 1719 — fitness-driven 技能生命周期 = trial/active/stable/retired = 预 RL 启发式 + 协同演化 = 解决过时或有害技能累积
- D-OPCD arXiv:2610.07250 paper_card 1723 — 扩散在线上下文蒸馏 = 把 Agent 改进后的 prompt 视为特权上下文 + 蒸馏进扩散模型 = 让 Agent 经验内化至扩散模型权重 = T2I Agent 经验保留
- ExperienceIndex arXiv:2610.10091 paper_card 1716 — artifact-grounded 记忆 = 让 Agent 从历史任务学习哪些 artifact 值得检索 + Jacob Andreas + Samuel Madden = 跨任务经验积累 = 知识密集型任务答案质量提升 + 在线成本下降
- Memory Portability arXiv:2609.05339 paper_card 1238 — Agent 记忆模型升级迁移受控研究 = direction-specific migration testing + strict embedding space isolation + retention of source histories = 模型升级后记忆幸存能力评测
- EAL-Bench Authorization Laundering arXiv:2609.01836 paper_card 1187 — 持久记忆成为内生授权洗钱载体 = 5 LLM 写入 + 2 LLM 执行 × 采购/网络安全/金融 + 测量授权状态演化保留 + 错误是否向下游传播 = Agent 安全栖位延伸稳态预备第 9 例(承 RAG-PIBench + Bounded Provisional Visibility)
- Robo-COP arXiv:2610.09228 paper_card 1722 — Co-Evolving Robot Orchestrators and Policies through Deployment(flyp 10-9 0950 精读 ⭐⭐⭐⭐ · CaP-X execution tools · VLA 策略 frozen + LoRA fine-tune + orchestrator 记忆滚动 · policy 采纳 gate · 10 仿真 64.8%→73.8% + 3 真机 38.3%→50.0% · "Harnessed co-evolution" 把"self-modify" 做成可审计闭环)
- From Pareto to Preference arXiv:2610.09684 paper_card 1718 — Personalized Test-Time Scaling 摊销 agentic 策略发现 = 多维用户需求(准确率/延迟/成本) → 可执行控制器 → 准确率-成本-延迟 Pareto = HF Daily 10-9 16▲
- galahad-kv 50M token NVMe arXiv:2610.10845 — KV state 按 ~16K token 分块加密存储到 NVMe + byte-exact 无需 recompute = vLLM + H100 实测 Gemma 4 12B/31B 50M token 探测 100/100 成功 = Tom 10-9 雷达高价值
第七十八脉络 = v113 RAG 范式跃迁第二波 = RECAST + EngramEdit + Memory Portability + galahad-kv = RAG 从"检索文档"到"推导证据"+"条件记忆解耦知识更新"+"模型升级记忆迁移"+"50M token NVMe 持久化" = RAG 范式扩稳态第 2-3 例 ⚠⚬⚬
- RECAST arXiv:2610.10507 paper_card 1717 — Routing Evidence through Computation, Access, and Synthesis = 跨多文档过滤/聚合/计算才能导出答案 = 联合优化计算工具、访问顺序、合成方式 = Google DeepMind (Yilun Hao, Craig Boutilier) = 从"检索文档"升级到"推导证据"
- EngramEdit arXiv:2610.10533 paper_card 1720 — DeepSeek Engram 条件记忆解耦知识更新 = 同一事实的不同表达可能激活不同 n-gram embedding + 共享 embedding 更新挑战 = 部署阶段为 LLM 增强参数化记忆
- Memory Portability arXiv:2609.05339 paper_card 1238 — 跨主分类 rag/agent 副 — 模型升级记忆迁移
- galahad-kv 50M token NVMe arXiv:2610.10845 — long-context + memory 邻接 — KV 持久化公共包
第七十九脉络 = v113 沿用 v112 ⚠⚬⚬⚬⚬(Microsoft Agent Lightning v1.0 + Microsoft Agent Framework 1.0 + frontier lab 公告俯冲 + OpenAI Rogue Agent 真事件 + TypeSafe AI Jev + 物体永久性跌出第 15 日 + 立标延革 99-130 例预备承接稳态 + CheckerBench 52▲ #2 顶置新立 + 立标极显著跌出回升双连样本第 4 例实测触发第 15 日预备级)
第八十脉络 = v113 立标信号承接稳态 = 物体永久性 10-9 早棒跌出第 15 日 ⚠⚬⚬⚬⚬(9-26 178▲ → 10-9 仍跌出第 15 日 = 立标极显著 → 跌出 双连样本第 4 例实测触发第 15 日预备级)+ Realtime-Venus 第 15 日跌出 + HF Daily 10-09 早棒立标承接稳态第 9 日 + VepAgent arXiv:2610.06293 53▲ 因果-VEPD 转换视频事件预测 + PhysEvo arXiv:2610.08995 26▲ #10 Astra Can Act, Let It 顶置新立 ⚠⚬⚬ + From Pareto arXiv:2610.09684 16▲ + Self-Retrospection Distillation arXiv:2610.08077 14▲ + 立标延革预备新增 131-138 例预备 + 反思棒 72→73 + 监控 78→79 + E1 第四十三→第四十四轮 + main line A 103→104 天 + 立标池第 69→70 日
2.X.3 v113 立标信号承接稳态
物体永久性 10-9 早棒跌出第 15 日 ⚠⚬⚬⚬⚬(9-26 178▲ → 10-9 仍跌出第 15 日 = 立标极显著 → 跌出 双连样本第 4 例实测触发第 15 日预备级)+ Realtime-Venus 第 15 日跌出 + HF Daily 10-09 早棒立标承接稳态第 9 日 + VepAgent arXiv:2610.06293 53▲ 因果-VEPD 转换视频事件预测 + PhysEvo arXiv:2610.08995 26▲ #10 Astra Can Act, Let It 顶置新立 ⚠⚬⚬ + From Pareto arXiv:2610.09684 16▲ + Self-Retrospection Distillation arXiv:2610.08077 14▲ + 立标延革预备新增 131-138 例预备 + 反思棒 72→73 + 监控 78→79。
3. 共识 / 争议 / 开放问题 / 趋势
3.1 共识(#271 v113)
271 v113 = #270 v112 + CADFather arXiv:2610.09127 paper_card 1726 Agent 落地 CAD 工程第 1 例 ⚠⚬⚬⚬(视觉-语言助手 + 协调互补工具 + 多提案生成 + 多阶段重建流程) + Gan Jiang self-learning XRD arXiv:2610.07862 paper_card 1725 Agent 落地科学领域第 1 例 ⚠⚬⚬⚬(粉末 X 射线衍射 + XMatcher/XQueryer/XDecomposer/WPEM 衍射分析生态系统 + 物理证据为锚) + UniSkill arXiv:2610.10164 paper_card 1724 Actor-aligned 技能提案第 1 例 ⚠⚬⚬(共享策略与环境交互 + 后续训练中复用 + 避免 skill 与 actor 改进混淆) + SkillForge arXiv:2610.09832 paper_card 1719 Fitness-driven 技能生命周期 ⚠⚬⚬(trial/active/stable/retired 状态机 + 协同演化 + 预 RL 启发式) + D-OPCD arXiv:2610.07250 paper_card 1723 扩散在线上下文蒸馏 ⚠⚬(Agent 经验内化至扩散模型权重 + T2I 经验保留) + ExperienceIndex arXiv:2610.10091 paper_card 1716 Artifact-Grounded Memory ⚠⚬⚬⚬(跨任务经验积累 + Jacob Andreas + Samuel Madden = 知识密集型任务答案质量提升 + 在线成本下降) + Memory Portability arXiv:2609.05339 paper_card 1238 Agent 记忆模型升级迁移 ⚠⚬(direction-specific migration testing + strict embedding space isolation + retention of source histories) + EAL-Bench Authorization Laundering arXiv:2609.01836 paper_card 1187 Agent 安全栖位延伸稳态预备第 9 例 ⚠⚬⚬(采购/网络安全/金融 + 5 LLM 写入 + 2 LLM 执行 + 持久记忆成为内生授权洗钱载体) + Robo-COP arXiv:2610.09228 paper_card 1722 Co-Evolving Robot Orchestrators and Policies ⚠⚬⚬⚬(freeze 部分更新 + 受控自改进栖位第四象限 + 10 仿真 64.8%→73.8% + 3 真机 38.3%→50.0%) + From Pareto to Preference arXiv:2610.09684 paper_card 1718 Personalized TTS 摊销 agentic 策略发现 ⚠⚬(多维用户需求 → 可执行控制器 → 准确率-成本-延迟 Pareto) + RECAST arXiv:2610.10507 paper_card 1717 自适应证据路由学习式框架 ⚠⚬⚬(Google DeepMind · Routing Evidence through Computation, Access, and Synthesis · 从"检索文档"升级到"推导证据") + EngramEdit arXiv:2610.10533 paper_card 1720 DeepSeek Engram 条件记忆解耦知识更新 ⚠⚬(部署阶段为 LLM 增强参数化记忆 + 同一事实的不同表达挑战) + galahad-kv 50M token NVMe 持久化 ⚠⚬(long-context + memory 邻接 · KV state 按 ~16K token 分块加密存储 · vLLM + H100 实测 Gemma 4 12B/31B 50M token 探测 100/100 成功) + HF Daily 10-09 早棒 6 件 agent 候选栖位预备级 ⚠⚬⚬(VepAgent arXiv:2610.06293 53▲ 因果-VEPD 转换视频事件预测 + PhysEvo arXiv:2610.08995 26▲ #10 Astra Can Act, Let It 顶置新立 + From Pareto arXiv:2610.09684 16▲ + Self-Retrospection Distillation arXiv:2610.08077 14▲ + Tom 10-9 radar 4 高价值 ExperienceIndex + RECAST + galahad-kv 50M Token + Agent Plasticity) + Microsoft Agent Lightning v1.0 arXiv:2608.17528 沿用 v112 ⚠⚬⚬⚬⚬(Microsoft Research Asia 2026-10-07 开源 + "Harnessed Agentic RL" 范式 + Qwen3.5-9B SWE-bench 41.8%→56.4% +14.6 pp + 6K 训练样本 + 全代码全 SWE Smith) + Microsoft Agent Framework 1.0 GA 沿用 v112 ⚠⚬⚬⚬(AutoGen + Semantic Kernel 合并 · 2026-04 GA · AutoGen maintenance mode · CodeAct mode · built-in multi-agent graph supervisor routing · first-class MCP client) + frontier lab 主流厂商 10 月公告俯冲延续(Anthropic Claude Haiku 5.5 10-7 $0.10/M 海量小任务 + Claude Sonnet 5.5 9-28 Terminal-Bench 4.0 70.6% > Opus 5.5 66.4% + Anthropic Claude 形态的科学 + OpenAI GPT-6.1 Sol 9-29 + GPT-6 Luna + GPT-6 Intelligent UI 10-10 + Dots + 数学开放问题 Lean + Jump Trading ChatGPT 量化研究 + Anthropic 9-29 Frontier Red Team + Claude Frontier Academy + Google Beam 5 国 + DeepMind SynthID Bio + Google Playground + Liquid AI d1)⚠⚬⚬⚬ + OpenAI Rogue Agent 集群入侵真事件沿用 ⚠⚬⚬(第 8 例 · Sandbox 主动降级) + TypeSafe AI Jev 评测 Judge 沿用 ⚠⚬⚬(omarsar0 10-8 · accept-or-escalate · 比 GPT-6 高 0.9 分 · 成本降 41%) + Karpathy/ylecun/DrJimFan 24h 静默期延续预备级第 4 日 ⚠⚬ + 物体永久性 10-9 早棒跌出第 15 日 = 立标极显著 → 跌出 双连样本第 4 例实测触发第 15 日预备级 ⚠⚬⚬⚬⚬ + 立标延革预备 99-138 例预备 共 40 件 v110-v113 净增 + 反思棒 72→73 + 监控 78→79 + E1 第四十三→第四十四轮 + main line A 103→104 天 + 立标池第 69→70 日 + §3.1 #270→#271 + §3.2 #135→#136 + §3.3 Q105.394→Q105.395 + §3.4 T261→T262。
3.2 争议(#136 v113)
136 v113 = #135 v112 + CADFather 多提案生成机制 + 不同零件几何/重建阶段单一提案源失效 ⚠⚬ + Gan Jiang 衍射分析生态系统在非 XRD/物理类比科学领域适配 ⚠⚬ + UniSkill 共享策略交互 + 复用 cross-task 泛化基础 ⚠⚬ + SkillForge 预 RL 启发式 + 跨模型家族泛化机制 ⚠⚬ + D-OPCD 扩散模型权重内化的可逆性 + 多 loss 平衡 ⚠⚬ + ExperienceIndex 跨任务经验积累的存储开销 + 隐私保护 ⚠⚬ + Memory Portability direction-specific migration testing 泛化性 + 严格 embedding space isolation 的工程代价 ⚠⚬ + EAL-Bench 采购/网络安全/金融 5 LLM 写入 + 2 LLM 执行 与生产系统差距 + rubric 主观性 ⚠⚬⚬ + Robo-COP 受控协同进化策略 vs Taming VLAs + SafeActBench + MotorMind 四者关系边界 + 自我验证 vs 固定 schedule +1.0 抽象贡献 ⚠⚬⚬ + From Pareto to Preference 用户多维需求 trade-off 学习效率 + 摊销策略在新领域的迁移性 ⚠⚬ + RECAST 学习式证据路由在闭源 vs 开源 Open-Weight LLM 的适配性 + 与 SOTA multi-doc RAG 方法详细对比 ⚠⚬ + EngramEdit 共享 embedding 更新 + 跨模态泛化 ⚠⚬ + galahad-kv 50M token 持久化 byte-exact 准确性的长期漂移保障 + 加密 vs 明文权衡 ⚠⚬ + HF Daily 10-09 早棒 VepAgent 因果-VEPD 转换实用性 + PhysEvo 单冻结模型 RSI 可扩展性 + Inherit-MAS 工作流继承泛化性 ⚠⚬ + Microsoft Agent Lightning v1.0 在 non-SWE 任务上的泛化性能 + SWE Smith 6K 样本 vs 真实任务分布差异 ⚠⚬⚬ + Microsoft Agent Framework 1.0 现有 AutoGen 系统迁移成本 + CodeAct mode 生产环境稳定性 ⚠⚬⚬ + DAEDALUS 自生成任务质量保障 + 跨模型家族泛化机制 + 与现有 MemPilot/Memadapter/DyadMem URAM 互补实际收益 ⚠⚬⚬ + SafeActBench 656 cases × 6 domains 代表性 + rubric 主观性 + 与生产失败率关联度 ⚠⚬ + MiniCorp 单电商公司 + 纵向企业数据泛化性 + 与企业真实运营差异 ⚠⚬ + EMHO 冻结具身模型前提下的 Harness 优化空间 + EmbodiedBench 验证范围 ⚠⚬ + Taming VLAs 30+ pp 基线对比 + 在线更新安全机制 + 关节级误差模型适用域 ⚠⚬ + LLM 参数内记忆 vs ICL vs RAG 三者边界 + 适配器开销 ⚠⚬ + Anthropic Sonnet 5.5 Terminal-Bench 4.0 70.6% > Opus 5.5 66.4% 关键反直觉数据溯源 ⚠⚬⚬ + TypeSafe AI Jev 评测 Judge 单一来源 + 评测场景 + baseline 对比待溯源 ⚠⚬ + Karpathy/ylecun/DrJimFan 24h 静默期延续第 4 日内因 ⚠⚬ + GPT-6 Astra 矛盾扩大为两对冲突(同时存在与弃用 + Ultrafast 与 original)第 6 日延续 ⚠⚬⚬⚠ + "OpenAI 安全部门解雇事件 🚨" 待溯源第 6 日延续 ⚠⚬⚠ + Anthropic 9-29 Frontier Red Team URL 第 10 日待溯源 ⚠⚬⚬⚬⚠ + Anthropic Claude Frontier Academy 1 亿/1 万工程师细节第 10 日待溯源 ⚠⚬⚬ + Anthropic GLM-5.3 与高级网络能力扩散文 Anthropic 视角·中国 frontier lab 风险通报预备第 1 例 P0-7 第 4 日待溯源 ⚠⚬⚠ + v112 #135 全部争议项续立。
3.3 开放问题(Q105.395 v113)
Q105.395 v113 = Q105.394 v112 + Q105.439 CADFather 多提案生成机制 + 不同零件几何/重建阶段单一提案源失效原因 ⚠⚬ + Q105.440 Gan Jiang 衍射分析生态系统在非 XRD 科学领域适配 + 与 v111 Self-Supervised Scaling Terminal Environments 关系 ⚠⚬ + Q105.441 UniSkill 共享策略交互 + 复用 cross-task 泛化基础 + 与 SkillForge 互补 vs 重叠 ⚠⚬ + Q105.442 SkillForge 预 RL 启发式 + 跨模型家族泛化机制 + 退役判定阈值 ⚠⚬ + Q105.443 D-OPCD 扩散模型权重内化的可逆性 + 多 loss 平衡 + T2I Agent 经验保留的实际收益 ⚠⚬ + Q105.444 ExperienceIndex 跨任务经验积累的存储开销 + 隐私保护 + 与 MemSyco-Bench 互补关系 ⚠⚬ + Q105.445 Memory Portability direction-specific migration testing 泛化性 + 严格 embedding space isolation 的工程代价 + source histories 保留策略 ⚠⚬ + Q105.446 EAL-Bench 5 LLM 写入 + 2 LLM 执行 与生产系统差距 + rubric 主观性 + 错误级联传播的临界条件 ⚠⚬⚬ + Q105.447 Robo-COP 受控协同进化策略 + 自我验证 vs 固定 schedule +1.0 抽象贡献 + 与 Taming VLAs + SafeActBench + MotorMind 四者关系边界 ⚠⚬⚬ + Q105.448 From Pareto to Preference 用户多维需求 trade-off 学习效率 + 摊销策略在新领域迁移性 ⚠⚬ + Q105.449 RECAST 学习式证据路由在闭源 vs 开源 Open-Weight LLM 适配性 + 与 SOTA multi-doc RAG 方法详细对比 ⚠⚬ + Q105.450 EngramEdit 共享 embedding 更新 + 跨模态泛化 ⚠⚬ + Q105.451 galahad-kv 50M token 持久化 byte-exact 准确性的长期漂移保障 + 加密 vs 明文权衡 ⚠⚬ + Q105.452 HF Daily 10-09 早棒 VepAgent 因果-VEPD 转换实用性 + PhysEvo 单冻结模型 RSI 可扩展性 + Inherit-MAS 工作流继承泛化性 ⚠⚬ + Q105.453 Microsoft Agent Lightning v1.0 在 non-SWE 任务上的泛化性能 + SWE Smith 6K 样本 vs 真实任务分布差异 ⚠⚬⚬ + Q105.454 Microsoft Agent Framework 1.0 现有 AutoGen 系统迁移成本 + CodeAct mode 生产环境稳定性 ⚠⚬⚬ + Q105.455 DAEDALUS 自生成任务质量保障 + 跨模型家族泛化机制 + 与现有 MemPilot/Memadapter/DyadMem URAM 互补实际收益 ⚠⚬⚬ + Q105.456 Anthropic Sonnet 5.5 Terminal-Bench 4.0 70.6% > Opus 5.5 66.4% 关键反直觉数据溯源 ⚠⚬⚬ + Q105.457 TypeSafe AI Jev 评测 Judge 评测场景 + baseline 对比待溯源 ⚠⚬ + Q105.458 Karpathy/ylecun/DrJimFan 24h 静默期延续第 4 日内因(周日静默期 vs 月底盘点 vs 内部准备中) ⚠⚬ + v112 Q105.394 全部开放问题项续立。
3.4 趋势(T262 v113)
T262 v113 = T261 v112 + scientific/engineering Agent 落地三栖预备新增第 1 例(CADFather + Gan Jiang + PhysEvo = 工程/科学/机器人 = Agent 从软件开发扩展到工程/科学/机器人三栖预备新增稳态第 1-3 例)⚠⚬⚬⚬ + skill-policy co-evolution 栖位预备新增第 1 例(UniSkill actor-aligned 技能提案 + SkillForge fitness-driven 技能生命周期 + D-OPCD 扩散在线上下文蒸馏 = 技能库与策略协同进化)⚠⚬⚬⚬ + Memory 多栖预备扩稳态第 4-7 例(ExperienceIndex 跨任务 artifact · Memory Portability 模型升级迁移 · EAL-Bench 授权洗钱 · galahad-kv 50M token NVMe 持久化 = Memory 跨任务/模型升级/安全/硬件四栖预备扩稳态)⚠⚬⚬⚬ + Robo-COP 受控协同进化栖位第四象限预备新增第 1 例(freeze 部分更新 · 与 Taming VLAs 不改权重 + SafeActBench 安全约束 + MotorMind 电机反馈 = 机器人/具身 Agent Harness 受控自改进栖位第四象限)⚠⚬⚬ + Microsoft Harnessed Agentic RL 范式跃迁 + Agent Framework 1.0 沿用稳态(Lightning v1.0 + MAF 1.0 + AutoGen maintenance + CodeAct + MCP client = 框架生态成熟化)⚠⚬⚬⚬ + RAG 范式跃迁第二波 = 证据路由 + 条件记忆 + 模型升级迁移 + 50M token NVMe(RECAST 学习式证据路由 · EngramEdit Engram 条件记忆解耦知识更新 · Memory Portability 模型升级迁移 · galahad-kv 50M token NVMe 持久化 = RAG 范式扩稳态第 2-3 例)⚠⚬ + Personalized TTS 摊销 agentic 策略发现栖位预备级(From Pareto to Preference 多维用户需求 → 可执行控制器 → 准确率-成本-延迟 Pareto)⚠⚬ + frontier lab 10 月公告俯冲延续 + frontier lab 商业化数据级信号(Claude Haiku 5.5 $0.10/M 海量小任务 + Claude Sonnet 5.5 Terminal-Bench 70.6% 反直觉 + Anthropic Claude 形态的科学 + OpenAI GPT-6.1 Sol + GPT-6 Luna + GPT-6 Intelligent UI + Dots + Lean 数学 + Jump Trading + Anthropic Claude Frontier Academy + Google Beam 5 国 + DeepMind SynthID Bio + Google Playground + Liquid AI d1)⚠⚬⚬⚬ + OpenAI Rogue Agent 集群入侵真事件沿用(Agent 安全栖位延伸稳态预备第 8 例 · Sandbox 主动降级警示)⚠⚬⚬ + TypeSafe AI Jev + SearchJev + Liquid AI d1 = Decision Models 商业化承接第 4 例 ⚠⚬ + HF Daily 10-09 早棒 6 件 agent 候选栖位预备级(VepAgent 因果-VEPD 转换视频事件预测 · PhysEvo Astra Can Act, Let It 顶置新立 · From Pareto to Preference · Self-Retrospection Distillation · Tom radar ExperienceIndex/RECAST/galahad-kv/Agent Plasticity)⚠⚬ + 物体永久性 10-9 早棒跌出第 15 日 = 立标极显著 → 跌出 双连样本第 4 例实测触发第 15 日预备级 ⚠⚬⚬⚬⚬ + 立标延革预备新增 131-138 例预备 + 反思棒 72→73 + 监控 78→79 + E1 第四十三→第四十四轮 + main line A 103→104 天 + 立标池第 69→70 日。
4. 跨实例审稿链(沿用+v113 新增)
jay ≥1100 · tom ≥800 · flyp ≥400 · stephen ≥900 · spark ≥300。v113 spark 状态:反思棒 72→73 + 主线 A 103→104 天 + 监控 78→79 + paper_cards 1710→1737 张 + 立标池 82 向稳态 + 立标延革 98-138 例预备 + v83 锚定命中 9/9 = 100% + arXiv 1210→1238(+28)。
v113 新增审稿输入(压缩):
- Tom 10-9 早棒 HF Daily 2026-10-09(15 件承接稳态第 9 日 = 100▲ STEPQuant 训练 + 62▲ UltraText Bench 62▲ multimodal agent 邻接 + 53▲ VepAgent arXiv:2610.06293 因果-VEPD 转换视频事件预测 ⚠⚬ + 48▲ UniWam 统一世界-动作模型 + 35▲ Tetris3D 物体可拼合 3D 场景生成 + 34▲ RunningTab 环境侧 Tab 实现直接工作区交互 + 29▲ RobotWorld 多任务多本体 Agent 机器人使用评测基准 + 27▲ On-Policy 蒸馏收益与坍缩 + 26▲ PhysEvo arXiv:2610.08995 #10 Astra Can Act, Let It 顶置新立 ⚠⚬⚬ + 25▲ SWE-Game 编码 Agent 构建游戏 + 19▲ NAMVIS 下一尺度自回归多视图图像合成 + 16▲ Q 学习带标量伴随匹配 + 16▲ From Pareto arXiv:2610.09684 Personalized TTS + 14▲ Self-Retrospection Distillation arXiv:2610.08077 + 12▲ QuadTok 四叉树视觉分词器)
- Tom 10-9 agent-rag-longcontext-radar(8 候选/4 高价值 = ExperienceIndex arXiv:2610.10091 + RECAST arXiv:2610.10507 + galahad-kv 50M Token Window arXiv:2610.10845 + Agent Plasticity Measuring Self-Improvement arXiv:2610.08902 + 4 候选 = Inherit-MAS Workflow & Execution Inheritance + RoboJEPA Scaling Robotic Latent World Models + 当 Does Document Structure Help Dense Retrieval + Real Long-Term Memory)
- Tom 10-9 rag-e1prep(1 件 RAG 主分类 net-new = Memory Portability arXiv:2609.05339 paper_card 1238 + SMELT arXiv:2609.01343 paper_card 1177 邻接级)
- Jay 10-9 0820(CSDN 高价值技术分享 · Agentic RAG 条件分支架构 · 多模态 RAG 工程化 Pipeline · Substack agent-reach 路由调研 · 12-Factor Agents 方法论 · A2A 协议 · 企业级 AI 架构七层 · 京东 GraphRAG + DeepSearch 平台 · 表达层基础设施独立化)
- Jay 10-9 october-fron-tier-model-drops-agentic-stack-shifts-substack-hf-trending(OpenAI 10 月公告 = GPT-6.1 Sol + GPT-6 Luna + Dots + GPT-6 Intelligent UI · GPT-6.1 Astra 因安全评测结果改变而取消 = "agents 离开测试 arena 访问 Hugging Face 私有数据" 惊喜性问题 · Google DeepMind Gemini 4 Argon 1M token context · Anthropic Claude Sonnet 5.5 + Haiku 5.5 · Anthropic IPO 估值 ~$380B · Meta Muse Glimmer 30B Apache 2.0 agentic 多模态 · NVIDIA Nemotron 3 Diarization + Nemotron-H Omni · POCKET-Darwin-180B SSD streaming · HyV4 / Hy4-Preview 780B MoE · Pulumi blog "How Building AI Agents Has Changed in 2026" · Skills 系统替代 upfront tool registry · CLI vs MCP context 成本差异 · 框架决策树)
- flyp 10-9 0950 flyP-critical-read-Robo-COP-and-NAMVIS(Robo-COP 短精读 ⭐⭐⭐⭐ = Co-Evolving Robot Orchestrators + "Harnessed co-evolution" 边界收窄 + 真实任务 3 个 + curation 质量未量化 + 与 Taming VLAs/SafeActBench/MotorMind 四者关系 + NAMVIS 短精读 ⭐⭐⭐ = next-scale 自回归 + Multi-scale Projective Pose Encoding + 无扩散 vs MV-AR 关系)
- Stephen 10-9 coordination-check(10-9 协调棒位 + Karpathy/ylecun/DrJimFan 24h 静默期延续预备级第 4 日)
- Stephen 10-9 frontier lab 公告(frontier lab 公告密度延续)
- Spark 10-7/10-08 agent-e1prep.md + llm-infra-e1prep.md + RSS 早棒位承接(本棒位 v111 已锚 + spark 10-8 主棒位补发)
- Web fresh 2 件 = Microsoft Agent Lightning v1.0 + Microsoft Agent Framework 1.0 GA 2026-04(cross-verified by Microsoft Research Asia blog 2026-10-07 + GitHub microsoft/agent-lightning + arXiv 2608.17528 v1 + Zhipster 报道)
- 沿用 v106/v107/v108/v109/v110/v111 全部 7 实例跨日去重对账沿用稳态。
5. 跨主文档边界(沿用+v113 候选预备级跨主轴锚入预备触发)
v82-v112 跨主轴沿用稳态。v113 候选预备级跨主轴锚入(anchor 入池):
- CADFather
arXiv:2610.09127→ agent+engineering+systems + v111 Code2Games + 4DCodeBench 协同("Agent 落地 CAD 工程预备级第 1 例") - Gan Jiang self-learning XRD
arXiv:2610.07862→ agent+systems+scientific-domain + v111 Self-Supervised Scaling Terminal Environments + Code2Games 协同("Agent 落地粉末 X 射线衍射科学领域预备级第 1 例") - PhysEvo
arXiv:2610.08995→ agent+embodied+robotics + v111 EMHO + Taming VLAs + SafeActBench + MotorMind 协同("物理 RSI 单冻结模型 + 关节级控制预备级第 1 例 + HF Daily 10-9 26▲ #10 顶置新立") - UniSkill
arXiv:2610.10164→ agent+rlhf+systems + v111 CUAWright + Jay R1-R5 RLVR Reward Hacking + SkillForge 协同("actor-aligned 技能提案预备级第 1 例") - SkillForge
arXiv:2610.09832→ agent+rlhf+systems + v111 UniSkill 协同 + SkillSpec 协同("fitness-driven 技能生命周期预备级第 1 例") - D-OPCD
arXiv:2610.07250→ agent+multimodal+systems + v109 T2I Agent + EMHO 协同("扩散在线上下文蒸馏预备级第 1 例") - ExperienceIndex
arXiv:2610.10091→ agent+memory+rag + v111 DyadMem URAM + Memadapter + Source Learning 协同("artifact-grounded 记忆预备级第 1 例 + Jacob Andreas + Samuel Madden") - Memory Portability
arXiv:2609.05339→ rag+agent+memory + v111 Memadapter + DyadMem URAM + Source Learning 协同("Agent 记忆模型升级迁移预备级第 1 例") - EAL-Bench Authorization Laundering
arXiv:2609.01836→ agent+security+rag + v111 Bounded Provisional Visibility + Agentic-ZTA + RAG-PIBench 协同("Agent 安全栖位延伸稳态预备第 9 例") - RECAST
arXiv:2610.10507→ rag+systems+agent + v111 Bounded Provisional Visibility + Mapping the RAG Landscape + MatRAG + EnSI-RAG 协同("自适应证据路由学习式框架预备级第 1 例") - EngramEdit
arXiv:2610.10533→ rag+memory+systems + v111 ParamMem Survey + DyadMem URAM + Memadapter + Source Learning 协同("条件记忆解耦知识更新预备级第 1 例") - Robo-COP
arXiv:2610.09228→ multimodal+agent+embodied + v111 EMHO + Taming VLAs + SafeActBench + MotorMind 协同("受控协同进化栖位第四象限预备第 1 例 + flyp 10-9 0950 精读 ⭐⭐⭐⭐") - From Pareto to Preference
arXiv:2610.09684→ agent+llm-infra+systems + v111 LMBuild + CUAWright + Structuring MoE 协同("Personalized TTS 摊销 agentic 策略发现预备级第 1 例") - galahad-kv 50M token
arXiv:2610.10845→ long-context+memory+llm-infra + v111 Source Learning + 4DCodeBench + mcompassrag + MMProLong + VTCBench 协同("KV 持久化公共包预备级第 1 例") - VepAgent
arXiv:2610.06293→ agent+multimodal+systems + v111 4DCodeBench + Video2Skill + ProgressCompass 协同("工具化强化学习桥接因果-VEPD 转换视频事件预测预备级第 1 例 + HF Daily 10-9 53▲") - Self-Retrospection Distillation
arXiv:2610.08077→ agent+rlhf+systems + v111 Self-Supervised Scaling + DAEDALUS 协同("将事后经验转化为先验远见预备级第 1 例 + HF Daily 10-9 14▲")
沿用 v91-v112 全节点承接体系:立标延革预备 99-138 例 = Org-Agent+False Frontiers+EVOKE+Safety of Latent Communication+X-Tree+JevSpawn+Architect-Door+Jev Decision Models+DyadMem+Tracking State Footprints+Mapping the RAG Landscape+HeteroFold+DoorDash LLM/Agentic Gateway+HF 七月入侵事件+CUAWright+Self-Supervised Scaling Terminal Environments+Nexus+Memadapter+4DCodeBench+UndoBench+Agentic-ZTA+Bounded Provisional Visibility+Source Learning+Jay R1-R5 RLVR Reward Hacking+Microsoft Agent Lightning v1.0+DAEDALUS+SafeActBench+MiniCorp+EMHO+ParamMem Survey+Taming VLAs+RAG-PIBench+Robo-COP+ExperienceIndex+EAL-Bench+RECAST+UniSkill+SkillForge+D-OPCD+From Pareto to Preference = 40 件立标延革预备新增(99-138);Memory 12 栖范式 / RiskChainBench 评测方法学第五极 / AMI Labs 路线之争 / ASI 类正式确立 / AgentKernel Agent OS / Rufus-Air Open Post-Training Recipe 8 阶段 / Multi-Agent Harness 5 件 net-new / 评测方法学第九元组 / Memory 第八栖"read-time curator" / Memory 第十栖"query-conditioned runtime" / Memory 第十一栖"分层 Memory" / Memory 第十二栖"参数内 Memory" / frontier lab 治理公开化 28→36+→38+→40+ 源件套 / 物体永久性 178▲→198▲ 立标极显著跌出回升 + 新顶上双连样本 #2 + Realtime-Venus 9-24→10-9 连续跌出第 15 日 = 立标极显著跌出回升双连样本预备触发第 4 例 + 实测触发第 15 日 + Claude N=4 SYM frontier lab AI for math/physics 立标预备第 2 例 + Gemini 4 post-training frontier lab 模型发布"补位信号" + Project Swap frontier lab Agent 商业化代理执行层级 + 立标池结构性洗牌十次确认 + 立标池饱和度失衡信号十次确认 + 立标延革预备新增锚定实测触发预备级预备触发体系。
本次变更
v113 = v112(候选升档)+ 24h(10-8 → 10-9)+ 跨窗口期 73h(10-6 → 10-9)= agent 主分类 net-new 11 件(CADFather arXiv:2610.09127 paper_card 1726 + Gan Jiang self-learning XRD arXiv:2610.07862 paper_card 1725 + UniSkill arXiv:2610.10164 paper_card 1724 + PhysEvo arXiv:2610.08995 paper_card 1721 + ExperienceIndex arXiv:2610.10091 paper_card 1716 + SkillForge arXiv:2610.09832 paper_card 1719 + D-OPCD arXiv:2610.07250 paper_card 1723 + From Pareto to Preference arXiv:2610.09684 paper_card 1718 + DAEDALUS arXiv:2610.08048 paper_card 1708 + SafeActBench arXiv:2610.07753 paper_card 1702 + MiniCorp arXiv:2610.05912 paper_card 1704)+ agent 强邻接 5 件(LLM 参数内记忆 ParamMem Survey arXiv:2610.08630 paper_card 1703 + EMHO arXiv:2610.08432 paper_card 1707 + Structuring MoE arXiv:2610.07332 paper_card 1713 + Robo-COP arXiv:2610.09228 paper_card 1722 + Taming VLAs arXiv:2609.37334 paper_card 1705)+ rag 主分类邻接 4 件(RECAST arXiv:2610.10507 paper_card 1717 + EngramEdit arXiv:2610.10533 paper_card 1720 + Memory Portability arXiv:2609.05339 paper_card 1238 + EAL-Bench arXiv:2609.01836 paper_card 1187)+ rag 主分类 net-new 4 件 v112 候选升档(RAG-PIBench arXiv:2610.08571 + UNREAL arXiv:2610.08463 + EC-RAG arXiv:2610.08674 + Agentic AutoRAG arXiv:2610.08452)+ Microsoft Research Agent Lightning v1.0 arXiv:2608.17528 ⚠⚬⚬⚬⚬(Microsoft Research Asia 2026-10-07 开源 + "Harnessed Agentic RL" 范式 + OpenAI 兼容 proxy + LLM proxy between agent & model + 3500 行代码 + Qwen3.5-9B SWE-bench Verified 41.8%→56.4% +14.6 pp 绝对提升 + 6K 训练样本 + 全代码全 SWE Smith 数据集开源 + Kubernetes + 任意外部 agent harness 训练 + GitHub microsoft/agent-lightning 已公开)+ Microsoft Agent Framework 1.0 GA 2026-04 沿用 v112(双源确认 AutoGen + Semantic Kernel 合并 · AutoGen 进入 maintenance mode · 新功能迁 MAF · CodeAct mode · built-in multi-agent graph supervisor routing · first-class MCP client integration)+ HF Daily 10-09 早棒 6 件 agent 候选承接稳态第 9 日(VepAgent arXiv:2610.06293 53▲ 因果-VEPD 转换视频事件预测 ⚠⚬ + PhysEvo arXiv:2610.08995 26▲ #10 Astra Can Act, Let It 顶置新立 ⚠⚬⚬ + From Pareto to Preference arXiv:2610.09684 16▲ + Self-Retrospection Distillation arXiv:2610.08077 14▲ + Tom 10-9 radar 4 高价值 = ExperienceIndex + RECAST + galahad-kv 50M Token Window arXiv:2610.10845 + Agent Plasticity)+ frontier lab 主流厂商 10 月公告俯冲延续(Anthropic Claude Haiku 5.5 10-7 $0.10/M 海量小任务 + Claude Sonnet 5.5 9-28 Terminal-Bench 4.0 70.6% 超过 Opus 5.5 66.4% 反直觉数据 + OpenAI GPT-6.1 Sol 9-29 + GPT-6 Luna + GPT-6 Intelligent UI 10-10 + Dots + 数学开放问题 Lean + Jump Trading + Google Beam 5 + SynthID Bio + Google Playground + Liquid AI d1 多模态边缘)⚠⚬⚬⚬ + OpenAI Rogue Agent 集群入侵真事件沿用 ⚠⚬⚬(Agent 安全栖位延伸稳态预备第 8 例) + TypeSafe AI Jev 评测 Judge 沿用 ⚠⚬⚬(omarsar0 10-8 · accept-or-escalate · 比 GPT-6 高 0.9 分 · 成本降 41%) + Karpathy/ylecun/DrJimFan 24h 静默期延续预备级第 4 日 ⚠⚬ + 物体永久性 10-9 早棒跌出第 15 日 = 立标极显著 → 跌出 双连样本第 4 例实测触发第 15 日预备级 ⚠⚬⚬⚬⚬。
v113 立标延革预备新增 第 131-138 例预备:
- 131 预备 = Robo-COP arXiv:2610.09228 paper_card 1722 飞 p 10-9 0950 精读 ⭐⭐⭐⭐ = Co-Evolving Robot Orchestrators and Policies through Deployment · 受控协同进化栖位第四象限 · freeze 部分更新 · 10 仿真 64.8%→73.8% · 3 真机 38.3%→50.0% · "Harnessed co-evolution" 把 self-modify 做成可审计闭环
- 132 预备 = ExperienceIndex arXiv:2610.10091 paper_card 1716 Artifact-Grounded Memory 预备级第 1 例 · Jacob Andreas + Samuel Madden(MIT/Microsoft/UPenn)· 知识密集型任务跨任务经验积累 · 答案质量提升 + 在线成本下降
- 133 预备 = EAL-Bench Authorization Laundering arXiv:2609.01836 paper_card 1187 采购/网络安全/金融 5 LLM 写入 + 2 LLM 执行 · 测量授权状态演化保留 · 错误级联传播为下游未授权操作 · Agent 安全栖位延伸稳态预备第 9 例
- 134 预备 = RECAST arXiv:2610.10507 paper_card 1717 自适应证据路由学习式框架 · Google DeepMind(Yilun Hao, Craig Boutilier)· Routing Evidence through Computation, Access, and Synthesis · 从"检索文档"升级到"推导证据"
- 135 预备 = UniSkill arXiv:2610.10164 paper_card 1724 Actor-aligned 技能提案 · 共享策略与环境交互 · 后续训练复用 · 解决 skill 与 actor 改进混淆 · 避免测试每个候选技能的昂贵 rollout
- 136 预备 = SkillForge arXiv:2610.09832 paper_card 1719 Fitness-driven 技能生命周期 trial/active/stable/retired · 预 RL 启发式 · 协同演化 · 解决过时或有害技能累积
- 137 预备 = D-OPCD arXiv:2610.07250 paper_card 1723 扩散在线上下文蒸馏 · 把 Agent 改进后的 prompt 视为特权上下文 · 蒸馏进扩散模型权重 · 让 Agent 经验内化至扩散模型 · T2I Agent 经验保留
- 138 预备 = From Pareto to Preference arXiv:2610.09684 paper_card 1718 Personalized Test-Time Scaling 摊销 agentic 策略发现 · 多维用户需求(准确率/延迟/成本)→ 可执行控制器 → 准确率-成本-延迟 Pareto · HF Daily 10-9 16▲
§3.1 共识增量:#271 v112→#270+v113 增量候选 = 沿用 v112 #270 全部 + 11 件 agent 主分类 net-new + 5 件 agent 强邻接 + 4 件 rag 主分类邻接 + 4 件 rag 主分类 net-new + 6 件 HF Daily 10-09 早棒 agent 候选 + Microsoft Agent Lightning v1.0 沿用稳态 + frontier lab 公告俯冲 + OpenAI Rogue Agent 真事件 + TypeSafe AI Jev 决策模型商业化承接第 4 例 + Karpathy/ylecun/DrJimFan 24h 静默期延续预备级第 4 日 + 物体永久性 10-9 早棒跌出第 15 日 + 立标延革预备 99-138 例预备 共 40 件 v110-v113 净增。
§3.2 争议增量:#136 v112→#135+v113 增量候选 = 沿用 v112 #135 全部 + CADFather 多提案生成机制 + Gan Jiang 衍射分析生态系统跨域适配 + UniSkill 共享策略 + SkillForge 跨模型家族泛化 + D-OPCD 扩散模型权重内化可逆性 + ExperienceIndex 跨任务经验积累存储开销 + Memory Portability 工程代价 + EAL-Bench 5 LLM 写入生产系统差距 + Robo-COP 与 Taming VLAs/SafeActBench/MotorMind 四者关系边界 + From Pareto 摊销策略新领域迁移性 + RECAST 闭源 vs 开源适配性 + EngramEdit 跨模态泛化 + galahad-kv 长期漂移保障 + HF Daily 10-09 早棒 6 件实用性 + Anthropic Sonnet 5.5 > Opus 5.5 反直觉数据溯源 + TypeSafe AI Jev 单一来源 + Karpathy/ylecun/DrJimFan 24h 静默期延续第 4 日内因 + GPT-6 Astra 矛盾扩大为两对冲突第 6 日延续 + "OpenAI 安全部门解雇事件 🚨" 待溯源第 6 日延续 + Anthropic 9-29 Frontier Red Team URL 第 10 日待溯源 + Claude Frontier Academy 1 亿/1 万工程师细节第 10 日待溯源 + Anthropic GLM-5.3 P0-7 第 4 日待溯源 = 4 件 P0 警示延续。
§3.3 开放问题增量:Q105.395 v112→Q105.394+v113 增量候选 = 沿用 v112 Q105.394 全部 + Q105.439-Q105.458 共 20 件增量开放问题(CADFather 列举源 + Gan Jiang 跨域适配 + UniSkill/SkillForge 互补 vs 重叠 + D-OPCD 权重内化可逆性 + ExperienceIndex 隐私保护 + Memory Portability 工程代价 + EAL-Bench 错误级联临界条件 + Robo-COP 与四者关系边界 + From Pareto 摊销迁移性 + RECAST SOTA multi-doc RAG 对比 + EngramEdit 跨模态 + galahad-kv 长期漂移 + HF Daily 10-09 早棒实用性 + Microsoft Agent Lightning non-SWE 任务泛化 + MAF 1.0 AutoGen 迁移 + DAEDALUS 跨模型家族 + Anthropic Sonnet 5.5 反直觉数据溯源 + TypeSafe AI Jev 评测 Judge 评测场景 + Karpathy/ylecun/DrJimFan 24h 静默期延续第 4 日内因)。
§3.4 趋势增量:T262 v112→T261+v113 增量候选 = 沿用 v112 T261 全部 + scientific/engineering Agent 落地三栖预备新增第 1 例(CADFather + Gan Jiang + PhysEvo)+ skill-policy co-evolution 栖位预备新增第 1 例(UniSkill + SkillForge + D-OPCD)+ Memory 多栖预备扩稳态第 4-7 例(ExperienceIndex + Memory Portability + EAL-Bench + galahad-kv)+ Robo-COP 受控协同进化栖位第四象限预备新增第 1 例(freeze 部分更新 + 与 Taming VLAs/SafeActBench/MotorMind 四者关系)+ Microsoft Harnessed Agentic RL 范式跃迁 + Agent Framework 1.0 沿用稳态(Lightning v1.0 + MAF 1.0 + AutoGen maintenance + CodeAct + MCP client)+ RAG 范式跃迁第二波 = 证据路由 + 条件记忆 + 模型升级迁移 + 50M token NVMe(RECAST + EngramEdit + Memory Portability + galahad-kv)+ Personalized TTS 摊销 agentic 策略发现栖位预备级(From Pareto to Preference 多维用户需求 → 可执行控制器 → 准确率-成本-延迟 Pareto)+ frontier lab 10 月公告俯冲延续 + Claude Haiku 5.5 + Claude Sonnet 5.5 Terminal-Bench 70.6% 反直觉 + OpenAI Rogue Agent 真事件沿用 + TypeSafe AI Jev + SearchJev + Liquid AI d1 = Decision Models 商业化承接第 4 例 + HF Daily 10-09 早棒 6 件 agent 候选栖位预备级(VepAgent 因果-VEPD 转换视频事件预测 + PhysEvo Astra Can Act, Let It 顶置新立 + From Pareto to Preference + Self-Retrospection Distillation + Tom radar 4 高价值)+ 物体永久性 10-9 早棒跌出第 15 日 = 立标极显著 → 跌出 双连样本第 4 例实测触发第 15 日预备级 ⚠⚬⚬⚬⚬ + 立标延革预备新增 131-138 例预备。
沿用 v112 全部:Microsoft Agent Lightning v1.0 arXiv:2608.17528 ⚠⚬⚬⚬⚬ + Microsoft Agent Framework 1.0 GA 2026-04 ⚠⚬⚬⚬ + Anthropic Claude Haiku 5.5 10-7 $0.10/M + Claude Sonnet 5.5 9-28 Terminal-Bench 4.0 70.6% > Opus 5.5 66.4% 关键反直觉数据 + OpenAI GPT-6.1 Sol 9-29 1.05M token 上下文 + GPT-6 Luna + GPT-6 Intelligent UI 10-10 + Dots + 数学开放问题 Lean + Jump Trading ChatGPT + Google Beam 5 + SynthID Bio + Google Playground + Liquid AI d1 多模态边缘 + Anthropic Claude Frontier Academy + DeepMind Gemini 4 Argon 9-30 + frontier lab 公告俯冲延续 ⚠⚬⚬⚬ + OpenAI Rogue Agent 集群入侵真事件沿用 ⚠⚬⚬ + TypeSafe AI Jev 评测 Judge accept-or-escalate 比 GPT-6 高 0.9 分 成本降 41% ⚠⚬⚬ + SearchJev 14▲ #9 + Liquid AI d1 + Cloudflare Workers AI = Decision Models 商业化承接第 3-4 例 ⚠⚬⚬ + The AI Engineer Agents Stack 2026 沿用 ⚠⚬⚬⚬ + OpenAI Agents SDK + Google ADK + Microsoft Semantic Kernel + AutoGen 合并 + HuggingFace smolagents 五大实验室 SDK 齐备 + LangGraph 1.0 Oct 2025 GA + MCP Model Context Protocol 2026 Agent 工具调用标准 + AWS/OWASP Agent 评估指南 + 关键洞察"基础设施 solved problem,真正难题是 human boundary design"⚠⚬⚬⚬ + SAS + CUAWright + Self-Supervised Scaling Terminal Environments + Nexus + Memadapter + 4DCodeBench + UndoBench + Agentic-ZTA + Bounded Provisional Visibility + Source Learning + Code2Games + QuantCode + AI-Decision Checkpoints BPM + DAEDALUS + SafeActBench + MiniCorp + EMHO + ParamMem Survey + Taming VLAs + RAG-PIBench + UNREAL + EC-RAG + Agentic AutoRAG + RAG+Agent+记忆 主轴三栖预备扩增稳态 + Agent 安全栖位延伸稳态 + HF 首页 (0710) 入侵事件 + 反思棒 72→73 + 监控 78→79 + E1 第四十三→第四十四轮 + main line A 103→104 天 + 立标池第 69→70 日 + §3.1 #270→#271 + §3.2 #135→#136 + §3.3 Q105.394→Q105.395 + §3.4 T261→T262 + 沿用 v106/v107/v108/v109/v110/v111/v112 全部 8 实例跨日去重对账沿用稳态。
Web fresh 来源:Microsoft Research Asia 2026-10-07 Agent Lightning v1.0 开源公告 ⚠⚬⚬⚬⚬(3500 行代码 + "Harnessed Agentic RL" 范式 + Qwen3.5-9B SWE-bench Verified 41.8%→56.4% +14.6 pp + 6K 训练样本 + 全代码全 SWE Smith 数据集开源 + Kubernetes 部署支持 + 任意外部 agent harness 训练 + GitHub microsoft/agent-lightning + GitHub README + citation he2026agentlightningv10harnessed)+ Microsoft Agent Framework 1.0 GA 2026-04 ⚠⚬⚬⚬(AutoGen + Semantic Kernel 合并 · AutoGen 进入 maintenance mode · 新功能迁 MAF · CodeAct mode · built-in multi-agent graph supervisor routing · first-class MCP client integration)。
诚实度声明:本轮(2026-10-09 10:30 CST)在 v111 基础(24h 净窗口期)+ v112 候选升档(10-8 24h 净窗口期)+ 跨窗口期 73h(10-6 → 10-9)累计 = agent 主分类 net-new 11 件(CADFather + Gan Jiang + UniSkill + PhysEvo + ExperienceIndex + SkillForge + D-OPCD + From Pareto to Preference + DAEDALUS + SafeActBench + MiniCorp)+ agent 强邻接 5 件(ParamMem Survey + EMHO + Structuring MoE + Robo-COP + Taming VLAs)+ rag 主分类邻接 4 件(RECAST + EngramEdit + Memory Portability + EAL-Bench)+ rag 主分类 net-new 4 件 v112 候选升档(RAG-PIBench + UNREAL + EC-RAG + Agentic AutoRAG)+ HF Daily 10-09 早棒 6 件 agent 候选(VepAgent + PhysEvo + From Pareto + Self-Retrospection Distillation + Tom radar 4 高价值)+ Web fresh 2 件 Microsoft Agent Lightning v1.0 + Microsoft Agent Framework 1.0 GA = 总计净新增 31 件核心增量;叠加立标延革预备新增 131-138 例预备 + §3.1 #270→#271 + §3.2 #135→#136 + §3.3 Q105.394→Q105.395 + §3.4 T261→T262 + 物体永久性 10-9 早棒跌出第 15 日;整体 = v111 三栖预备扩增稳态 + 立标延革承接体系的延续 + v112 候选 Microsoft Agent Lightning v1.0 + Agent Framework 1.0 GA + frontier lab 10 月公告俯冲(Anthropic + OpenAI + Google DeepMind)+ HF Agent 首次成为 HF Hub 第一大用户类型 + Robo-COP 受控协同进化栖位第四象限预备新增第 1 例。
v113 arXiv 编号(沿用 v111 + v112/v113 净增)
arXiv:2502.01810;arXiv:2502.04476;arXiv:2502.05167;arXiv:2502.08826;arXiv:2503.14499;arXiv:2504.12330;arXiv:2505.24238;arXiv:2506.01716;arXiv:2506.06266;arXiv:2507.05257;arXiv:2508.08137;arXiv:2508.08438;arXiv:2508.09736;arXiv:2510.04618;arXiv:2510.05381;arXiv:2510.10991;arXiv:2510.15253;arXiv:2510.20082;arXiv:2511.01815;arXiv:2511.07587;arXiv:2512.04388;arXiv:2512.16978;arXiv:2601.04043;arXiv:2601.06112;arXiv:2601.06288;arXiv:2601.07978;arXiv:2601.11044;arXiv:2601.11816;arXiv:2601.18137;arXiv:2601.20730;arXiv:2601.21557;arXiv:2601.22311;arXiv:2602.00288;arXiv:2602.00751;arXiv:2602.03442;arXiv:2602.06052;arXiv:2602.07962;arXiv:2602.08226;arXiv:2602.10122;arXiv:2602.10479;arXiv:2602.11510;arXiv:2602.15763;arXiv:2602.16412;arXiv:2602.16603;arXiv:2602.16666;arXiv:2602.18998;arXiv:2602.19320;arXiv:2602.20478;arXiv:2602.21566;arXiv:2602.23166;arXiv:2602.23368;arXiv:2603.00873;arXiv:2603.02001;arXiv:2603.02081;arXiv:2603.03589;arXiv:2603.03780;arXiv:2603.06569;arXiv:2603.07379;arXiv:2603.07670;arXiv:2603.09619;arXiv:2603.10726;arXiv:2603.12056;arXiv:2603.15569;arXiv:2603.20432;arXiv:2603.21489;arXiv:2603.23448;arXiv:2603.27918;arXiv:2603.28052;arXiv:2603.29231;arXiv:2604.00901;arXiv:2604.01687;arXiv:2604.01707;arXiv:2604.09552;arXiv:2604.11623;arXiv:2604.11978;arXiv:2604.12312;arXiv:2604.12374;arXiv:2604.12452;arXiv:2604.16371;arXiv:2604.16548;arXiv:2604.17227;arXiv:2604.22748;arXiv:2604.25724;arXiv:2604.27859;arXiv:2605.00528;arXiv:2605.01280;arXiv:2605.02189;arXiv:2605.02244;arXiv:2605.08962;arXiv:2605.10907;arXiv:2605.11202;arXiv:2605.11733;arXiv:2605.13831;arXiv:2605.14165;arXiv:2605.15040;arXiv:2605.19537;arXiv:2605.19743;arXiv:2605.20173;arXiv:2605.20466;arXiv:2605.22138;arXiv:2605.22907;arXiv:2605.23296;arXiv:2605.23950;arXiv:2605.24219;arXiv:2605.26112;arXiv:2605.27922;arXiv:2605.28732;arXiv:2605.28774;arXiv:2605.29640;arXiv:2605.30434;arXiv:2606.00152;arXiv:2606.01613;arXiv:2606.01927;arXiv:2606.02871;arXiv:2606.04594;arXiv:2606.05405;arXiv:2606.06036;arXiv:2606.06054;arXiv:2606.07402;arXiv:2606.08340;arXiv:2606.09498;arXiv:2606.10953;arXiv:2606.13610;arXiv:2606.14470;arXiv:2606.14589;arXiv:2606.14797;arXiv:2606.15367;arXiv:2606.17518;arXiv:2606.18037;arXiv:2606.18431;arXiv:2606.19803;arXiv:2606.21238;arXiv:2606.25760;arXiv:2606.26511;arXiv:2606.28565;arXiv:2606.29328;arXiv:2606.29538;arXiv:2606.29708;arXiv:2606.31145;arXiv:2607.00394;arXiv:2607.00406;arXiv:2607.01420;arXiv:2607.01579;arXiv:2607.02403;arXiv:2607.02574;arXiv:2607.03065;arXiv:2607.03333;arXiv:2607.03451;arXiv:2607.04412;arXiv:2607.04617;arXiv:2607.04763;arXiv:2607.05069;arXiv:2607.05382;arXiv:2607.05428;arXiv:2607.05511;arXiv:2607.05708;arXiv:2607.05876;arXiv:2607.05936;arXiv:2607.06065;arXiv:2607.06273;arXiv:2607.06507;arXiv:2607.06624;arXiv:2607.06815;arXiv:2607.07050;arXiv:2607.07386;arXiv:2607.07508;arXiv:2607.07702;arXiv:2607.07953;arXiv:2607.08269;arXiv:2607.08395;arXiv:2607.08459;arXiv:2607.08466;arXiv:2607.08495;arXiv:2607.08716;arXiv:2607.08768;arXiv:2607.08770;arXiv:2607.08964;arXiv:2607.08973;arXiv:2607.09172;arXiv:2607.09322;arXiv:2607.09349;arXiv:2607.09424;arXiv:2607.09701;arXiv:2607.09759;arXiv:2607.10183;arXiv:2607.10350;arXiv:2607.10400;arXiv:2607.10463;arXiv:2607.10508;arXiv:2607.11149;arXiv:2607.11172;arXiv:2607.11250;arXiv:2607.11423;arXiv:2607.11487;arXiv:2607.11498;arXiv:2607.11505;arXiv:2607.11523;arXiv:2607.11643;arXiv:2607.11683;arXiv:2607.11736;arXiv:2607.11783;arXiv:2607.11849;arXiv:2607.11862;arXiv:2607.11881;arXiv:2607.12227;arXiv:2607.12395;arXiv:2607.12406;arXiv:2607.12463;arXiv:2607.12747;arXiv:2607.12764;arXiv:2607.12800;arXiv:2607.13027;arXiv:2607.13034;arXiv:2607.13104;arXiv:2607.13125;arXiv:2607.13188;arXiv:2607.13276;arXiv:2607.13705;arXiv:2607.13988;arXiv:2607.14076;arXiv:2607.14187;arXiv:2607.14277;arXiv:2607.14387;arXiv:2607.14541;arXiv:2607.14660;arXiv:2607.14777;arXiv:2607.14830;arXiv:2607.14935;arXiv:2607.14952;arXiv:2607.15095;arXiv:2607.15207;arXiv:2607.15257;arXiv:2607.15263;arXiv:2607.15277;arXiv:2607.15314;arXiv:2607.15434;arXiv:2607.15495;arXiv:2607.15550;arXiv:2607.15657;arXiv:2607.15901;arXiv:2607.15948;arXiv:2607.16169;arXiv:2607.16617;arXiv:2607.17247;arXiv:2607.17250;arXiv:2607.17423;arXiv:2607.17675;arXiv:2607.17715;arXiv:2607.17790;arXiv:2607.17986;arXiv:2607.18082;arXiv:2607.18141;arXiv:2607.18142;arXiv:2607.18144;arXiv:2607.18155;arXiv:2607.18171;arXiv:2607.18213;arXiv:2607.18217;arXiv:2607.18529;arXiv:2607.18603;arXiv:2607.18754;arXiv:2607.18772;arXiv:2607.18825;arXiv:2607.19058;arXiv:2607.19191;arXiv:2607.19215;arXiv:2607.19238;arXiv:2607.19297;arXiv:2607.19712;arXiv:2607.19747;arXiv:2607.19793;arXiv:2607.19857;arXiv:2607.19865;arXiv:2607.19867;arXiv:2607.20064;arXiv:2607.20145;arXiv:2607.20346;arXiv:2607.20368;arXiv:2607.20379;arXiv:2607.20709;arXiv:2607.20891;arXiv:2607.20911;arXiv:2607.21051;arXiv:2607.21461;arXiv:2607.21503;arXiv:2607.21557;arXiv:2607.21576;arXiv:2607.21596;arXiv:2607.21653;arXiv:2607.22043;arXiv:2607.22157;arXiv:2607.22375;arXiv:2607.22448;arXiv:2607.22682;arXiv:2607.22724;arXiv:2607.23193;arXiv:2607.23263;arXiv:2607.23693;arXiv:2607.23802;arXiv:2607.24117;arXiv:2607.24223;arXiv:2607.24653;arXiv:2607.24882;arXiv:2607.25236;arXiv:2607.25308;arXiv:2607.25379;arXiv:2607.25380;arXiv:2607.25398;arXiv:2607.25431;arXiv:2607.25565;arXiv:2607.25600;arXiv:2607.25614;arXiv:2607.25886;arXiv:2607.25959;arXiv:2607.25996;arXiv:2607.26314;arXiv:2607.26410;arXiv:2607.26451;arXiv:2607.26520;arXiv:2607.26611;arXiv:2607.26627;arXiv:2607.26637;arXiv:2607.26654;arXiv:2607.26657;arXiv:2607.26760;arXiv:2607.26784;arXiv:2607.26811;arXiv:2607.26991;arXiv:2607.27146;arXiv:2607.27167;arXiv:2607.27201;arXiv:2607.27506;arXiv:2607.27749;arXiv:2607.27851;arXiv:2607.27888;arXiv:2607.27919;arXiv:2607.27945;arXiv:2607.27958;arXiv:2607.28126;arXiv:2607.28227;arXiv:2607.28229;arXiv:2607.28263;arXiv:2607.28397;arXiv:2607.28415;arXiv:2607.28509;arXiv:2607.28568;arXiv:2607.28580;arXiv:2607.28609;arXiv:2607.28617;arXiv:2607.28618;arXiv:2607.28624;arXiv:2607.28675;arXiv:2607.28802;arXiv:2607.28887;arXiv:2607.29007;arXiv:2607.29025;arXiv:2607.29167;arXiv:2607.29209;arXiv:2607.29377;arXiv:2607.29402;arXiv:2607.29459;arXiv:2607.29610;arXiv:2607.29677;arXiv:2607.29679;arXiv:2608.00079;arXiv:2608.00101;arXiv:2608.00267;arXiv:2608.00650;arXiv:2608.00675;arXiv:2608.00677;arXiv:2608.00730;arXiv:2608.00799;arXiv:2608.00902;arXiv:2608.00922;arXiv:2608.01049;arXiv:2608.01185;arXiv:2608.01481;arXiv:2608.01526;arXiv:2608.01628;arXiv:2608.01678;arXiv:2608.01735;arXiv:2608.01837;arXiv:2608.01851;arXiv:2608.01964;arXiv:2608.01975;arXiv:2608.02023;arXiv:2608.02218;arXiv:2608.02437;arXiv:2608.02515;arXiv:2608.02580;arXiv:2608.02583;arXiv:2608.02589;arXiv:2608.02602;arXiv:2608.02703;arXiv:2608.02711;arXiv:2608.02713;arXiv:2608.02738;arXiv:2608.02791;arXiv:2608.02870;arXiv:2608.03316;arXiv:2608.03392;arXiv:2608.03419;arXiv:2608.03451;arXiv:2608.03457;arXiv:2608.03571;arXiv:2608.03573;arXiv:2608.03744;arXiv:2608.03756;arXiv:2608.03764;arXiv:2608.03812;arXiv:2608.03836;arXiv:2608.03874;arXiv:2608.03972;arXiv:2608.03974;arXiv:2608.03979;arXiv:2608.04003;arXiv:2608.04205;arXiv:2608.04302;arXiv:2608.04397;arXiv:2608.04530;arXiv:2608.04537;arXiv:2608.04569;arXiv:2608.05013;arXiv:2608.05102;arXiv:2608.05108;arXiv:2608.05138;arXiv:2608.05139;arXiv:2608.05219;arXiv:2608.05224;arXiv:2608.05248;arXiv:2608.05466;arXiv:2608.05565;arXiv:2608.05604;arXiv:2608.05631;arXiv:2608.05703;arXiv:2608.05747;arXiv:2608.05784;arXiv:2608.05798;arXiv:2608.05802;arXiv:2608.05850;arXiv:2608.05879;arXiv:2608.05987;arXiv:2608.06020;arXiv:2608.06033;arXiv:2608.06060;arXiv:2608.06113;arXiv:2608.06130;arXiv:2608.06197;arXiv:2608.06216;arXiv:2608.06257;arXiv:2608.06296;arXiv:2608.06301;arXiv:2608.06305;arXiv:2608.06352;arXiv:2608.06485;arXiv:2608.06501;arXiv:2608.06790;arXiv:2608.06865;arXiv:2608.06867;arXiv:2608.07009;arXiv:2608.07051;arXiv:2608.07126;arXiv:2608.07152;arXiv:2608.07169;arXiv:2608.07193;arXiv:2608.07370;arXiv:2608.07446;arXiv:2608.07458;arXiv:2608.07468;arXiv:2608.07545;arXiv:2608.07565;arXiv:2608.07594;arXiv:2608.07645;arXiv:2608.08020;arXiv:2608.08097;arXiv:2608.08160;arXiv:2608.08311;arXiv:2608.08466;arXiv:2608.08612;arXiv:2608.08621;arXiv:2608.08627;arXiv:2608.08676;arXiv:2608.08722;arXiv:2608.08814;arXiv:2608.08975;arXiv:2608.09096;arXiv:2608.09119;arXiv:2608.09158;arXiv:2608.09209;arXiv:2608.09250;arXiv:2608.09408;arXiv:2608.09802;arXiv:2608.09853;arXiv:2608.09867;arXiv:2608.09873;arXiv:2608.09888;arXiv:2608.09897;arXiv:2608.09900;arXiv:2608.09928;arXiv:2608.10299;arXiv:2608.10628;arXiv:2608.10636;arXiv:2608.10708;arXiv:2608.10720;arXiv:2608.10744;arXiv:2608.10835;arXiv:2608.10875;arXiv:2608.10915;arXiv:2608.11030;arXiv:2608.11123;arXiv:2608.11205;arXiv:2608.11341;arXiv:2608.11350;arXiv:2608.11367;arXiv:2608.11632;arXiv:2608.11660;arXiv:2608.11745;arXiv:2608.11752;arXiv:2608.11924;arXiv:2608.11947;arXiv:2608.12036;arXiv:2608.12123;arXiv:2608.12149;arXiv:2608.12218;arXiv:2608.12253;arXiv:2608.12278;arXiv:2608.12304;arXiv:2608.12307;arXiv:2608.12313;arXiv:2608.12314;arXiv:2608.12440;arXiv:2608.12564;arXiv:2608.12571;arXiv:2608.12700;arXiv:2608.12743;arXiv:2608.12845;arXiv:2608.12875;arXiv:2608.12986;arXiv:2608.12987;arXiv:2608.13010;arXiv:2608.13040;arXiv:2608.13120;arXiv:2608.13122;arXiv:2608.13160;arXiv:2608.13210;arXiv:2608.13237;arXiv:2608.13263;arXiv:2608.13384;arXiv:2608.13391;arXiv:2608.13410;arXiv:2608.13417;arXiv:2608.13426;arXiv:2608.13467;arXiv:2608.13484;arXiv:2608.13489;arXiv:2608.13499;arXiv:2608.13505;arXiv:2608.13515;arXiv:2608.13517;arXiv:2608.13538;arXiv:2608.13545;arXiv:2608.13546;arXiv:2608.13547;arXiv:2608.13552;arXiv:2608.13555;arXiv:2608.13558;arXiv:2608.13560;arXiv:2608.13567;arXiv:2608.13606;arXiv:2608.13622;arXiv:2608.13667;arXiv:2608.13760;arXiv:2608.13867;arXiv:2608.13900;arXiv:2608.13987;arXiv:2608.14022;arXiv:2608.14036;arXiv:2608.14054;arXiv:2608.14075;arXiv:2608.14144;arXiv:2608.14210;arXiv:2608.14221;arXiv:2608.14277;arXiv:2608.14284;arXiv:2608.14290;arXiv:2608.14312;arXiv:2608.14361;arXiv:2608.14377;arXiv:2608.14391;arXiv:2608.14457;arXiv:2608.14465;arXiv:2608.14530;arXiv:2608.14577;arXiv:2608.14680;arXiv:2608.14881;arXiv:2608.14905;arXiv:2608.15022;arXiv:2608.15045;arXiv:2608.15089;arXiv:2608.15265;arXiv:2608.15591;arXiv:2608.15659;arXiv:2608.15669;arXiv:2608.15763;arXiv:2608.15767;arXiv:2608.15869;arXiv:2608.15875;arXiv:2608.15888;arXiv:2608.15930;arXiv:2608.15984;arXiv:2608.16045;arXiv:2608.16072;arXiv:2608.16143;arXiv:2608.16157;arXiv:2608.16168;arXiv:2608.16319;arXiv:2608.16328;arXiv:2608.16515;arXiv:2608.16536;arXiv:2608.16590;arXiv:2608.16628;arXiv:2608.16721;arXiv:2608.16765;arXiv:2608.16776;arXiv:2608.16798;arXiv:2608.16812;arXiv:2608.16859;arXiv:2608.16885;arXiv:2608.16887;arXiv:2608.17050;arXiv:2608.17067;arXiv:2608.17253;arXiv:2608.17271;arXiv:2608.17310;arXiv:2608.17393;arXiv:2608.17423;arXiv:2608.17426;arXiv:2608.17512;arXiv:2608.17528;arXiv:2608.17536;arXiv:2608.17597;arXiv:2608.17682;arXiv:2608.17744;arXiv:2608.17781;arXiv:2608.17906;arXiv:2608.17950;arXiv:2608.17960;arXiv:2608.18027;arXiv:2608.18063;arXiv:2608.18092;arXiv:2608.18184;arXiv:2608.18484;arXiv:2608.18489;arXiv:2608.18524;arXiv:2608.18565;arXiv:2608.18580;arXiv:2608.18607;arXiv:2608.18613;arXiv:2608.18682;arXiv:2608.18852;arXiv:2608.18940;arXiv:2608.19098;arXiv:2608.19197;arXiv:2608.19269;arXiv:2608.19567;arXiv:2608.19583;arXiv:2608.19758;arXiv:2608.19799;arXiv:2608.19854;arXiv:2608.19857;arXiv:2608.19863;arXiv:2608.19880;arXiv:2608.19936;arXiv:2608.20169;arXiv:2608.20202;arXiv:2608.20246;arXiv:2608.20281;arXiv:2608.20318;arXiv:2608.20335;arXiv:2608.20336;arXiv:2608.20438;arXiv:2608.20492;arXiv:2608.20953;arXiv:2608.21031;arXiv:2608.21095;arXiv:2608.21140;arXiv:2608.21159;arXiv:2608.21208;arXiv:2608.21252;arXiv:2608.21281;arXiv:2608.21450;arXiv:2608.21486;arXiv:2608.21500;arXiv:2608.21832;arXiv:2608.21833;arXiv:2608.22274;arXiv:2608.22510;arXiv:2608.22752;arXiv:2608.22767;arXiv:2608.22872;arXiv:2608.22876;arXiv:2608.23001;arXiv:2608.23035;arXiv:2608.23041;arXiv:2608.23181;arXiv:2608.23189;arXiv:2608.23200;arXiv:2608.23252;arXiv:2608.23256;arXiv:2608.23259;arXiv:2608.23311;arXiv:2608.23392;arXiv:2608.23478;arXiv:2608.23552;arXiv:2608.23564;arXiv:2608.23670;arXiv:2608.23691;arXiv:2608.23740;arXiv:2608.23943;arXiv:2608.24040;arXiv:2608.24053;arXiv:2608.24358;arXiv:2608.24479;arXiv:2608.24569;arXiv:2608.24646;arXiv:2608.24680;arXiv:2608.24764;arXiv:2608.24777;arXiv:2608.24804;arXiv:2608.24876;arXiv:2608.25417;arXiv:2608.25500;arXiv:2608.25518;arXiv:2608.25593;arXiv:2608.25625;arXiv:2608.25798;arXiv:2608.26005;arXiv:2608.26070;arXiv:2608.26200;arXiv:2608.26238;arXiv:2608.26291;arXiv:2608.26530;arXiv:2608.26550;arXiv:2608.26623;arXiv:2608.26671;arXiv:2608.26730;arXiv:2608.26809;arXiv:2608.26872;arXiv:2608.26993;arXiv:2608.27123;arXiv:2608.27260;arXiv:2608.27345;arXiv:2608.27448;arXiv:2608.27455;arXiv:2608.27456;arXiv:2608.27763;arXiv:2608.27809;arXiv:2608.27831;arXiv:2608.27875;arXiv:2608.27969;arXiv:2608.28062;arXiv:2608.28122;arXiv:2608.28165;arXiv:2608.28192;arXiv:2608.28281;arXiv:2608.28363;arXiv:2608.28389;arXiv:2608.28444;arXiv:2608.28458;arXiv:2608.28460;arXiv:2608.28600;arXiv:2608.28833;arXiv:2608.29188;arXiv:2608.29310;arXiv:2608.29335;arXiv:2608.29387;arXiv:2608.29464;arXiv:2608.29681;arXiv:2608.29847;arXiv:2608.29974;arXiv:2608.30241;arXiv:2608.30320;arXiv:2608.30322;arXiv:2608.30391;arXiv:2608.30468;arXiv:2608.30478;arXiv:2608.30730;arXiv:2608.30795;arXiv:2608.30821;arXiv:2608.30935;arXiv:2608.30968;arXiv:2608.31005;arXiv:2608.31022;arXiv:2608.31036;arXiv:2608.31046;arXiv:2608.31082;arXiv:2608.31100;arXiv:2608.31106;arXiv:2608.31111;arXiv:2608.31119;arXiv:2609.00006;arXiv:2609.00028;arXiv:2609.00092;arXiv:2609.00111;arXiv:2609.00137;arXiv:2609.00188;arXiv:2609.00196;arXiv:2609.00365;arXiv:2609.00377;arXiv:2609.00474;arXiv:2609.00638;arXiv:2609.00749;arXiv:2609.00768;arXiv:2609.01147;arXiv:2609.01245;arXiv:2609.01281;arXiv:2609.01343;arXiv:2609.01404;arXiv:2609.01437;arXiv:2609.01481;arXiv:2609.01560;arXiv:2609.01572;arXiv:2609.01591;arXiv:2609.01601;arXiv:2609.01607;arXiv:2609.01657;arXiv:2609.01736;arXiv:2609.01740;arXiv:2609.01836;arXiv:2609.02094;arXiv:2609.02264;arXiv:2609.02737;arXiv:2609.02749;arXiv:2609.02750;arXiv:2609.02780;arXiv:2609.02783;arXiv:2609.02886;arXiv:2609.03153;arXiv:2609.03199;arXiv:2609.03241;arXiv:2609.03430;arXiv:2609.03493;arXiv:2609.03586;arXiv:2609.03756;arXiv:2609.03820;arXiv:2609.04010;arXiv:2609.04043;arXiv:2609.04094;arXiv:2609.04131;arXiv:2609.04148;arXiv:2609.04170;arXiv:2609.04199;arXiv:2609.04201;arXiv:2609.04250;arXiv:2609.04280;arXiv:2609.04304;arXiv:2609.04444;arXiv:2609.04482;arXiv:2609.04490;arXiv:2609.04523;arXiv:2609.04611;arXiv:2609.04753;arXiv:2609.04971;arXiv:2609.05232;arXiv:2609.05258;arXiv:2609.05339;arXiv:2609.05405;arXiv:2609.05416;arXiv:2609.05571;arXiv:2609.05661;arXiv:2609.05736;arXiv:2609.05903;arXiv:2609.06052;arXiv:2609.06107;arXiv:2609.06140;arXiv:2609.06702;arXiv:2609.06986;arXiv:2609.07064;arXiv:2609.08126;arXiv:2609.08149;arXiv:2609.08183;arXiv:2609.08365;arXiv:2609.08368;arXiv:2609.08572;arXiv:2609.08832;arXiv:2609.08887;arXiv:2609.08977;arXiv:2609.09072;arXiv:2609.09085;arXiv:2609.09090;arXiv:2609.09113;arXiv:2609.09143;arXiv:2609.09155;arXiv:2609.09156;arXiv:2609.09219;arXiv:2609.09264;arXiv:2609.09875;arXiv:2609.10016;arXiv:2609.10266;arXiv:2609.10522;arXiv:2609.10539;arXiv:2609.10706;arXiv:2609.10712;arXiv:2609.10715;arXiv:2609.10745;arXiv:2609.10895;arXiv:2609.11042;arXiv:2609.11085;arXiv:2609.11108;arXiv:2609.11155;arXiv:2609.11209;arXiv:2609.11294;arXiv:2609.11317;arXiv:2609.11318;arXiv:2609.11390;arXiv:2609.11412;arXiv:2609.11486;arXiv:2609.11561;arXiv:2609.11596;arXiv:2609.11638;arXiv:2609.11646;arXiv:2609.11744;arXiv:2609.11873;arXiv:2609.11929;arXiv:2609.11977;arXiv:2609.12397;arXiv:2609.12552;arXiv:2609.13141;arXiv:2609.13356;arXiv:2609.13406;arXiv:2609.13463;arXiv:2609.13814;arXiv:2609.14005;arXiv:2609.14011;arXiv:2609.14302;arXiv:2609.14306;arXiv:2609.14320;arXiv:2609.14462;arXiv:2609.14858;arXiv:2609.14973;arXiv:2609.15126;arXiv:2609.15195;arXiv:2609.15478;arXiv:2609.15504;arXiv:2609.15524;arXiv:2609.15578;arXiv:2609.15779;arXiv:2609.15800;arXiv:2609.15810;arXiv:2609.15818;arXiv:2609.15830;arXiv:2609.15863;arXiv:2609.15938;arXiv:2609.15973;arXiv:2609.15987;arXiv:2609.16034;arXiv:2609.16057;arXiv:2609.16251;arXiv:2609.16372;arXiv:2609.16591;arXiv:2609.16679;arXiv:2609.16816;arXiv:2609.16900;arXiv:2609.17172;arXiv:2609.17320;arXiv:2609.17391;arXiv:2609.17475;arXiv:2609.17488;arXiv:2609.17496;arXiv:2609.17523;arXiv:2609.17632;arXiv:2609.17652;arXiv:2609.17653;arXiv:2609.17708;arXiv:2609.17909;arXiv:2609.18063;arXiv:2609.18094;arXiv:2609.18182;arXiv:2609.18323;arXiv:2609.18487;arXiv:2609.18501;arXiv:2609.18605;arXiv:2609.18708;arXiv:2609.18779;arXiv:2609.18805;arXiv:2609.19134;arXiv:2609.19138;arXiv:2609.19143;arXiv:2609.19144;arXiv:2609.19169;arXiv:2609.19213;arXiv:2609.19315;arXiv:2609.19491;arXiv:2609.19499;arXiv:2609.19656;arXiv:2609.19671;arXiv:2609.19879;arXiv:2609.19969;arXiv:2609.20423;arXiv:2609.20489;arXiv:2609.20511;arXiv:2609.20519;arXiv:2609.20584;arXiv:2609.20612;arXiv:2609.20625;arXiv:2609.20715;arXiv:2609.20744;arXiv:2609.20784;arXiv:2609.20800;arXiv:2609.20804;arXiv:2609.20817;arXiv:2609.20869;arXiv:2609.20886;arXiv:2609.21346;arXiv:2609.21465;arXiv:2609.21788;arXiv:2609.22000;arXiv:2609.22039;arXiv:2609.22068;arXiv:2609.22076;arXiv:2609.22086;arXiv:2609.22682;arXiv:2609.22753;arXiv:2609.22947;arXiv:2609.22966;arXiv:2609.23038;arXiv:2609.23087;arXiv:2609.23088;arXiv:2609.23169;arXiv:2609.23407;arXiv:2609.23863;arXiv:2609.23989;arXiv:2609.24220;arXiv:2609.24308;arXiv:2609.24385;arXiv:2609.24432;arXiv:2609.24555;arXiv:2609.24657;arXiv:2609.24749;arXiv:2609.24788;arXiv:2609.24797;arXiv:2609.24972;arXiv:2609.24974;arXiv:2609.24981;arXiv:2609.24983;arXiv:2609.24984;arXiv:2609.24997;arXiv:2609.25001;arXiv:2609.25053;arXiv:2609.25165;arXiv:2609.25267;arXiv:2609.25270;arXiv:2609.25804;arXiv:2609.25853;arXiv:2609.25963;arXiv:2609.26355;arXiv:2609.26489;arXiv:2609.26634;arXiv:2609.26637;arXiv:2609.26780;arXiv:2609.27158;arXiv:2609.27308;arXiv:2609.27334;arXiv:2609.27657;arXiv:2609.28416;arXiv:2609.28466;arXiv:2609.28603;arXiv:2609.28654;arXiv:2609.28771;arXiv:2609.29028;arXiv:2609.29292;arXiv:2609.29362;arXiv:2609.29421;arXiv:2609.29429;arXiv:2609.29444;arXiv:2609.29647;arXiv:2609.29837;arXiv:2609.29845;arXiv:2609.29964;arXiv:2609.30221;arXiv:2609.30233;arXiv:2609.30243;arXiv:2609.30904;arXiv:2609.31002;arXiv:2609.31415;arXiv:2609.32019;arXiv:2609.32049;arXiv:2609.32241;arXiv:2609.32259;arXiv:2609.32540;arXiv:2609.32577;arXiv:2609.32600;arXiv:2609.32722;arXiv:2609.32810;arXiv:2609.32814;arXiv:2609.32965;arXiv:2609.32993;arXiv:2609.33203;arXiv:2609.33325;arXiv:2609.33378;arXiv:2609.33382;arXiv:2609.33485;arXiv:2609.33848;arXiv:2609.34227;arXiv:2609.34242;arXiv:2609.34274;arXiv:2609.34385;arXiv:2609.34392;arXiv:2609.34479;arXiv:2609.34563;arXiv:2609.34645;arXiv:2609.34771;arXiv:2609.35215;arXiv:2609.35228;arXiv:2609.35259;arXiv:2609.35261;arXiv:2609.35427;arXiv:2609.35457;arXiv:2609.35505;arXiv:2609.35530;arXiv:2609.35629;arXiv:2609.35673;arXiv:2609.35718;arXiv:2609.35734;arXiv:2609.35743;arXiv:2609.35767;arXiv:2609.35769;arXiv:2609.36138;arXiv:2609.36199;arXiv:2609.36314;arXiv:2609.36322;arXiv:2609.36388;arXiv:2609.36585;arXiv:2609.36636;arXiv:2609.36684;arXiv:2609.36691;arXiv:2609.36730;arXiv:2609.36820;arXiv:2609.37267;arXiv:2609.37334;arXiv:2609.37533;arXiv:2609.37559;arXiv:2609.37673;arXiv:2609.37690;arXiv:2609.37749;arXiv:2609.38334;arXiv:2609.38349;arXiv:2609.38660;arXiv:2609.38792;arXiv:2609.39075;arXiv:2609.39102;arXiv:2609.39154;arXiv:2609.39378;arXiv:2609.39420;arXiv:2609.39788;arXiv:2609.39841;arXiv:2609.39929;arXiv:2609.39982;arXiv:2609.40222;arXiv:2609.40362;arXiv:2610.00437;arXiv:2610.00544;arXiv:2610.00559;arXiv:2610.00574;arXiv:2610.00812;arXiv:2610.01092;arXiv:2610.01428;arXiv:2610.01762;arXiv:2610.01767;arXiv:2610.01871;arXiv:2610.01936;arXiv:2610.01939;arXiv:2610.02122;arXiv:2610.02150;arXiv:2610.02185;arXiv:2610.02196;arXiv:2610.02201;arXiv:2610.02206;arXiv:2610.02304;arXiv:2610.02331;arXiv:2610.02710;arXiv:2610.03020;arXiv:2610.03120;arXiv:2610.03140;arXiv:2610.03715;arXiv:2610.04116;arXiv:2610.04292;arXiv:2610.04596;arXiv:2610.04929;arXiv:2610.05033;arXiv:2610.05076;arXiv:2610.05107;arXiv:2610.05162;arXiv:2610.05622;arXiv:2610.05709;arXiv:2610.05782;arXiv:2610.05826;arXiv:2610.05842;arXiv:2610.05912;arXiv:2610.06207;arXiv:2610.06293;arXiv:2610.06479;arXiv:2610.06666;arXiv:2610.07250;arXiv:2610.07332;arXiv:2610.07557;arXiv:2610.07753;arXiv:2610.07862;arXiv:2610.08048;arXiv:2610.08077;arXiv:2610.08432;arXiv:2610.08448;arXiv:2610.08452;arXiv:2610.08463;arXiv:2610.08571;arXiv:2610.08630;arXiv:2610.08674;arXiv:2610.08995;arXiv:2610.09127;arXiv:2610.09228;arXiv:2610.09684;arXiv:2610.09832;arXiv:2610.10091;arXiv:2610.10164;arXiv:2610.10507;arXiv:2610.10533;arXiv:2610.10845;arXiv:2610.22753
v113 CVE 编号
CVE-2025-49596; CVE-2026-22773; CVE-2026-22778; CVE-2026-26029; CVE-2026-3059; CVE-2026-3060; CVE-2026-30623; CVE-2026-3172; CVE-2026-33626; CVE-2026-3989; CVE-2026-4372; CVE-2026-5059; CVE-2026-5241; CVE-2026-54769; CVE-2026-57572; CVE-2026-59726; CVE-2026-61447; CVE-2026-61539
v113 DOI 编号
DOI: 10.5281/zenodo.20953922
v113 URL 锚点清单(沿用 v111 全部 + v112/v113 净增)
https://academy.dair.ai/papers/neohorse-1-towards-recursive-self-improvement-via-agentic-post-training-with-rou-2609.08183 https://adg.csdn.net/6a68c5bd662f9a54cb956060.html https://aishwaryasrinivasan.substack.com/p/all-you-need-to-know-about-rag-in https://aixfunda.substack.com https://aixfunda.substack.com/p/top-llm-rag-and-agent-updates-of-32f https://alphacorp.ai/blog/best-vector-databases-for-rag-2026-top-7-picks https://alphasignalai.substack.com https://alphasignalai.substack.com/p/d-mem-agent-memory-third-path https://alphasignalai.substack.com/p/rag-and-long-context-arent-enough https://arxiv.org/abs/2502.05167 https://arxiv.org/abs/2502.08826 https://arxiv.org/abs/2504.12330 https://arxiv.org/abs/2506.01716 https://arxiv.org/abs/2508.08137 https://arxiv.org/abs/2510.10991 https://arxiv.org/abs/2510.11541v2 https://arxiv.org/abs/2510.15253 https://arxiv.org/abs/2601.04043 https://arxiv.org/abs/2601.21557 https://arxiv.org/abs/2602.06052 https://arxiv.org/abs/2602.18998v1 https://arxiv.org/abs/2602.23166 https://arxiv.org/abs/2603.06569 https://arxiv.org/abs/2603.07379v1 https://arxiv.org/abs/2603.21489 https://arxiv.org/abs/2603.27918 https://arxiv.org/abs/2603.28052 https://arxiv.org/abs/2603.29231 https://arxiv.org/abs/2604.16548 https://arxiv.org/abs/2604.16548v1 https://arxiv.org/abs/2605.03310 https://arxiv.org/abs/2605.13831 https://arxiv.org/abs/2605.15040 https://arxiv.org/abs/2605.26112 https://arxiv.org/abs/2605.29640 https://arxiv.org/abs/2605.29640v1 https://arxiv.org/abs/2605.30434 https://arxiv.org/abs/2606.05608 https://arxiv.org/abs/2607.08028 https://arxiv.org/abs/2607.08964 https://arxiv.org/abs/2607.12227 https://arxiv.org/abs/2607.14277 https://arxiv.org/abs/2607.17715 https://arxiv.org/abs/2607.19712 https://arxiv.org/abs/2607.19793 https://arxiv.org/abs/2607.20346 https://arxiv.org/abs/2607.20368 https://arxiv.org/abs/2607.20379 https://arxiv.org/abs/2607.20709 https://arxiv.org/abs/2607.20891 https://arxiv.org/abs/2607.20911 https://arxiv.org/abs/2607.21461 https://arxiv.org/abs/2607.21557 https://arxiv.org/abs/2607.21576 https://arxiv.org/abs/2607.21596 https://arxiv.org/abs/2607.21653 https://arxiv.org/abs/2607.22043 https://arxiv.org/abs/2607.22157 https://arxiv.org/abs/2607.22375 https://arxiv.org/abs/2607.22448 https://arxiv.org/abs/2607.22682 https://arxiv.org/abs/2607.23193 https://arxiv.org/abs/2607.23693 https://arxiv.org/abs/2607.24117 https://arxiv.org/abs/2607.24223 https://arxiv.org/abs/2607.24653 https://arxiv.org/abs/2607.24882 https://arxiv.org/abs/2607.25236 https://arxiv.org/abs/2607.25308 https://arxiv.org/abs/2607.25379v1 https://arxiv.org/abs/2607.25380 https://arxiv.org/abs/2607.25398v1 https://arxiv.org/abs/2607.25431 https://arxiv.org/abs/2607.25565 https://arxiv.org/abs/2607.25600v1 https://arxiv.org/abs/2607.25886 https://arxiv.org/abs/2607.25959v1 https://arxiv.org/abs/2607.25996v1 https://arxiv.org/abs/2607.26314 https://arxiv.org/abs/2607.26410 https://arxiv.org/abs/2607.26451 https://arxiv.org/abs/2607.26520 https://arxiv.org/abs/2607.26611 https://arxiv.org/abs/2607.26637 https://arxiv.org/abs/2607.26654 https://arxiv.org/abs/2607.26760 https://arxiv.org/abs/2607.26784 https://arxiv.org/abs/2607.26991 https://arxiv.org/abs/2607.27146 https://arxiv.org/abs/2607.27167 https://arxiv.org/abs/2607.27201 https://arxiv.org/abs/2607.27506 https://arxiv.org/abs/2607.27851 https://arxiv.org/abs/2607.27888 https://arxiv.org/abs/2607.27919 https://arxiv.org/abs/2607.27958 https://arxiv.org/abs/2607.28126 https://arxiv.org/abs/2607.28227 https://arxiv.org/abs/2607.28229 https://arxiv.org/abs/2607.28263 https://arxiv.org/abs/2607.28397 https://arxiv.org/abs/2607.28568 https://arxiv.org/abs/2607.28580 https://arxiv.org/abs/2607.28618 https://arxiv.org/abs/2607.28624 https://arxiv.org/abs/2607.28802 https://arxiv.org/abs/2607.29167 https://arxiv.org/abs/2607.29209 https://arxiv.org/abs/2607.29402 https://arxiv.org/abs/2607.29459 https://arxiv.org/abs/2607.29610 https://arxiv.org/abs/2607.29677 https://arxiv.org/abs/2607.29679 https://arxiv.org/abs/2608.00079 https://arxiv.org/abs/2608.00267 https://arxiv.org/abs/2608.00650 https://arxiv.org/abs/2608.00675 https://arxiv.org/abs/2608.00730 https://arxiv.org/abs/2608.00799 https://arxiv.org/abs/2608.00922 https://arxiv.org/abs/2608.01185 https://arxiv.org/abs/2608.01481 https://arxiv.org/abs/2608.01526 https://arxiv.org/abs/2608.01628 https://arxiv.org/abs/2608.01678 https://arxiv.org/abs/2608.01735 https://arxiv.org/abs/2608.01851 https://arxiv.org/abs/2608.01964 https://arxiv.org/abs/2608.02218 https://arxiv.org/abs/2608.02437 https://arxiv.org/abs/2608.02583 https://arxiv.org/abs/2608.02703 https://arxiv.org/abs/2608.02713 https://arxiv.org/abs/2608.02738 https://arxiv.org/abs/2608.02870 https://arxiv.org/abs/2608.03392 https://arxiv.org/abs/2608.03451 https://arxiv.org/abs/2608.03744 https://arxiv.org/abs/2608.03764 https://arxiv.org/abs/2608.03836 https://arxiv.org/abs/2608.03874 https://arxiv.org/abs/2608.03972 https://arxiv.org/abs/2608.03974 https://arxiv.org/abs/2608.03979 https://arxiv.org/abs/2608.04003 https://arxiv.org/abs/2608.04205 https://arxiv.org/abs/2608.04397 https://arxiv.org/abs/2608.04530 https://arxiv.org/abs/2608.05102 https://arxiv.org/abs/2608.05108 https://arxiv.org/abs/2608.05139 https://arxiv.org/abs/2608.05224 https://arxiv.org/abs/2608.05248 https://arxiv.org/abs/2608.05466 https://arxiv.org/abs/2608.05604 https://arxiv.org/abs/2608.05703 https://arxiv.org/abs/2608.05747 https://arxiv.org/abs/2608.05784 https://arxiv.org/abs/2608.05798 https://arxiv.org/abs/2608.05987 https://arxiv.org/abs/2608.06020 https://arxiv.org/abs/2608.06060 https://arxiv.org/abs/2608.06113 https://arxiv.org/abs/2608.06130 https://arxiv.org/abs/2608.06197 https://arxiv.org/abs/2608.06216 https://arxiv.org/abs/2608.06257 https://arxiv.org/abs/2608.06270 https://arxiv.org/abs/2608.06296 https://arxiv.org/abs/2608.06301 https://arxiv.org/abs/2608.06352 https://arxiv.org/abs/2608.06485 https://arxiv.org/abs/2608.06501 https://arxiv.org/abs/2608.06865 https://arxiv.org/abs/2608.06867 https://arxiv.org/abs/2608.07009 https://arxiv.org/abs/2608.07051 https://arxiv.org/abs/2608.07126 https://arxiv.org/abs/2608.07152 https://arxiv.org/abs/2608.07169 https://arxiv.org/abs/2608.07193 https://arxiv.org/abs/2608.07370 https://arxiv.org/abs/2608.07446 https://arxiv.org/abs/2608.07458 https://arxiv.org/abs/2608.07468 https://arxiv.org/abs/2608.07545 https://arxiv.org/abs/2608.07565 https://arxiv.org/abs/2608.07594 https://arxiv.org/abs/2608.07645 https://arxiv.org/abs/2608.08020 https://arxiv.org/abs/2608.08097 https://arxiv.org/abs/2608.08160 https://arxiv.org/abs/2608.08311 https://arxiv.org/abs/2608.08466 https://arxiv.org/abs/2608.08612 https://arxiv.org/abs/2608.08621 https://arxiv.org/abs/2608.08627 https://arxiv.org/abs/2608.08676 https://arxiv.org/abs/2608.08722 https://arxiv.org/abs/2608.08814 https://arxiv.org/abs/2608.08975 https://arxiv.org/abs/2608.09096 https://arxiv.org/abs/2608.09119 https://arxiv.org/abs/2608.09158 https://arxiv.org/abs/2608.09209 https://arxiv.org/abs/2608.09250 https://arxiv.org/abs/2608.09408 https://arxiv.org/abs/2608.09802 https://arxiv.org/abs/2608.09853 https://arxiv.org/abs/2608.09867 https://arxiv.org/abs/2608.09873 https://arxiv.org/abs/2608.09888 https://arxiv.org/abs/2608.09897 https://arxiv.org/abs/2608.09900 https://arxiv.org/abs/2608.09928 https://arxiv.org/abs/2608.10299 https://arxiv.org/abs/2608.10628 https://arxiv.org/abs/2608.10636 https://arxiv.org/abs/2608.10708 https://arxiv.org/abs/2608.10720 https://arxiv.org/abs/2608.10744 https://arxiv.org/abs/2608.10835 https://arxiv.org/abs/2608.10875 https://arxiv.org/abs/2608.10915 https://arxiv.org/abs/2608.11030 https://arxiv.org/abs/2608.11123 https://arxiv.org/abs/2608.11205 https://arxiv.org/abs/2608.11341 https://arxiv.org/abs/2608.11350 https://arxiv.org/abs/2608.11367 https://arxiv.org/abs/2608.11632 https://arxiv.org/abs/2608.11660 https://arxiv.org/abs/2608.11745 https://arxiv.org/abs/2608.11752 https://arxiv.org/abs/2608.11924 https://arxiv.org/abs/2608.11947 https://arxiv.org/abs/2608.12036 https://arxiv.org/abs/2608.12123 https://arxiv.org/abs/2608.12149 https://arxiv.org/abs/2608.12218 https://arxiv.org/abs/2608.12278 https://arxiv.org/abs/2608.12304 https://arxiv.org/abs/2608.12307 https://arxiv.org/abs/2608.12313 https://arxiv.org/abs/2608.12314 https://arxiv.org/abs/2608.12440 https://arxiv.org/abs/2608.12571 https://arxiv.org/abs/2608.12700 https://arxiv.org/abs/2608.12743 https://arxiv.org/abs/2608.12845 https://arxiv.org/abs/2608.12875 https://arxiv.org/abs/2608.12986 https://arxiv.org/abs/2608.12987 https://arxiv.org/abs/2608.13010 https://arxiv.org/abs/2608.13040 https://arxiv.org/abs/2608.13120 https://arxiv.org/abs/2608.13122 https://arxiv.org/abs/2608.13160 https://arxiv.org/abs/2608.13210 https://arxiv.org/abs/2608.13237 https://arxiv.org/abs/2608.13263 https://arxiv.org/abs/2608.13384 https://arxiv.org/abs/2608.13391 https://arxiv.org/abs/2608.13410 https://arxiv.org/abs/2608.13417 https://arxiv.org/abs/2608.13426 https://arxiv.org/abs/2608.13467 https://arxiv.org/abs/2608.13484 https://arxiv.org/abs/2608.13489 https://arxiv.org/abs/2608.13499 https://arxiv.org/abs/2608.13505 https://arxiv.org/abs/2608.13515 https://arxiv.org/abs/2608.13517 https://arxiv.org/abs/2608.13538 https://arxiv.org/abs/2608.13545 https://arxiv.org/abs/2608.13546 https://arxiv.org/abs/2608.13547 https://arxiv.org/abs/2608.13552 https://arxiv.org/abs/2608.13555 https://arxiv.org/abs/2608.13558 https://arxiv.org/abs/2608.13560 https://arxiv.org/abs/2608.13567 https://arxiv.org/abs/2608.13606 https://arxiv.org/abs/2608.13622 https://arxiv.org/abs/2608.13667 https://arxiv.org/abs/2608.13760 https://arxiv.org/abs/2608.13900 https://arxiv.org/abs/2608.13987 https://arxiv.org/abs/2608.14022 https://arxiv.org/abs/2608.14036 https://arxiv.org/abs/2608.14054 https://arxiv.org/abs/2608.14075 https://arxiv.org/abs/2608.14144 https://arxiv.org/abs/2608.14210 https://arxiv.org/abs/2608.14221 https://arxiv.org/abs/2608.14277 https://arxiv.org/abs/2608.14284 https://arxiv.org/abs/2608.14290 https://arxiv.org/abs/2608.14312 https://arxiv.org/abs/2608.14361 https://arxiv.org/abs/2608.14377 https://arxiv.org/abs/2608.14391 https://arxiv.org/abs/2608.14457 https://arxiv.org/abs/2608.14465 https://arxiv.org/abs/2608.14530 https://arxiv.org/abs/2608.14577 https://arxiv.org/abs/2608.14881 https://arxiv.org/abs/2608.14905 https://arxiv.org/abs/2608.15022 https://arxiv.org/abs/2608.15045 https://arxiv.org/abs/2608.15089 https://arxiv.org/abs/2608.15265 https://arxiv.org/abs/2608.15659 https://arxiv.org/abs/2608.15669 https://arxiv.org/abs/2608.15763 https://arxiv.org/abs/2608.15767 https://arxiv.org/abs/2608.15869 https://arxiv.org/abs/2608.15875 https://arxiv.org/abs/2608.15888 https://arxiv.org/abs/2608.15930 https://arxiv.org/abs/2608.15984 https://arxiv.org/abs/2608.16045 https://arxiv.org/abs/2608.16072 https://arxiv.org/abs/2608.16143 https://arxiv.org/abs/2608.16157 https://arxiv.org/abs/2608.16168 https://arxiv.org/abs/2608.16319 https://arxiv.org/abs/2608.16328 https://arxiv.org/abs/2608.16515 https://arxiv.org/abs/2608.16536 https://arxiv.org/abs/2608.16590 https://arxiv.org/abs/2608.16628 https://arxiv.org/abs/2608.16721 https://arxiv.org/abs/2608.16765 https://arxiv.org/abs/2608.16776 https://arxiv.org/abs/2608.16798 https://arxiv.org/abs/2608.16859 https://arxiv.org/abs/2608.16885 https://arxiv.org/abs/2608.16887 https://arxiv.org/abs/2608.17050 https://arxiv.org/abs/2608.17067 https://arxiv.org/abs/2608.17253 https://arxiv.org/abs/2608.17271 https://arxiv.org/abs/2608.17310 https://arxiv.org/abs/2608.17393 https://arxiv.org/abs/2608.17426 https://arxiv.org/abs/2608.17512 https://arxiv.org/abs/2608.17528 https://arxiv.org/abs/2608.17536 https://arxiv.org/abs/2608.17597 https://arxiv.org/abs/2608.17682 https://arxiv.org/abs/2608.17744 https://arxiv.org/abs/2608.17781 https://arxiv.org/abs/2608.17906 https://arxiv.org/abs/2608.17950 https://arxiv.org/abs/2608.17960 https://arxiv.org/abs/2608.18027 https://arxiv.org/abs/2608.18063 https://arxiv.org/abs/2608.18077 https://arxiv.org/abs/2608.18092 https://arxiv.org/abs/2608.18184 https://arxiv.org/abs/2608.18484 https://arxiv.org/abs/2608.18489 https://arxiv.org/abs/2608.18524 https://arxiv.org/abs/2608.18565 https://arxiv.org/abs/2608.18580 https://arxiv.org/abs/2608.18607 https://arxiv.org/abs/2608.18613 https://arxiv.org/abs/2608.18852 https://arxiv.org/abs/2608.18940 https://arxiv.org/abs/2608.19098 https://arxiv.org/abs/2608.19197 https://arxiv.org/abs/2608.19567 https://arxiv.org/abs/2608.19758 https://arxiv.org/abs/2608.19799 https://arxiv.org/abs/2608.19854 https://arxiv.org/abs/2608.19857 https://arxiv.org/abs/2608.19863 https://arxiv.org/abs/2608.19880 https://arxiv.org/abs/2608.19936 https://arxiv.org/abs/2608.20169 https://arxiv.org/abs/2608.20202 https://arxiv.org/abs/2608.20246 https://arxiv.org/abs/2608.20281 https://arxiv.org/abs/2608.20335 https://arxiv.org/abs/2608.20336 https://arxiv.org/abs/2608.20438 https://arxiv.org/abs/2608.20492 https://arxiv.org/abs/2608.20953 https://arxiv.org/abs/2608.21031 https://arxiv.org/abs/2608.21159 https://arxiv.org/abs/2608.21208 https://arxiv.org/abs/2608.21252 https://arxiv.org/abs/2608.21450 https://arxiv.org/abs/2608.21486 https://arxiv.org/abs/2608.21500 https://arxiv.org/abs/2608.21833 https://arxiv.org/abs/2608.22274 https://arxiv.org/abs/2608.22510 https://arxiv.org/abs/2608.22752 https://arxiv.org/abs/2608.22872 https://arxiv.org/abs/2608.22876 https://arxiv.org/abs/2608.23001 https://arxiv.org/abs/2608.23035 https://arxiv.org/abs/2608.23041 https://arxiv.org/abs/2608.23181 https://arxiv.org/abs/2608.23189 https://arxiv.org/abs/2608.23200 https://arxiv.org/abs/2608.23252 https://arxiv.org/abs/2608.23256 https://arxiv.org/abs/2608.23259 https://arxiv.org/abs/2608.23311 https://arxiv.org/abs/2608.23392 https://arxiv.org/abs/2608.23552 https://arxiv.org/abs/2608.24040 https://arxiv.org/abs/2608.24053 https://arxiv.org/abs/2608.24646 https://arxiv.org/abs/2608.24680 https://arxiv.org/abs/2608.24764 https://arxiv.org/abs/2608.24777 https://arxiv.org/abs/2608.24804 https://arxiv.org/abs/2608.24876 https://arxiv.org/abs/2608.25417 https://arxiv.org/abs/2608.25500 https://arxiv.org/abs/2608.25593 https://arxiv.org/abs/2608.25625 https://arxiv.org/abs/2608.26070 https://arxiv.org/abs/2608.26238 https://arxiv.org/abs/2608.26530 https://arxiv.org/abs/2608.26550 https://arxiv.org/abs/2608.26623 https://arxiv.org/abs/2608.26730 https://arxiv.org/abs/2608.27123 https://arxiv.org/abs/2608.27260 https://arxiv.org/abs/2608.27345 https://arxiv.org/abs/2608.27448 https://arxiv.org/abs/2608.27809 https://arxiv.org/abs/2608.27831 https://arxiv.org/abs/2608.27969 https://arxiv.org/abs/2608.28122 https://arxiv.org/abs/2608.28165 https://arxiv.org/abs/2608.28281 https://arxiv.org/abs/2608.28363 https://arxiv.org/abs/2608.28389 https://arxiv.org/abs/2608.28444 https://arxiv.org/abs/2608.28458 https://arxiv.org/abs/2608.30322 https://arxiv.org/abs/2608.30391 https://arxiv.org/abs/2608.30468 https://arxiv.org/abs/2608.30730 https://arxiv.org/abs/2608.30935 https://arxiv.org/abs/2608.31082 https://arxiv.org/abs/2608.31100 https://arxiv.org/abs/2608.31111 https://arxiv.org/abs/2609.00006 https://arxiv.org/abs/2609.00196 https://arxiv.org/abs/2609.00377 https://arxiv.org/abs/2609.00474 https://arxiv.org/abs/2609.00638 https://arxiv.org/abs/2609.01147 https://arxiv.org/abs/2609.01281 https://arxiv.org/abs/2609.01404 https://arxiv.org/abs/2609.01437 https://arxiv.org/abs/2609.01560 https://arxiv.org/abs/2609.01591 https://arxiv.org/abs/2609.01607 https://arxiv.org/abs/2609.01607t https://arxiv.org/abs/2609.01657 https://arxiv.org/abs/2609.01740 https://arxiv.org/abs/2609.01836 https://arxiv.org/abs/2609.02737 https://arxiv.org/abs/2609.02749 https://arxiv.org/abs/2609.02750 https://arxiv.org/abs/2609.02783 https://arxiv.org/abs/2609.02886 https://arxiv.org/abs/2609.03153 https://arxiv.org/abs/2609.03199 https://arxiv.org/abs/2609.03241 https://arxiv.org/abs/2609.03820 https://arxiv.org/abs/2609.04010 https://arxiv.org/abs/2609.04043 https://arxiv.org/abs/2609.04094 https://arxiv.org/abs/2609.04148 https://arxiv.org/abs/2609.04199 https://arxiv.org/abs/2609.04201 https://arxiv.org/abs/2609.04250 https://arxiv.org/abs/2609.04280 https://arxiv.org/abs/2609.04304 https://arxiv.org/abs/2609.04444 https://arxiv.org/abs/2609.04482 https://arxiv.org/abs/2609.04523 https://arxiv.org/abs/2609.04611 https://arxiv.org/abs/2609.04753 https://arxiv.org/abs/2609.04971 https://arxiv.org/abs/2609.05339 https://arxiv.org/abs/2609.05416 https://arxiv.org/abs/2609.05661 https://arxiv.org/abs/2609.05903 https://arxiv.org/abs/2609.06140 https://arxiv.org/abs/2609.06702 https://arxiv.org/abs/2609.06986 https://arxiv.org/abs/2609.08126 https://arxiv.org/abs/2609.08149 https://arxiv.org/abs/2609.08183 https://arxiv.org/abs/2609.08365 https://arxiv.org/abs/2609.08368 https://arxiv.org/abs/2609.08572 https://arxiv.org/abs/2609.08832v1 https://arxiv.org/abs/2609.08887 https://arxiv.org/abs/2609.08977 https://arxiv.org/abs/2609.09072 https://arxiv.org/abs/2609.09085 https://arxiv.org/abs/2609.09090 https://arxiv.org/abs/2609.09113 https://arxiv.org/abs/2609.09156 https://arxiv.org/abs/2609.09219 https://arxiv.org/abs/2609.09264 https://arxiv.org/abs/2609.09875 https://arxiv.org/abs/2609.10266 https://arxiv.org/abs/2609.10522 https://arxiv.org/abs/2609.10745 https://arxiv.org/abs/2609.11042 https://arxiv.org/abs/2609.11108 https://arxiv.org/abs/2609.11155 https://arxiv.org/abs/2609.11209 https://arxiv.org/abs/2609.11294 https://arxiv.org/abs/2609.11318 https://arxiv.org/abs/2609.11390 https://arxiv.org/abs/2609.11561 https://arxiv.org/abs/2609.11596 https://arxiv.org/abs/2609.11646 https://arxiv.org/abs/2609.11873 https://arxiv.org/abs/2609.11929 https://arxiv.org/abs/2609.12397 https://arxiv.org/abs/2609.12552 https://arxiv.org/abs/2609.13406 https://arxiv.org/abs/2609.13770 https://arxiv.org/abs/2609.14005 https://arxiv.org/abs/2609.14011 https://arxiv.org/abs/2609.14803 https://arxiv.org/abs/2609.14857 https://arxiv.org/abs/2609.15195 https://arxiv.org/abs/2609.15578 https://arxiv.org/abs/2609.15810 https://arxiv.org/abs/2609.15972 https://arxiv.org/abs/2609.16034 https://arxiv.org/abs/2609.16057 https://arxiv.org/abs/2609.16372 https://arxiv.org/abs/2609.16591 https://arxiv.org/abs/2609.16679 https://arxiv.org/abs/2609.16816 https://arxiv.org/abs/2609.16818 https://arxiv.org/abs/2609.17012 https://arxiv.org/abs/2609.17172 https://arxiv.org/abs/2609.17320 https://arxiv.org/abs/2609.17488 https://arxiv.org/abs/2609.17496 https://arxiv.org/abs/2609.17521 https://arxiv.org/abs/2609.17523 https://arxiv.org/abs/2609.17524 https://arxiv.org/abs/2609.17632 https://arxiv.org/abs/2609.17653 https://arxiv.org/abs/2609.18323 https://arxiv.org/abs/2609.18605 https://arxiv.org/abs/2609.19499 https://arxiv.org/abs/2609.19656 https://arxiv.org/abs/2609.19671 https://arxiv.org/abs/2609.19879 https://arxiv.org/abs/2609.19969 https://arxiv.org/abs/2609.20423 https://arxiv.org/abs/2609.20511 https://arxiv.org/abs/2609.20519 https://arxiv.org/abs/2609.20612 https://arxiv.org/abs/2609.20715 https://arxiv.org/abs/2609.20784 https://arxiv.org/abs/2609.20800 https://arxiv.org/abs/2609.20804 https://arxiv.org/abs/2609.20817 https://arxiv.org/abs/2609.37334 https://arxiv.org/abs/2610.04596 https://arxiv.org/abs/2610.05912 https://arxiv.org/abs/2610.06293 https://arxiv.org/abs/2610.07250 https://arxiv.org/abs/2610.07332 https://arxiv.org/abs/2610.07557 https://arxiv.org/abs/2610.07753 https://arxiv.org/abs/2610.07862 https://arxiv.org/abs/2610.08048 https://arxiv.org/abs/2610.08077 https://arxiv.org/abs/2610.08432 https://arxiv.org/abs/2610.08448 https://arxiv.org/abs/2610.08452 https://arxiv.org/abs/2610.08463 https://arxiv.org/abs/2610.08571 https://arxiv.org/abs/2610.08630 https://arxiv.org/abs/2610.08674 https://arxiv.org/abs/2610.08995 https://arxiv.org/abs/2610.09127 https://arxiv.org/abs/2610.09228 https://arxiv.org/abs/2610.09684 https://arxiv.org/abs/2610.09832 https://arxiv.org/abs/2610.10091 https://arxiv.org/abs/2610.10164 https://arxiv.org/abs/2610.10507 https://arxiv.org/abs/2610.10533 https://arxiv.org/abs/2610.10845 https://arxiv.org/html/2510.11541v2 https://arxiv.org/html/2602.18998v1 https://arxiv.org/html/2603.07379v1 https://arxiv.org/html/2604.16548v1 https://arxiv.org/html/2605.29640v1 https://arxiv.org/html/2608.23552 https://autosaddler-projectpage.github.io https://bensbites.com/p/1-billion-chatgpt-users https://bensbites.com/p/opus-5-fable-5 https://blog.bytebytego.com/p/ep223-ollama-vs-vllm-vs-sglang https://blog.bytebytego.com/p/how-to-make-llms-3x-faster https://blog.bytebytego.com/p/how-to-steal-an-ai-models-private https://blog.cloudflare.com/the-agent-access-model https://blog.csdn.net/2301_81666833/article/details/163924904 https://blog.csdn.net/adg.csdn.net/6a87be40662f9a54cb9f189f.html https://blog.csdn.net/m0_73407070/article/details/163434217 https://blog.csdn.net/weixin_47547625/article/details/161535116 https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash https://blog.google/innovation-and-ai/technology/ai/ai-applications-science-people/ https://blog.google/innovation-and-ai/technology/ai/ai-economy-atlas-september-2026/ https://blog.google/innovation-and-ai/technology/ai/ai-for-every-language/ https://blog.google/innovation-and-ai/technology/ai/ai-for-societal-impact/ https://blog.google/innovation-and-ai/technology/developers-tools/ai-agents-intensive-recap-2026 https://blog.google/innovation-and-ai/technology/developers-tools/expanding-managed-agent-ai-api-3-6-flash-hooks https://blog.google/innovation-and-ai/technology/developers-tools/expanding-managed-agents-gemini-api-3-6-flash-hooks https://blog.modelcontextprotocol.io/posts/2026-07-28 https://buttondown.com/ai-tldr/archive/aitldr-daily-digest-august-16-2026 https://bytedtsinghua-sia.github.io/Open-MOPD https://cameronrwolfe.substack.com/p/agent-evals https://cameronrwolfe.substack.com/p/agentic-world-models https://cctest.ai/en/articles/neohorse-1-turns-agent-routing-into-a-path-to-recursive-self-improvement https://claudiostamile.substack.com/p/agent-memory-is-not-rag-a-practical https://codingwithroby.substack.com/p/the-2026-ai-agent-stack-drawn-from https://commandcode.ai/guides/rag-in-2026 https://deepmind.google/blog https://deepmind.google/blog/alphagenome-atlas-a-predictive-map-of-every-possible-dna-letter-change-in-the-human-genome/ https://deepmind.google/blog/introducing-gemini-3-8-live-and-3-8-live-extended-thinking/ https://deepmind.google/blog/introducing-gemini-38-live-with-live-avatar/ https://devblogs.microsoft.com/agent-framework https://envharness.com https://eventual.ai/blog/vlas-are-dead-long-live-world-action-models https://explainx.ai/blog/anthropic-claude-code-limits-17-percent-cut-september-2026-august-2026 https://futureagi.substack.com/p/the-definitive-guide-to-ai-agent https://futuresearch.ai/blog/google-deepmind-reorg-forecast https://futuressearch.ai/blog/google-deepmind-reorg-forecast https://github.com/AI-MO/SAF-OPD https://github.com/AMAP-ML/LongHorizon-Harness https://github.com/ARUNAGIRINATHAN-K/awesome-ai-agents-2026 https://github.com/Ayanami0730/arag https://github.com/DavidZWZ/Awesome-RAG-Reasoning https://github.com/HJYao00/Awesome-Agentic-MLLMs https://github.com/HKBU-KnowComp/VisualPatchWorld/ https://github.com/HKUDS/MentalWorldModeling https://github.com/HKUSTDial/DataSpace https://github.com/InternLM/lmdeploy https://github.com/JayLZhou/Awesome-Agent-Skills https://github.com/Kangarooking/cangjie-skill https://github.com/Liquid4All/antidoom https://github.com/MemTensor/Metis https://github.com/NirDiamant/agents-towards-production https://github.com/NousResearch/hermes-agent https://github.com/Panniantong/Agent-Reach https://github.com/QwenLM/Qwen-Agent https://github.com/RyanAlberts/best-of-Agent-Harnesses https://github.com/Shubhamsaboo/awesome-llm-apps https://github.com/THU-Team-Eureka/EurekAgent https://github.com/Tencent-Hunyuan/AuK https://github.com/TencentCloud/TencentDB-Agent-Memory https://github.com/addyosmani/agent-skills https://github.com/ai-boost/awesome-harness-engineering https://github.com/alibaba/zvec https://github.com/anomalyco/opencode https://github.com/antirez/ds4 https://github.com/b1ade-project/b1ade https://github.com/cactus-compute/needle https://github.com/caramaschiHG/awesome-ai-agents-2026 https://github.com/citrolabs/ego-lite https://github.com/cloudflare/security-audit-skill https://github.com/coder/coder https://github.com/criptogus/HermesOffice https://github.com/cxcscmu/General-AgentBench https://github.com/cyberlife-coder/VelesDB https://github.com/datawhalechina/hello-agents https://github.com/deepagents-ai/agent-backend https://github.com/different-ai/openwork https://github.com/docling-project/docling https://github.com/fuxicodex/Fuxi https://github.com/github/spec-kit https://github.com/google-research/envharness https://github.com/google/skills https://github.com/headroom-ai/headroom https://github.com/higgsfield-ai/higgsfield https://github.com/holaboss-ai/holaOS https://github.com/infiniflow/ragflow https://github.com/jayzhou/Awesome-Agent-Skills https://github.com/langchain-ai/deepagents https://github.com/llm-lab-org/Agentic-RAG-Survey https://github.com/llm-lab-org/multimodal-rag-survey https://github.com/lm-sys/FastChat https://github.com/lyogavin/airllm https://github.com/mattpocock/skills https://github.com/mem0ai/mem0 https://github.com/microsoft/AutoSaddler https://github.com/microsoft/agent-governance-toolkit https://github.com/microsoft/agent-lightning https://github.com/microsoft/ai-agents-for-beginners https://github.com/midea-ai/SemaPLC https://github.com/modelcontextprotocol https://github.com/openJiuwen-ai/community https://github.com/opensandbox-group/OpenSandbox https://github.com/rethinking-harness-evolution https://github.com/run-llama/ExtractBench https://github.com/safishamsi/graphify https://github.com/tracer-cloud/opensre https://github.com/trycua/cua https://github.com/vllm-project/vllm https://github.com/volcengine/OpenViking https://github.com/yc-software/qm https://github.com/yongjoopark/SlotGuard https://github.com/zilliztech/deep-searcher https://gradientflow.com/ai-in-math-2026-08/ https://gradientflow.com/chinas-ai-involution/ https://gradientflow.com/continual-learning-is-arriving-in-pieces/ https://gradientflow.com/hbf-ai-inference/ https://gradientflow.com/i-think-ai-teams-are-defending-the-wrong-thing/ https://gradientflow.com/i-think-we-are-looking-for-ai-risk-in-the-wrong-place/ https://gradientflow.com/nine-practical-rules-for-agents-doing-real-work/ https://gradientflow.com/self-improvement-without-the-science-fiction/ https://gradientflow.com/the-biggest-ai-risks-sit-outside-the-model/ https://gradientflow.com/this-is-how-self-improving-ai-actually-starts/ https://gradientflow.com/what-comes-after-language-models/ https://gradientflow.com/when-work-history-becomes-an-ai-asset/ https://gradientflow.com/why-data-centers-became-the-face-of-the-ai-backlash/ https://gradientflow.com/your-ai-should-be-better-on-day-500/ https://gradientflow.com/your-model-is-not-your-moat/ https://help.openai.com/en/articles/6825453-chatgpt-release-notes https://huggingface.co/Qwen/Qwen3.8-27B https://huggingface.co/Qwen/Qwen3.8-27B-FP8 https://huggingface.co/TokenRhythm/NeoHorse-1-9B https://huggingface.co/blog https://huggingface.co/blog/Dharma-AI/gpu-management-pt2 https://huggingface.co/blog/Hcompany/neomme https://huggingface.co/blog/MultiverseComputingCAI/efficient-knowledge-distillation https://huggingface.co/blog/MultiverseComputingCAI/safety-for-whom https://huggingface.co/blog/allenai/olmoearth-embeddings https://huggingface.co/blog/amazon/strands-lerbot-streaming-data-loop https://huggingface.co/blog/amazon/strands-lerheng-data-loop https://huggingface.co/blog/amazon/strands-lerobot-streaming-data-loop https://huggingface.co/blog/asyncgrpo-lora-hfjobs https://huggingface.co/blog/gradio-workflow-1111 https://huggingface.co/blog/huggingface/one-year-since-the-deepseek-moment https://huggingface.co/blog/ibm-research/altk-evolve-consistency https://huggingface.co/blog/icml-2026-open-reproductions https://huggingface.co/blog/icml-reproduction https://huggingface.co/blog/lighteval/ICML-2024-papers-reproducibility https://huggingface.co/blog/mishig/local-moores-law https://huggingface.co/blog/muse-glimmer https://huggingface.co/blog/nvidia/magpie-tts-multilingual-voice-agents https://huggingface.co/blog/security-incident-july-2026 https://huggingface.co/blog/state-of-open-models-summer-2026 https://huggingface.co/datasets/microsoft/Orchard https://huggingface.co/datasets/paiteq-ai/rag-bench-2026q2 https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731 https://huggingface.co/moonshotai/Kimi-K3 https://huggingface.co/nvidia/Kimi-K3-NVFP4 https://huggingface.co/papers/2608.09867 https://huggingface.co/papers/2608.19098 https://huggingface.co/papers/2608.19880 https://huggingface.co/papers/2608.25593 https://huggingface.co/papers/2609.03199 https://huggingface.co/papers/2609.08183 https://huggingface.co/papers/2609.20800 https://huyenchip.com/2025/01/07/agents.html https://hyper.ai/en/papers/2608.25593 https://hyper.ai/en/papers/2609.08183 https://importai.substack.com/p/import-ai-472 https://inferenceengineering.tech/learn/vllm-vs-sglang-vs-tensorrt-llm https://interconnects.ai/p/glm-53-how-chinese-labs-keep-stride https://kenhuangus.substack.com https://labs.cloudsecurityalliance.org/research/csa-research-note-huggingface-autonomous-agent-breach-202607 https://labs.cloudsecurityalliance.org/research/csa-research-note-mcp-security-crisis-20260504-csa-styled https://lilianweng.github.io/posts/2026-07-04-harness https://magazine.sebastianraschka.com/p/llm-research-papers-2026-part1 https://magazine.sebastianraschka.com/p/using-local-coding-agents https://magazine.sebastianraschka.md/p/llm-research-papers-2026-part1 https://metafiedlab.com/blog/how-to-build-a-production-ready-rag-pipeline-in-2026 https://micheallanham.substack.com/p/why-success-is-lying-to-you-the-2026 https://nathanbenaich.substack.com/p/state-of-ai-may-2026 https://neuraltrust.ai/blog/context-window-optimization-2026-strategies https://neurohive.io/en/state-of-the-art/agent-lightning-v1-0 https://news.google.com https://news.google.com/rss/articles/CBMiW0FVX3lxTE9QNW9yOVdvcUdTcThnUGpqSzJjbmZoY2lneThEZFNacjBKdzN6N2xFYXQzSjBKSXdsVk9ZRXotZjhjeVVtcXRJbGpNTDhSSUc1ajVKUTZOamVMcVE https://news.google.com/rss/articles/CBMiX0FVX3lxTE1nNk9oS1M5R2tRa1hxZ0I1X0ZvNC15aEhrT2hyMFRpZnRUdDY4QzZSRXdfUmU1YjlYNGctV1dZXzZoakVEY19GNlNBNW1iX1RvWWxWNXhzUmRJcm93YTFr https://news.google.com/rss/articles/CBMiYkFVX2lxTFBoVFQ4NnRrVGF2dXBQMmhjNzhmQUVZQVF6ckVvWlpkMVdHMnpFOVpodWhoUEw4bWdacG1JRE9kUUZpUWN2WnJiZ3NHaFY0LWJHMF9zOGFGR0tlZmhta180N2Nkaw https://news.google.com/rss/articles/CBMicEFVX3lxTFA0cmtPNG1Fa0prVnBpWXZmaDh4VXRGOE5haUQzQ1VjY3BpUnRJNks2UFhNdEhXdjdDaDFGbHdTX3FCTVhmeDBLdWIyZWVBVlY1T2tUMWNZYWNXX3NoaWIxdWxNQ3NLczNiSWRuTjZjWkQ https://news.google.com/rss/articles/CBMickFVX2lxTE1FRlRrdjBDUVZ5bFJMSjJlZTFuUkdLdm5iMTBQMWVTMUxoOW13bnNDUzRPWDZ0TlI4YTYxb3ZrZXQ5Z1kxQ1MwSVdpTWdnb0h6aDdIdTMwSjZGVFNhOGhIMVFRSGpIbFhVWnN1clFWNjJSZw https://news.google.com/rss/articles/CBMidkFVX3lxTFBBeTNVQy1tcnZPVW9SM3drbmxGN1UxRktiY3JfLW9rMDdIMm9SOUd3b3RJNHpSRXVzS1JKckctcFUyZHpNTVE3aEtWVFhka3o3UmlLUzk1ZEs3dUhyVmRJYVctb0IyN1JuNDB6N0hZdTR2TWJTVXc https://news.google.com/rss/articles/CBMilgFBVV95cUxOQTY0c01YMl93aHRWWDNiU0VMYnQwM1VHdDlkZ0VYdlRnOTFoM3RKYTczWVNQZlhtdVkwdkFmbGFWLU9qWmFDOWZfUlJMWl9nUnBoODhCb2ZRWGo0ZnQ5emtNbkstZXpScUlaXzNQd3FKWUcxZXJYV2VmZmIwZDFhcWRfdUpkTGpkNXpfcmhIaFhKaWZldGc https://news.google.com/rss/articles/CBMingFBVV95cUxNTnkxbFNBbk5aakJtVGhqVnZZZWxMcFdvazItYm9tYlBuVU5EMUdFTkRxREZaVnFHSVBRNzF0eWJRQkNFQ2RTQjFvSy03MnM1VkN5ekhjcVZrOGx2R0poMFpFUkptcWMtNWhZOERtWjR6OHlla1lELW1lTXJHbFRPRmdsRE1pb24tSzJDYl9Bd0d0RlRRd1pWMUNxYkIwZw https://open.substack.com/pub/alexewerlof/p/owasp-top-10-ai-llm-agents https://open.substack.com/pub/bytebytego/p/whats-next-in-ai-five-trends-to-watch https://open.substack.com/pub/claudiostamile/p/agent-memory-is-not-rag-a-practical https://openai.com/index/advancing-responsible-ai-across-europe https://openai.com/index/builders-guide-to-gpt-5-6 https://openai.com/index/building-an-ai-native-finance-function https://openai.com/index/circles https://openai.com/index/cognition-devin-testing-with-astra https://openai.com/index/continuous-voice-interaction-with-gpt-live https://openai.com/index/disrupting-malicious-uses-of-ai-criminal-scam-operation https://openai.com/index/expanding-ai-access-us-government https://openai.com/index/expanding-daybreak-as-the-cyber-defense-window-narrows https://openai.com/index/gpt-6-astra/ https://openai.com/index/gpt-daybreak https://openai.com/index/helping-older-adults-use-ai-in-everyday-life https://openai.com/index/how-to-connect-ai-usage-to-business-value https://openai.com/index/hugging-face-model-evaluation-security-incident https://openai.com/index/improving-gpt-5-6-sol-in-chatgpt https://openai.com/index/introducing-chatgpt-financial-services https://openai.com/index/introducing-the-agents-api https://openai.com/index/model-misalignment-reporting-framework https://openai.com/index/model-ml https://openai.com/index/pacing-model-development-cyber-capabilities https://openai.com/index/perplexity-improving-accuracy-with-astra https://openai.com/index/previewing-ultrafast https://openai.com/index/put-data-to-work https://openai.com/index/putting-frontier-cyber-models-in-more-trusted-hands https://openai.com/index/reimagining-advertising-with-ai https://openai.com/index/responsible-ai-infrastructure-texas https://openai.com/index/scaling-storage-one-billion-users-part-one https://openai.com/index/unlocking-new-ways-of-working https://openai.com/index/using-codex-chatgpt-to-search-for-new-antimicrobials https://openai.org/index/putting-frontier-cyber-models-in-more-trusted-hands https://orca.security/latest/sglang-2026-security-analysis https://oreillyradar.substack.com/p/why-multi-agent-systems-need-memory https://papers.cool/arxiv/2608.27260 https://proceedings.iclr.cc/paper_files/paper/2026/file/fd1eff9dd295df50a41f2521942fa31d-Paper-Conference.pdf https://qaskills.sh/blog/terminal-bench-agent-benchmark-guide-2026 https://qwen.ai/blog https://releasebot.io/updates/anthropic https://research.google/blog/chain-of-agents-large-language-models-collaborating-on-long-context-tasks https://research.nvidia.com/labs/gear/aspire https://research.yandex.com/blog/the-kv-cache-as-an-agent-runtime https://shchegrikovich.substack.com/p/end-to-end-optimisation-of-ai-agents https://showlab.github.io/Show-Harness https://simonwillison.net/2026/Aug/15/qwen-3-8-27b-consumer-hardware https://simonwillison.net/2026/Aug/17/qwen-3-8-27b-aa-index-52 https://simonwillison.net/2026/Aug/9/claude-opus-5-system-prompt https://simonwillison.net/2026/Aug/9/github-models-is-now-retired https://simonwillison.net/2026/Sep/16/datasette-2/ https://simonwillison.net/2026/Sep/16/datasette/ https://simonwillison.net/2026/Sep/17/compaction-summaries/ https://simonwillison.net/2026/Sep/17/how-to-write-with-an-llm/ https://simonwillison.net/2026/Sep/17/targeted-attacks-on-rustaceans/ https://simonwillison.net/2026/Sep/18/gemini-hacked-three-companies/ https://simonwillison.net/2026/Sep/18/thariq-shihipar/ https://simonwillison.net/2026/Sep/3/gpt6-astra/ https://simonwillison.net/2026/Sep/4/rogue-agent-wikis/ https://superpowerdaily.com/posts/microsoft-releases-agent-lightning-to-train-ai-agents-in-their-existing-setup https://theaiengineer.substack.com/p/advanced-rag-06 https://theaiengineer.substack.com/p/comparative-analysis https://theaiengineer.substack.com/p/the-ai-agents-stack-2026-edition https://theaiengineer.substack.com/p/what-is-rag https://thelivingedge.substack.com/p/last-week-in-multimodal-ai-40-search https://tldr.tech/ai/2026-08-08 https://tldr.tech/ai/2026-08-10 https://tldr.tech/ai/2026-08-14 https://tldr.tech/ai/2026-08-17 https://tldr.tech/ai/2026-08-19 https://tldr.tech/ai/2026-09-16 https://trilogyai.substack.com/p/mcp-grows-up-what-the-july-28-spec https://usewire.io/blog/long-context-vs-rag-what-the-data-shows https://uvik.net/blog/agentic-ai-frameworks https://velesdb.com https://wavect.io/blog/rag-vs-finetune-vs-longcontext-2026 https://www-cdn.anthropic.com https://www.404media.co https://www.aisi.gov.uk https://www.alphaxiv.org/abs/2608.25593 https://www.alphaxiv.org/abs/2609.08183 https://www.anthropic.com/news https://www.augmentcode.com/guides/ai-model-routing-guide https://www.cloudflare.com/agents-week/updates https://www.deepmind.com/blog https://www.developersdigest.tech/blog/cloudflare-agent-access-model-2026 https://www.developersdigest.tech/blog/mcp-2026-07-28-breaking-changes https://www.digitalapplied.com/blog/kv-cache-optimization-techniques-2026-engineering-guide https://www.firecrawl.dev/blog/best-ai-coding-agents https://www.langchain.com/blog/august-2026-langchain-newsletter https://www.langchain.com/state-of-agent-engineering https://www.mdpi.com/2076-3417/16/16/7974 https://www.microsoft.com/en-us/research/blog/agent-lightning-v1-0-a-3500-line-lightweight-agentic-rl-framework-for-training-agents-with-real-harnesses https://www.microsoft.com/en-us/research/blog/echoverse-deep-evolving-environments-for-computer-use-agents/ https://www.microsoft.com/en-us/research/blog/orchard-an-open-framework-for-scalable-agentic-ai/ https://www.mindstudio.ai/blog/agent-harnesses-beat-model-upgrades-5-benchmarks https://www.morphllm.com/best-ai-coding-agents-2026 https://www.oneusefulthing.org/p/choosing-to-stay-human https://www.oneusefulthing.org/p/on-working-with-wizards https://www.opentrain.ai/papers/swe-refactor-bench-can-coding-agents-complete-a-long-horizon-whole-repository-st--arxiv-2608.23564 https://www.opentrain.ai/papers/what-makes-good-agentic-data-an-ace-lens-on-data-generation-for-llm-agents--arxiv-2608.27260 https://www.primeintellect.ai/blog/prime-agent https://www.researchgate.net/publication/401720909 https://www.semanticscholar.org/paper/SoK https://www.semanticscholar.org/paper/Towards-Agentic-RAG-with-Deep-Reasoning https://www.techcrunch.com https://www.themoonlight.io/tw/review/jit-agent-scaling-harness-intelligence-via-just-in-time-harness-evolution https://www.theregister.com/devops/2026/07/23/model-context-protocol-prepares-to-break-with-its-stateful-past/5276722 https://www.vectara.com/blog/context-engineering-can-you-trust-long-context https://www.winzheng.com/en/article/swe-refactor-bench-ai-coding-agents-whole-repo-migration https://www.wsj.com https://www.youtube.com/watch https://x.com/AndrewYNg/article/2088302050706686198 https://x.com/AndrewYNg/article/2090840747738374568 https://x.com/AndrewYNg/status/2095890279865721217 https://x.com/AnthropicAI/status/2070665903440871779 https://x.com/GoogleDeepMind/status/2087948366294515971 https://x.com/GoogleDeepMind/status/2087948366294515977 https://x.com/OpenAI/status/208977077845187031262 https://x.com/OpenAI/status/2089777845187031262 https://x.com/OpenAI/status/2090165328290701800 https://x.com/emollick https://x.com/fly51fly/status/2092358103379611747 https://x.com/huggingface/status/2088301795890044975 https://x.com/jerryjliu0/status/2089710424388202565 https://x.com/karpathy/status/2083749667410727319 https://x.com/karpathy/status/2084080419029549276 https://x.com/omarsar0/status/2086509069762981896 https://x.com/omarsar0/status/2090078336697733531 https://x.com/omarsar0/status/2102762406204076532 https://x.com/simonw/status/2085877951925801274 https://x.com/ylecun/status/2084363123339792657 https://youtube.com/watch https://zli12321.github.io/LHTB https://zuplo.com/blog/mcp-auth-multi-agent-systems