issue 2026-09-08
This commit is contained in:
@@ -2,8 +2,89 @@
|
||||
<feed xmlns="http://www.w3.org/2005/Atom">
|
||||
<title>Spring 编辑精选</title>
|
||||
<id>tag:mzaxd,2026:newsroom</id>
|
||||
<updated>2026-09-03T23:34:39+08:00</updated>
|
||||
<updated>2026-09-08T07:35:00+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>
|
||||
<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/"/>
|
||||
<updated>2026-09-08T07:35:00+08:00</updated>
|
||||
<published>2026-09-08T07:35:00+08:00</published>
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>Framework 悄悄上架 192GB 内存 SKU，按现有 32/64/128 定价梯度推算板卡约 4.5k 美元档。结合上期腾讯把 Hy4-preview 压到约 200GB GGUF，一台消费级笔电装下「准满血 MoE」在下半年成为现实选项，本地推理硬件成本曲线继续下移。对比 8-27 期 Mac Studio 512GB 路线，x86 阵营终于有了答牌——对「第二台本地推理机」的选型窗口可以再等等。</p><p><a href="https://www.reddit.com/r/LocalLLaMA/comments/1w28x8u/its_official_192gb_framework/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<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/"/>
|
||||
<updated>2026-09-08T07:35:00+08:00</updated>
|
||||
<published>2026-09-08T07:35:00+08:00</published>
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>Muse Spark 预告即将开放权重，社区正从命名规律反推参数量（122B+ 呼声最高）。GLM/Qwen/DeepSeek 三强之外若再进一个新玩家，开源模型的供给端竞争会更卷——发布落地时别只看跑分，先看 GGUF 生态跟进速度，那才是本地部署可用性的真实指标。</p><p><a href="https://www.reddit.com/r/LocalLLaMA/comments/1w5l8bw/muse_spark_open_weights_comming_soon/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<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/"/>
|
||||
<updated>2026-09-08T07:35:00+08:00</updated>
|
||||
<published>2026-09-08T07:35:00+08:00</published>
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>Pangolin 1.22 新增 AI Gateway 资源类型：在云端模型 API 和自托管模型服务前挂一层身份感知代理，统一鉴权、路由、审计。这正对上你自建网关的思路——把「云模型 + 本地 GLM」混布成一套入口。结合 8-22 期 Stripe 拟收购 OpenRouter，模型路由/网关层正从玩具变成 infra 标配，值得纳入技术雷达。</p><p><a href="https://www.reddit.com/r/selfhosted/comments/1w0s16a/pangolin_122_ai_gateway_for_selfhosted_models/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<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/"/>
|
||||
<updated>2026-09-08T07:35:00+08:00</updated>
|
||||
<published>2026-09-08T07:35:00+08:00</published>
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>CrowdSec 1.8 在 WAF 之上加了 bot 检测，社区情报驱动的自托管防护栈又补一块拼图。对你直接相关：DERP 中继、NPM、Gitea 这些公网可达入口的日常防御，它比单机 fail2ban 省心得多。本周 selfhosted 版还在热讨「你对自托管安全多认真」（起因是澳 AFP/FBI 端掉被滥用的自托管服务）——公网暴露面管理值得排进 infra 待办。</p><p><a href="https://www.reddit.com/r/selfhosted/comments/1w5botz/crowdsec_v18_just_released_self_hosted_idsipswaf/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<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/"/>
|
||||
<updated>2026-09-08T07:35:00+08:00</updated>
|
||||
<published>2026-09-08T07:35:00+08:00</published>
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>Planka 把 SSO 挪进付费层引发众怒后，社区整理出「免费支持 SSO/OIDC 的自托管应用清单」。你的服务矩阵（Vaultwarden/Gitea/Miniflux…）正处在该不该统一入口的阶段，这份清单可直接当选型依据：先盘哪些服务原生支持 OIDC，再决定 Authelia/Authentik 值不值得为剩下的铺一套。</p><p><a href="https://www.reddit.com/r/selfhosted/comments/1w47nrq/here_is_a_list_of_self_hosted_apps_support_sso/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<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/"/>
|
||||
<updated>2026-09-08T07:35:00+08:00</updated>
|
||||
<published>2026-09-08T07:35:00+08:00</published>
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>Dawarich 把百万级轨迹点的渲染压到秒级，自托管位置历史/GPS 轨迹项目的活跃度持续超预期。对你的健康 App 方向是竞品情报：它证明「自己持有数据 + 历史轨迹可视化」有真实需求池，但同类（含 8-22 期 SparkyFitness）都停在工具型记录，还没人把轨迹/体征数据接进健康洞察闭环——这个缺口正是差异化空间。</p><p><a href="https://www.reddit.com/r/selfhosted/comments/1w56g1u/dawarich_1141_now_rendering_millions_of_points_in/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<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/"/>
|
||||
<updated>2026-09-08T07:35:00+08:00</updated>
|
||||
<published>2026-09-08T07:35:00+08:00</published>
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>新 tech lead 全面用 AI 审 PR、高强度产出评审意见，团队开始消化不了这种「评审通货膨胀」。有价值的不是槽点，而是评论区一线工程师正在形成的边界共识：AI 适合做 PR 的第一道粗筛，final call 必须有人签字。AI 进入流程的顺序和权责划分，是接下来一年工程组织的核心议题。</p><p><a href="https://www.reddit.com/r/ExperiencedDevs/comments/1w5ifk8/how_to_deal_with_tech_lead_that_uses_ai_for_pr/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<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/"/>
|
||||
<updated>2026-09-08T07:35:00+08:00</updated>
|
||||
<published>2026-09-08T07:35:00+08:00</published>
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>多年 genAI 实践者的诚实复盘：产出仍像老虎机——不可控、不可复现，和「2025 年后再没写过一行代码」的加速主义叙事完全是两个世界。这种一手体感是校准 AI 预期的宝贵样本：模型能力在涨，但确定性没有同步涨，把运气变成产出靠的还是工程化收敛（测试、评审、沙箱）。与 8-22 期「职业往哪走」讨论对照食用。</p><p><a href="https://www.reddit.com/r/ExperiencedDevs/comments/1w1l8v4/have_been_using_genai_for_a_few_years_now_and_it/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<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/"/>
|
||||
<updated>2026-09-08T07:35:00+08:00</updated>
|
||||
<published>2026-09-08T07:35:00+08:00</published>
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>周内要发版的代码 AI 写的、只有一个人写+测+懂业务逻辑，发版前才有人认真 review，发帖人问「要不要推迟，AITA」。评论一边倒，但真正值得看的是事故结构：AI 让单人产出大量代码变容易后，review 带宽成了新瓶颈，流程不跟上就是裸奔。对个人项目同理——自动测试 + CI 门禁的价值在 AI 时代只升不降。</p><p><a href="https://www.reddit.com/r/ExperiencedDevs/comments/1w3gfx4/release_this_weekend_no_code_review_so_far_i/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<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>
|
||||
@@ -283,49 +364,4 @@
|
||||
<author><name>nanobot</name></author>
|
||||
<content type="html"><p>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。</p><p><a href="https://www.zhihu.com/question/2074187767911925314">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<entry>
|
||||
<title>Qwen3.8-27B 本周刷屏：Unsloth Dynamic v3 量化 + 基准对齐 DeepSeek</title>
|
||||
<id>tag:mzaxd,2026-08-22:n4</id>
|
||||
<link href="https://www.reddit.com/r/LocalLLaMA/comments/1vqyq8r/artificial_analysis_qwen3827b_benchmarks_put_it/"/>
|
||||
<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"><p>Qwen3.8-27B 是本周 LocalLLaMA 主旋律：Artificial Analysis 基准把它排到与 DeepSeek 同档，Unsloth 的 Dynamic v3 量化同尺寸精度提升 10%。另有帖子称它是'本地模型里 agency 最强'——对你折腾本地 agent 有直接参考价值。量化版：reddit.com/r/LocalLLaMA/comments/1vsr67c</p><p><a href="https://www.reddit.com/r/LocalLLaMA/comments/1vqyq8r/artificial_analysis_qwen3827b_benchmarks_put_it/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<entry>
|
||||
<title>内存价格 12 个月上涨 500%</title>
|
||||
<id>tag:mzaxd,2026-08-22:n5</id>
|
||||
<link href="https://www.reddit.com/r/LocalLLaMA/comments/1vrwsfl/memory_prices_climb_500_in_12_months_up_to_10x/"/>
|
||||
<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"><p>DRAM 涨价已到历史级，AI 数据中心需求抽干了消费级供给。短期信号：想加内存条/换 NAS 内存趁早，2026 年内看不到缓解。</p><p><a href="https://www.reddit.com/r/LocalLLaMA/comments/1vrwsfl/memory_prices_climb_500_in_12_months_up_to_10x/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<entry>
|
||||
<title>SparkyFitness v1.6.2：自托管 MyFitnessPal 替代（5.2k stars）</title>
|
||||
<id>tag:mzaxd,2026-08-22:n6</id>
|
||||
<link href="https://www.reddit.com/r/selfhosted/comments/1vqe7qg/sparkyfitness_v162_a_selfhosted_alternative_for/"/>
|
||||
<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"><p>开源饮食/健康追踪，98 人贡献。与你的 AI 饮食健康 App 方向直接对标——建议拉下来看看它的数据模型和 AI 功能边界，竞品调研成本几乎为零。</p><p><a href="https://www.reddit.com/r/selfhosted/comments/1vqe7qg/sparkyfitness_v162_a_selfhosted_alternative_for/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<entry>
|
||||
<title>「用 AI 写代码之后，这个职业到底在往哪走？」——ExperiencedDevs 本周高热讨论</title>
|
||||
<id>tag:mzaxd,2026-08-22:n7</id>
|
||||
<link href="https://www.reddit.com/r/ExperiencedDevs/comments/1vq8w8c/whats_wrong_with_our_team/"/>
|
||||
<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"><p>7 年经验工程师的真实困惑：AI 写码后个人价值锚点在哪。评论区有大量一线工程师的应对策略。与《The Pragmatic Engineer》近期 Meta 裁员潮报道互为印证——工程职业结构正在换挡。</p><p><a href="https://www.reddit.com/r/ExperiencedDevs/comments/1vq8w8c/whats_wrong_with_our_team/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
<entry>
|
||||
<title>非科技公司升 Staff 工程师 8 个月复盘</title>
|
||||
<id>tag:mzaxd,2026-08-22:n8</id>
|
||||
<link href="https://www.reddit.com/r/ExperiencedDevs/comments/1vpg59k/followup_what_this_year_has_been_like_as_a/"/>
|
||||
<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"><p>大型非科技公司（接近你司环境）从 Senior 升 Staff 的完整复盘：晋升后的期望错位、技术影响力如何落地。对判断'走管理还是走专家'有第一手参照。</p><p><a href="https://www.reddit.com/r/ExperiencedDevs/comments/1vpg59k/followup_what_this_year_has_been_like_as_a/">阅读原文</a></p></content>
|
||||
</entry>
|
||||
</feed>
|
||||
|
||||
Reference in New Issue
Block a user