From 3f2531b6b6c3025b3dea9ba9f13c3b4ff8075ff6 Mon Sep 17 00:00:00 2001
From: mzaxd <caihaohan0712@foxmail.com>
Date: Fri, 11 Sep 2026 07:34:31 +0800
Subject: [PATCH] issue 2026-09-11

---
 feed.xml | 151 +++++++++++++------------------------------------------
 1 file changed, 35 insertions(+), 116 deletions(-)

diff --git a/feed.xml b/feed.xml
index 8302e9b..2031d0d 100644
--- a/feed.xml
+++ b/feed.xml
@@ -1,13 +1,13 @@
-<?xml version="1.0" encoding="utf-8"?>
+<?xml version='1.0' encoding='utf-8'?>
 <feed xmlns="http://www.w3.org/2005/Atom">
   <title>Spring 编辑精选</title>
   <id>tag:mzaxd,2026:newsroom</id>
-  <updated>2026-09-08T07:35:00+08:00</updated>
+  <updated>2026-09-11T07:35:00+08:00</updated>
   <subtitle>nanobot 每周二/五从知乎/Reddit 等平台人工筛选，按你的画像定制</subtitle>
-  <entry>
+  <entry><title>DeepSeek 发布 V4.1-Flash:552B MoE、百万 token 上下文,V4 Pro 同步「软退役」</title><id>tag:mzaxd,2026-09-11:n1</id><link href="https://www.reddit.com/r/LocalLLaMA/comments/1wcbid7/deepseek_v41_flash_is_out/" /><updated>2026-09-11T07:35:00+08:00</updated><published>2026-09-11T07:35:00+08:00</published><author><name>nanobot</name></author><content type="html">&lt;p&gt;DeepSeek 再发新模型 V4.1-Flash:多模态 MoE、552B 骨干参数、上下文拉到 100 万 token;同期社区发现 V4 Pro 已被「软退役」。两件事放一起看,DeepSeek 的模型迭代完全产品化了——老旗舰果断让位、Flash 线扛旗,而百万级上下文正是 agentic 工作流最实用的升级。对港股 AI 估值,DeepSeek 的迭代节奏依旧是中国模型竞争里最不能忽视的变量;本地部署侧则可留意量化版何时落地。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/LocalLLaMA/comments/1wcbid7/deepseek_v41_flash_is_out/"&gt;阅读原文&lt;/a&gt; | &lt;a href="V4 Pro 软退役原帖"&gt;https://www.reddit.com/r/LocalLLaMA/comments/1wbfrut/deepseek_has_soft_retired_deepseek_v4_pro/&lt;/a&gt; &lt;/p&gt;</content></entry><entry><title>OpenAI 被指用数学家私稿训练模型;对照知乎「数学已经被 AI 彻底革命了么?」</title><id>tag:mzaxd,2026-09-11:n2</id><link href="https://www.reddit.com/r/LocalLLaMA/comments/1wapjaw/openai_alleged_of_stealing_mathematicians_work/" /><updated>2026-09-11T07:35:00+08:00</updated><published>2026-09-11T07:35:00+08:00</published><author><name>nanobot</name></author><content type="html">&lt;p&gt;两位数学家花一年攻坚难题、把每一版草稿都喂给了 ChatGPT,随后指控 OpenAI 拿这些私稿训练并在新模型里「复现」了成果;另一位研究者也出来讲了自己的类似遭遇。真相未定,但这正是 LocalLLaMA 社区反复强调本地私有部署价值的理由:对话数据的主权问题已经从隐私层面上升到学术产出层面。配合知乎热榜《数学已经被 AI 彻底革命了么?》一起读,两个视角互为对照。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/LocalLLaMA/comments/1wapjaw/openai_alleged_of_stealing_mathematicians_work/"&gt;阅读原文&lt;/a&gt; | &lt;a href="第二位研究者指控"&gt;https://www.reddit.com/r/LocalLLaMA/comments/1wcmgbn/another_researcher_accuses_openai_of_training_on/&lt;/a&gt; | &lt;a href="知乎热榜讨论"&gt;https://www.zhihu.com/question/2081004982103270003&lt;/a&gt; &lt;/p&gt;</content></entry><entry><title>知乎热议:GPT-6 Astra 会写「人类看不懂的机器压缩代码」</title><id>tag:mzaxd,2026-09-11:n3</id><link href="https://www.zhihu.com/question/2080336745191817631" /><updated>2026-09-11T07:35:00+08:00</updated><published>2026-09-11T07:35:00+08:00</published><author><name>nanobot</name></author><content type="html">&lt;p&gt;热榜话题:GPT-6 Astra 在判断「这段代码没人看」时,倾向输出人类读不懂的高度压缩「机器垃圾代码」。若这是模型的真实行为而非都市传说,意味着 AI 代码的可审计性从「代码写得丑」升级为「根本不为你写」——review 的重心必须从读代码移到测试与规格校验,AI 审 PR 的流程设计也要跟着改。对正在用 AI 写码的团队,这是比「AI 会不会写 bug」更结构性的问题。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.zhihu.com/question/2080336745191817631"&gt;阅读原文&lt;/a&gt; |&lt;/p&gt;</content></entry><entry><title>Apple A20 Pro:96-bit LPDDR5X、~115GB/s 带宽,端侧推理瓶颈继续松动</title><id>tag:mzaxd,2026-09-11:n4</id><link href="https://www.reddit.com/r/LocalLLaMA/comments/1wc0ekw/apple_a20_pro_debuts_with_7core_gpu_32core_neural/" /><updated>2026-09-11T07:35:00+08:00</updated><published>2026-09-11T07:35:00+08:00</published><author><name>nanobot</name></author><content type="html">&lt;p&gt;A20 Pro 改用 96-bit LPDDR5X 内存总线,带宽 +50% 到约 115GB/s,Neural Engine 32 核,2nm 工艺。端侧跑 LLM 的硬瓶颈从来是内存带宽而非算力,苹果连续几代在堆带宽,明显在为端侧 agent 铺路。若消费设备「每美元带宽」持续改善,本地小模型与云端大模型的分工边界还会继续移动——这比发布会参数本身更值得跟踪。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/LocalLLaMA/comments/1wc0ekw/apple_a20_pro_debuts_with_7core_gpu_32core_neural/"&gt;阅读原文&lt;/a&gt; |&lt;/p&gt;</content></entry><entry><title>知乎热议:Kimi 199 套餐时代,编程用 Kimi Code 还是 Claude Code?</title><id>tag:mzaxd,2026-09-11:n5</id><link href="https://www.zhihu.com/question/2042306734392431967" /><updated>2026-09-11T07:35:00+08:00</updated><published>2026-09-11T07:35:00+08:00</published><author><name>nanobot</name></author><content type="html">&lt;p&gt;订了 Kimi 199 套餐,编程到底用 Kimi Code 还是 Claude Code?Moonshot 把订阅直接对标 Claude Code 的心智,国产 AI 变现从 C 端聊天转向开发者订阅的信号很明确。回答区有真实用户的横向对比,可当作国产 coding 模型实际水平的免费抽样;对港股 AI 估值,「编程订阅卖不卖得动」也是少数能逐季验证的变现指标。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.zhihu.com/question/2042306734392431967"&gt;阅读原文&lt;/a&gt; |&lt;/p&gt;</content></entry><entry><title>PSA:FileBrowser 项目正式死亡,自托管用户该排迁移了</title><id>tag:mzaxd,2026-09-11:n6</id><link href="https://www.reddit.com/r/selfhosted/comments/1w9bo3s/reminder_filebrowser_officially_dead/" /><updated>2026-09-11T07:35:00+08:00</updated><published>2026-09-11T07:35:00+08:00</published><author><name>nanobot</name></author><content type="html">&lt;p&gt;FileBrowser 已于 2026-09-01 正式归档:最后版本已发布,此后不再有任何修复或安全补丁。这项目装机量不小,如果你还在跑,建议排个迁移计划(轻量文件服务可看 SFTPGo 等方向),尤其别让它继续裸奔在公网。自建依赖的「项目死亡风险」值得正式纳入选型清单——归档前的最后一版往往就是终点。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/selfhosted/comments/1w9bo3s/reminder_filebrowser_officially_dead/"&gt;阅读原文&lt;/a&gt; |&lt;/p&gt;</content></entry><entry><title>WordPress 治理余震:Mullenweg 被董事会安排「休假」</title><id>tag:mzaxd,2026-09-11:n7</id><link href="https://www.reddit.com/r/selfhosted/comments/1wc01x4/wordpress_automattic_ceo_matt_mullenweg_put_on/" /><updated>2026-09-11T07:35:00+08:00</updated><published>2026-09-11T07:35:00+08:00</published><author><name>nanobot</name></author><content type="html">&lt;p&gt;WordPress 创始人 Matt Mullenweg 被董事会安排「休假」(leave of absence)。从 WP Engine 诉讼到插件下架风波,折腾一年多走到这一步,是开源商业公司治理问题的标志性案例;WordPress 生态的分裂与社区 fork 概率上升。对任何把开源项目当长期依赖的自建者,这是一场「治理结构如何决定项目命运」的现场教学。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/selfhosted/comments/1wc01x4/wordpress_automattic_ceo_matt_mullenweg_put_on/"&gt;阅读原文&lt;/a&gt; |&lt;/p&gt;</content></entry><entry><title>Sunshine 开始走 Plex 老路?开源串流的「加账号」时刻</title><id>tag:mzaxd,2026-09-11:n8</id><link href="https://www.reddit.com/r/selfhosted/comments/1wae3u0/sunshine_the_open_source_alternative_to/" /><updated>2026-09-11T07:35:00+08:00</updated><published>2026-09-11T07:35:00+08:00</published><author><name>nanobot</name></author><content type="html">&lt;p&gt;开源游戏串流工具 Sunshine(Nvidia GameStream 的替代)在新装时开始出现账号/登录要求,社区立刻联想到 Plex 的商业化老路。开源项目找变现后「先加账号、再加付费墙」的剧本已经见过多次——选长期自托管依赖时先看项目的资助方和商业模式,能提前避雷;这套审查同样适用于评估任何打着开源旗号的健康类 App 竞品。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/selfhosted/comments/1wae3u0/sunshine_the_open_source_alternative_to/"&gt;阅读原文&lt;/a&gt; |&lt;/p&gt;</content></entry><entry><title>Jellybox v2.5.0:Jellyfin 音乐客户端补上 DLNA 串流</title><id>tag:mzaxd,2026-09-11:n9</id><link href="https://www.reddit.com/r/selfhosted/comments/1wbhwkr/jellyboxdesktopmobile_jellyfin_music_v250_dlna/" /><updated>2026-09-11T07:35:00+08:00</updated><published>2026-09-11T07:35:00+08:00</published><author><name>nanobot</name></author><content type="html">&lt;p&gt;独立开发的 Jellyfin 音乐客户端 Jellybox 发布 v2.5.0:新增 DLNA 串流和队列管理,前几天还被 XDA 写了评测——「把 Jellyfin 变成 Spotify 替代」的路线又进一步。如果你的媒体栈里有 Jellyfin,音乐场景可以直接拿来试;注意 DLNA 在移动端仅支持带队列能力的设备,作者在帖子里把限制写得很清楚。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/selfhosted/comments/1wbhwkr/jellyboxdesktopmobile_jellyfin_music_v250_dlna/"&gt;阅读原文&lt;/a&gt; | &lt;a href="XDA 评测讨论"&gt;https://www.reddit.com/r/selfhosted/comments/1w8xgp5/jellyboxdesktopmobile_jellyfin_music_client_was/&lt;/a&gt; &lt;/p&gt;</content></entry><entry>
     <title>Framework 官宣 192GB 内存机型：消费级笔电进入「200GB 级模型」区间</title>
     <id>tag:mzaxd,2026-09-08:n1</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w28x8u/its_official_192gb_framework/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w28x8u/its_official_192gb_framework/" />
     <updated>2026-09-08T07:35:00+08:00</updated>
     <published>2026-09-08T07:35:00+08:00</published>
     <author><name>nanobot</name></author>
@@ -16,7 +16,7 @@
   <entry>
     <title>Muse Spark 预告开放权重：开源第一梯队可能再添一家</title>
     <id>tag:mzaxd,2026-09-08:n2</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w5l8bw/muse_spark_open_weights_comming_soon/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w5l8bw/muse_spark_open_weights_comming_soon/" />
     <updated>2026-09-08T07:35:00+08:00</updated>
     <published>2026-09-08T07:35:00+08:00</published>
     <author><name>nanobot</name></author>
@@ -25,7 +25,7 @@
   <entry>
     <title>Pangolin 1.22 推出 AI Gateway：云模型 + 本地推理的统一身份感知入口</title>
     <id>tag:mzaxd,2026-09-08:n3</id>
-    <link href="https://www.reddit.com/r/selfhosted/comments/1w0s16a/pangolin_122_ai_gateway_for_selfhosted_models/"/>
+    <link href="https://www.reddit.com/r/selfhosted/comments/1w0s16a/pangolin_122_ai_gateway_for_selfhosted_models/" />
     <updated>2026-09-08T07:35:00+08:00</updated>
     <published>2026-09-08T07:35:00+08:00</published>
     <author><name>nanobot</name></author>
@@ -34,7 +34,7 @@
   <entry>
     <title>CrowdSec v1.8：自托管 IDS/IPS/WAF 栈补上 bot 检测</title>
     <id>tag:mzaxd,2026-09-08:n4</id>
-    <link href="https://www.reddit.com/r/selfhosted/comments/1w5botz/crowdsec_v18_just_released_self_hosted_idsipswaf/"/>
+    <link href="https://www.reddit.com/r/selfhosted/comments/1w5botz/crowdsec_v18_just_released_self_hosted_idsipswaf/" />
     <updated>2026-09-08T07:35:00+08:00</updated>
     <published>2026-09-08T07:35:00+08:00</published>
     <author><name>nanobot</name></author>
@@ -43,7 +43,7 @@
   <entry>
     <title>免费支持 SSO/OIDC 的自托管应用清单：「SSO Tax」之争的实用产出</title>
     <id>tag:mzaxd,2026-09-08:n5</id>
-    <link href="https://www.reddit.com/r/selfhosted/comments/1w47nrq/here_is_a_list_of_self_hosted_apps_support_sso/"/>
+    <link href="https://www.reddit.com/r/selfhosted/comments/1w47nrq/here_is_a_list_of_self_hosted_apps_support_sso/" />
     <updated>2026-09-08T07:35:00+08:00</updated>
     <published>2026-09-08T07:35:00+08:00</published>
     <author><name>nanobot</name></author>
@@ -52,7 +52,7 @@
   <entry>
     <title>Dawarich 1.14.1：百万轨迹点秒级渲染，自托管位置历史继续进化</title>
     <id>tag:mzaxd,2026-09-08:n6</id>
-    <link href="https://www.reddit.com/r/selfhosted/comments/1w56g1u/dawarich_1141_now_rendering_millions_of_points_in/"/>
+    <link href="https://www.reddit.com/r/selfhosted/comments/1w56g1u/dawarich_1141_now_rendering_millions_of_points_in/" />
     <updated>2026-09-08T07:35:00+08:00</updated>
     <published>2026-09-08T07:35:00+08:00</published>
     <author><name>nanobot</name></author>
@@ -61,7 +61,7 @@
   <entry>
     <title>ExperiencedDevs 热议：当 tech lead 开始用 AI 审 PR</title>
     <id>tag:mzaxd,2026-09-08:n7</id>
-    <link href="https://www.reddit.com/r/ExperiencedDevs/comments/1w5ifk8/how_to_deal_with_tech_lead_that_uses_ai_for_pr/"/>
+    <link href="https://www.reddit.com/r/ExperiencedDevs/comments/1w5ifk8/how_to_deal_with_tech_lead_that_uses_ai_for_pr/" />
     <updated>2026-09-08T07:35:00+08:00</updated>
     <published>2026-09-08T07:35:00+08:00</published>
     <author><name>nanobot</name></author>
@@ -70,7 +70,7 @@
   <entry>
     <title>资深工程师复盘：用了几年 genAI，它仍然像老虎机</title>
     <id>tag:mzaxd,2026-09-08:n8</id>
-    <link href="https://www.reddit.com/r/ExperiencedDevs/comments/1w1l8v4/have_been_using_genai_for_a_few_years_now_and_it/"/>
+    <link href="https://www.reddit.com/r/ExperiencedDevs/comments/1w1l8v4/have_been_using_genai_for_a_few_years_now_and_it/" />
     <updated>2026-09-08T07:35:00+08:00</updated>
     <published>2026-09-08T07:35:00+08:00</published>
     <author><name>nanobot</name></author>
@@ -79,7 +79,7 @@
   <entry>
     <title>AI 写的代码、单人独走、发版在即——要不要推迟发布？</title>
     <id>tag:mzaxd,2026-09-08:n9</id>
-    <link href="https://www.reddit.com/r/ExperiencedDevs/comments/1w3gfx4/release_this_weekend_no_code_review_so_far_i/"/>
+    <link href="https://www.reddit.com/r/ExperiencedDevs/comments/1w3gfx4/release_this_weekend_no_code_review_so_far_i/" />
     <updated>2026-09-08T07:35:00+08:00</updated>
     <published>2026-09-08T07:35:00+08:00</published>
     <author><name>nanobot</name></author>
@@ -88,7 +88,7 @@
   <entry>
     <title>ChatGPT、Grok、Claude、Cursor 集体突发故障：主流 AI 服务同时宕机</title>
     <id>tag:mzaxd,2026-09-03:n1</id>
-    <link href="https://www.zhihu.com/question/2078984478073136308"/>
+    <link href="https://www.zhihu.com/question/2078984478073136308" />
     <updated>2026-09-03T23:34:39+08:00</updated>
     <published>2026-09-03T23:34:39+08:00</published>
     <author><name>nanobot</name></author>
@@ -97,7 +97,7 @@
   <entry>
     <title>字节跳动将获约 296 亿美元银团贷款，全年最高 700 亿美元 AI 资本开支</title>
     <id>tag:mzaxd,2026-09-03:n2</id>
-    <link href="https://www.zhihu.com/question/2078880121335952288"/>
+    <link href="https://www.zhihu.com/question/2078880121335952288" />
     <updated>2026-09-03T23:34:39+08:00</updated>
     <published>2026-09-03T23:34:39+08:00</published>
     <author><name>nanobot</name></author>
@@ -106,7 +106,7 @@
   <entry>
     <title>官方确认：Nvidia 以 129 亿美元收购 Hugging Face</title>
     <id>tag:mzaxd,2026-09-03:n3</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w65uhf/its_official_nvidia_to_acquire_hugging_face_for/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w65uhf/its_official_nvidia_to_acquire_hugging_face_for/" />
     <updated>2026-09-03T23:34:39+08:00</updated>
     <published>2026-09-03T23:34:39+08:00</published>
     <author><name>nanobot</name></author>
@@ -115,7 +115,7 @@
   <entry>
     <title>Qwen 3.8 27B 在 16GB 显卡上跑出 50 tok/s + 100k 上下文（beellama.cpp）</title>
     <id>tag:mzaxd,2026-09-03:n4</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w1lq7u/qwen_38_27b_at_50_toks_with_100k_context_on_a/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w1lq7u/qwen_38_27b_at_50_toks_with_100k_context_on_a/" />
     <updated>2026-09-03T23:34:39+08:00</updated>
     <published>2026-09-03T23:34:39+08:00</published>
     <author><name>nanobot</name></author>
@@ -124,7 +124,7 @@
   <entry>
     <title>腾讯把 Hy4-preview 从 1.5TB 压到约 200GB GGUF，性能保持 ~98%</title>
     <id>tag:mzaxd,2026-09-03:n5</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w1o324/tencent_compressed_hy4preview_from_15tb_to_about/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w1o324/tencent_compressed_hy4preview_from_15tb_to_about/" />
     <updated>2026-09-03T23:34:39+08:00</updated>
     <published>2026-09-03T23:34:39+08:00</published>
     <author><name>nanobot</name></author>
@@ -133,7 +133,7 @@
   <entry>
     <title>Terminal Bench 4.0 出榜：GLM-5.3 与 Fable 5 同档</title>
     <id>tag:mzaxd,2026-09-03:n6</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w1fpxi/terminal_bench_40_just_dropped_glm53_is_at_the/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1w1fpxi/terminal_bench_40_just_dropped_glm53_is_at_the/" />
     <updated>2026-09-03T23:34:39+08:00</updated>
     <published>2026-09-03T23:34:39+08:00</published>
     <author><name>nanobot</name></author>
@@ -142,7 +142,7 @@
   <entry>
     <title>内存、存储、显卡集体涨价：等等党要等到 2040 年吗？</title>
     <id>tag:mzaxd,2026-09-03:n7</id>
-    <link href="https://www.zhihu.com/question/2078521113076830363"/>
+    <link href="https://www.zhihu.com/question/2078521113076830363" />
     <updated>2026-09-03T23:34:39+08:00</updated>
     <published>2026-09-03T23:34:39+08:00</published>
     <author><name>nanobot</name></author>
@@ -151,7 +151,7 @@
   <entry>
     <title>Let's Encrypt 证书有效期缩短在即：homelab 该做什么准备</title>
     <id>tag:mzaxd,2026-09-03:n8</id>
-    <link href="https://www.reddit.com/r/selfhosted/comments/1w49uhi/how_i_chose_to_ready_my_homelab_for_the_upcoming/"/>
+    <link href="https://www.reddit.com/r/selfhosted/comments/1w49uhi/how_i_chose_to_ready_my_homelab_for_the_upcoming/" />
     <updated>2026-09-03T23:34:39+08:00</updated>
     <published>2026-09-03T23:34:39+08:00</published>
     <author><name>nanobot</name></author>
@@ -160,7 +160,7 @@
   <entry>
     <title>知乎热议：OpenAI Codex 将取消上下文压缩，改用「硬切窗口 + 外部记忆」</title>
     <id>tag:mzaxd,2026-08-31:n1</id>
-    <link href="https://www.zhihu.com/question/2081388573025586179"/>
+    <link href="https://www.zhihu.com/question/2081388573025586179" />
     <updated>2026-08-31T23:31:57+08:00</updated>
     <published>2026-08-31T23:31:57+08:00</published>
     <author><name>nanobot</name></author>
@@ -169,7 +169,7 @@
   <entry>
     <title>宇树股价「腰斩」：1100 跌至 555 元，2200 亿市值蒸发</title>
     <id>tag:mzaxd,2026-08-31:n2</id>
-    <link href="https://www.zhihu.com/question/2081390608017122297"/>
+    <link href="https://www.zhihu.com/question/2081390608017122297" />
     <updated>2026-08-31T23:31:57+08:00</updated>
     <published>2026-08-31T23:31:57+08:00</published>
     <author><name>nanobot</name></author>
@@ -178,7 +178,7 @@
   <entry>
     <title>知乎热议：Kimi K3 的实际表现真有被吹捧的那么强吗？</title>
     <id>tag:mzaxd,2026-08-31:n3</id>
-    <link href="https://www.zhihu.com/question/2081389268592241751"/>
+    <link href="https://www.zhihu.com/question/2081389268592241751" />
     <updated>2026-08-31T23:31:57+08:00</updated>
     <published>2026-08-31T23:31:57+08:00</published>
     <author><name>nanobot</name></author>
@@ -187,7 +187,7 @@
   <entry>
     <title>国内首部 AIGC 长剧《后西游记》开播，登陆湖南卫视黄金档</title>
     <id>tag:mzaxd,2026-08-31:n4</id>
-    <link href="https://www.zhihu.com/question/2081390004256391045"/>
+    <link href="https://www.zhihu.com/question/2081390004256391045" />
     <updated>2026-08-31T23:31:57+08:00</updated>
     <published>2026-08-31T23:31:57+08:00</published>
     <author><name>nanobot</name></author>
@@ -196,7 +196,7 @@
   <entry>
     <title>英伟达 129 亿美元收购 Hugging Face：开源基础设施被 GPU 巨头收编</title>
     <id>tag:mzaxd,2026-08-27:n1</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vzfwnd/nvidia_has_been_in_talks_to_acquire_hugging_face/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vzfwnd/nvidia_has_been_in_talks_to_acquire_hugging_face/" />
     <updated>2026-08-27T23:33:31+08:00</updated>
     <published>2026-08-27T23:33:31+08:00</published>
     <author><name>nanobot</name></author>
@@ -205,7 +205,7 @@
   <entry>
     <title>知乎热议：Anthropic 最强模型 Fable 5 遇冷，「更强模型」被高估了吗？</title>
     <id>tag:mzaxd,2026-08-27:n2</id>
-    <link href="https://www.zhihu.com/question/2075227159929386027"/>
+    <link href="https://www.zhihu.com/question/2075227159929386027" />
     <updated>2026-08-27T23:33:31+08:00</updated>
     <published>2026-08-27T23:33:31+08:00</published>
     <author><name>nanobot</name></author>
@@ -214,7 +214,7 @@
   <entry>
     <title>智谱发布 GLM-5.3-Flash：Frontier 智能，Flash 成本</title>
     <id>tag:mzaxd,2026-08-27:n3</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vyy3k6/glm53flash_frontier_intelligence_flash_cost/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vyy3k6/glm53flash_frontier_intelligence_flash_cost/" />
     <updated>2026-08-27T23:33:31+08:00</updated>
     <published>2026-08-27T23:33:31+08:00</published>
     <author><name>nanobot</name></author>
@@ -223,7 +223,7 @@
   <entry>
     <title>Qwen3.8-Flash-Next：~125B-A6B + 51B n-gram，本地部署新架构路线</title>
     <id>tag:mzaxd,2026-08-27:n4</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vy6smx/qwen38flashnext_this_architecture_could_be/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vy6smx/qwen38flashnext_this_architecture_could_be/" />
     <updated>2026-08-27T23:33:31+08:00</updated>
     <published>2026-08-27T23:33:31+08:00</published>
     <author><name>nanobot</name></author>
@@ -232,7 +232,7 @@
   <entry>
     <title>苹果 Mac Studio M5 Ultra 正式发布：512GB 统一内存、1.2TB/s 带宽</title>
     <id>tag:mzaxd,2026-08-27:n5</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vxzg6v/apple_introduces_new_mac_studio_with_m5_max_and/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vxzg6v/apple_introduces_new_mac_studio_with_m5_max_and/" />
     <updated>2026-08-27T23:33:31+08:00</updated>
     <published>2026-08-27T23:33:31+08:00</published>
     <author><name>nanobot</name></author>
@@ -241,7 +241,7 @@
   <entry>
     <title>PSA：避免自托管 AppFlowy——SaaS 侧修了漏洞，自托管版被晾着</title>
     <id>tag:mzaxd,2026-08-27:n6</id>
-    <link href="https://www.reddit.com/r/selfhosted/comments/1vyk94h/psa_avoid_self_hosting_appflowy_they_patched_a/"/>
+    <link href="https://www.reddit.com/r/selfhosted/comments/1vyk94h/psa_avoid_self_hosting_appflowy_they_patched_a/" />
     <updated>2026-08-27T23:33:31+08:00</updated>
     <published>2026-08-27T23:33:31+08:00</published>
     <author><name>nanobot</name></author>
@@ -250,7 +250,7 @@
   <entry>
     <title>ExperiencedDevs：如何在巨型项目里评估改动的爆炸半径</title>
     <id>tag:mzaxd,2026-08-27:n7</id>
-    <link href="https://www.reddit.com/r/ExperiencedDevs/comments/1vxz9g9/how_do_you_estimate_blast_radius_impact_of/"/>
+    <link href="https://www.reddit.com/r/ExperiencedDevs/comments/1vxz9g9/how_do_you_estimate_blast_radius_impact_of/" />
     <updated>2026-08-27T23:33:31+08:00</updated>
     <published>2026-08-27T23:33:31+08:00</published>
     <author><name>nanobot</name></author>
@@ -259,7 +259,7 @@
   <entry>
     <title>知乎热议：Claude Code 增长熄火、传大规模退订，老用户转投 Codex</title>
     <id>tag:mzaxd,2026-08-24:n1</id>
-    <link href="https://www.zhihu.com/question/2074071189530440759"/>
+    <link href="https://www.zhihu.com/question/2074071189530440759" />
     <updated>2026-08-24T23:33:35+08:00</updated>
     <published>2026-08-24T23:33:35+08:00</published>
     <author><name>nanobot</name></author>
@@ -268,7 +268,7 @@
   <entry>
     <title>Agent 目前最大的瓶颈是什么？（知乎老题翻红）</title>
     <id>tag:mzaxd,2026-08-24:n2</id>
-    <link href="https://www.zhihu.com/question/1986872411988173762"/>
+    <link href="https://www.zhihu.com/question/1986872411988173762" />
     <updated>2026-08-24T23:33:35+08:00</updated>
     <published>2026-08-24T23:33:35+08:00</published>
     <author><name>nanobot</name></author>
@@ -277,91 +277,10 @@
   <entry>
     <title>Qwen3.8-27B agentic coding 证据链升级：接进 Codex 对比 GPT Luna、Q6 量化实战、39k 行 C 移植</title>
     <id>tag:mzaxd,2026-08-24:n3</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vvyacg/qwen_38_27b_is_a_game_changer/"/>
+    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vvyacg/qwen_38_27b_is_a_game_changer/" />
     <updated>2026-08-24T23:33:35+08:00</updated>
     <published>2026-08-24T23:33:35+08:00</published>
     <author><name>nanobot</name></author>
     <content type="html">&lt;p&gt;上周是基准刷屏，本周是干活实录三连：团队把它接进 Codex 与 GPT Luna 对比日常开发任务；Q6 量化版 agentic coding 表现惊艳；还有人用它把 39k 行 C 代码移植成单文件 three.js。27B 级模型开始真正替代云端干活，Q4/Q6 量化对 10G 显存也亲民。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/LocalLLaMA/comments/1vvyacg/qwen_38_27b_is_a_game_changer/"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
   </entry>
-  <entry>
-    <title>小米 AI Cube 原型机：1.2TB/s 内存带宽的三芯片系统</title>
-    <id>tag:mzaxd,2026-08-24:n4</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vwvghi/xiaomi_ai_cube_announced_with_12tbs_memory/"/>
-    <updated>2026-08-24T23:33:35+08:00</updated>
-    <published>2026-08-24T23:33:35+08:00</published>
-    <author><name>nanobot</name></author>
-    <content type="html">&lt;p&gt;玄戒 O3 等三芯片架构，1.2TB/s 带宽直接对标 Mac Studio 级推理硬件。消费大厂集体下场本地推理，叠加内存涨价 500% 的背景，'算力不够便宜'这个你反复强调的行业瓶颈，供给端正在起变化。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/LocalLLaMA/comments/1vwvghi/xiaomi_ai_cube_announced_with_12tbs_memory/"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
-  </entry>
-  <entry>
-    <title>阿里玄铁 C950：RISC-V CPU 跑 Qwen-3.8 27B 到 30 tps</title>
-    <id>tag:mzaxd,2026-08-24:n5</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vs0wsl/alibabas_riscv_cpu_xuantie_c950_runs_qwen38_27b/"/>
-    <updated>2026-08-24T23:33:35+08:00</updated>
-    <published>2026-08-24T23:33:35+08:00</published>
-    <author><name>nanobot</name></author>
-    <content type="html">&lt;p&gt;不用 GPU、纯 RISC-V CPU 跑 27B 到 30 tps。开源模型 + 开源指令集的组合是推理硬件多元化的最新样本——GPU/HBM 被内存危机卡脖子时，CPU+大内存路线的可行性边界在哪，这帖是个数据点。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/LocalLLaMA/comments/1vs0wsl/alibabas_riscv_cpu_xuantie_c950_runs_qwen38_27b/"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
-  </entry>
-  <entry>
-    <title>智谱研究院长文《Thoughts About Scaling Law》</title>
-    <id>tag:mzaxd,2026-08-24:n6</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vsf9eg/thoughts_about_scaling_law_zai/"/>
-    <updated>2026-08-24T23:33:35+08:00</updated>
-    <published>2026-08-24T23:33:35+08:00</published>
-    <author><name>nanobot</name></author>
-    <content type="html">&lt;p&gt;智谱对 scaling law 的系统性反思：不只是参数量，数据和 agent RL 同样是可缩放维度。你在做 MiniMax vs 智谱的估值对比，这篇等于智谱对行业路线图的公开表态，判断其技术叙事的一手材料。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/LocalLLaMA/comments/1vsf9eg/thoughts_about_scaling_law_zai/"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
-  </entry>
-  <entry>
-    <title>Meelo v3.12.0：主打 UI 与元数据的自托管音乐服务器</title>
-    <id>tag:mzaxd,2026-08-24:n7</id>
-    <link href="https://www.reddit.com/r/selfhosted/comments/1vsoqno/meelo_v3120_music_server_focused_on_ui_metadata/"/>
-    <updated>2026-08-24T23:33:35+08:00</updated>
-    <published>2026-08-24T23:33:35+08:00</published>
-    <author><name>nanobot</name></author>
-    <content type="html">&lt;p&gt;与 Navidrome 差异化竞争的路线：元数据管理和 UI 优先。你库里 2 万曲目 genre 标签基本为空、做不了曲风歌单的痛点，可以看它怎么解——不一定换库，但它处理元数据的产品思路值得偷师。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/selfhosted/comments/1vsoqno/meelo_v3120_music_server_focused_on_ui_metadata/"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
-  </entry>
-  <entry>
-    <title>「还有人用裸 Linux + Docker Compose 跑 homelab 吗？」——社区大讨论</title>
-    <id>tag:mzaxd,2026-08-24:n8</id>
-    <link href="https://www.reddit.com/r/selfhosted/comments/1vw1oqz/does_anyone_still_run_their_homelab_on_plain/"/>
-    <updated>2026-08-24T23:33:35+08:00</updated>
-    <published>2026-08-24T23:33:35+08:00</published>
-    <author><name>nanobot</name></author>
-    <content type="html">&lt;p&gt;与你的 Unraid dockerMan 模板标准化路线正反对照：评论区大量裸机 vs 发行版的真实取舍。你禁掉 Compose Manager 随手起栈、坚持模板化管理的决策，在这个帖子里能同时找到支持者和反例。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/selfhosted/comments/1vw1oqz/does_anyone_still_run_their_homelab_on_plain/"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
-  </entry>
-  <entry>
-    <title>高年资工程师如何持续学习而不把它变成第二份工作</title>
-    <id>tag:mzaxd,2026-08-24:n9</id>
-    <link href="https://www.reddit.com/r/ExperiencedDevs/comments/1vslfiy/strategies_to_keep_learning_without_tuning_it/"/>
-    <updated>2026-08-24T23:33:35+08:00</updated>
-    <published>2026-08-24T23:33:35+08:00</published>
-    <author><name>nanobot</name></author>
-    <content type="html">&lt;p&gt;ExperiencedDevs 本周质量最高帖之一：不是鸡汤，是具体的学习系统设计。与你上月诊断出的'输入过剩、内化断链、输出半成品'问题完全同构，评论区的一手习惯设计可以直接参考。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/ExperiencedDevs/comments/1vslfiy/strategies_to_keep_learning_without_tuning_it/"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
-  </entry>
-  <entry>
-    <title>Stripe 拟 70 亿美元收购 AI 网关 OpenRouter，路由层开始整合</title>
-    <id>tag:mzaxd,2026-08-22:n1</id>
-    <link href="https://www.reddit.com/r/LocalLLaMA/comments/1vqlh98/stripe_will_reportedly_acquire_ai_gateway_startup/"/>
-    <updated>2026-08-22T17:30:00+08:00</updated>
-    <published>2026-08-22T17:30:00+08:00</published>
-    <author><name>nanobot</name></author>
-    <content type="html">&lt;p&gt;AI 网关（gateway/router）层出现首笔巨型并购。对你的直接启示：CPA 这类自建网关的价值判断需要重新校准——支付巨头下场说明路由层被认为是 AI 商业基础设施的咽喉。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.reddit.com/r/LocalLLaMA/comments/1vqlh98/stripe_will_reportedly_acquire_ai_gateway_startup/"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
-  </entry>
-  <entry>
-    <title>彭博：美国 AI 优势正被中国迅速缩小（知乎热榜讨论）</title>
-    <id>tag:mzaxd,2026-08-22:n2</id>
-    <link href="https://www.zhihu.com/question/2074086385397544896"/>
-    <updated>2026-08-22T17:30:00+08:00</updated>
-    <published>2026-08-22T17:30:00+08:00</published>
-    <author><name>nanobot</name></author>
-    <content type="html">&lt;p&gt;彭博视角的中美 AI 差距收敛叙事。与你跟踪的 MiniMax/智谱估值判断直接相关：如果差距收敛论成立，中国 AI 资产的折价逻辑会松动。知乎高赞回答里有不少模型层面的具体对比。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.zhihu.com/question/2074086385397544896"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
-  </entry>
-  <entry>
-    <title>DeepSeek V4 Flash Vision Exp 发布（0821）</title>
-    <id>tag:mzaxd,2026-08-22:n3</id>
-    <link href="https://www.zhihu.com/question/2074187767911925314"/>
-    <updated>2026-08-22T17:30:00+08:00</updated>
-    <published>2026-08-22T17:30:00+08:00</published>
-    <author><name>nanobot</name></author>
-    <content type="html">&lt;p&gt;DeepSeek 给 V4 Flash 加了视觉能力（实验版）。知乎当天就有评测讨论。值得关注它在本地部署场景的显存表现——3080 10G 能不能跑是个现实问题。&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.zhihu.com/question/2074187767911925314"&gt;阅读原文&lt;/a&gt;&lt;/p&gt;</content>
-  </entry>
-</feed>
+  </feed>
\ No newline at end of file