Source-backed discovery
New Term Radar
Discover names and concepts appearing in public sources, with dated evidence, source coverage and comparable growth.
Saved terms stay in this browser and do not sync to your account. Source collection continues on the server without this page being open.
Collection coverage · Comparable windowsCurrent / previous observed days 7 / 7
Counts use complete UTC days. First observed means first seen by this collection, not when a term was coined.
Current window：2026-10-31 → 2026-11-07 (end excluded) · History 19 days
| Source | Current days | Previous days | Last success |
|---|---|---|---|
| arXiv | 6 | 5 | Nov 7, 2026, 12:45 AM UTC |
| arXiv | 7 | 4 | Nov 7, 2026, 5:30 AM UTC |
| Ask HN | 7 | 5 | Nov 7, 2026, 4:30 PM UTC |
| Ask HN | 7 | 5 | Nov 7, 2026, 4:30 PM UTC |
| Ask HN | 7 | 5 | Nov 7, 2026, 4:45 PM UTC |
| Ask HN | 7 | 5 | Nov 7, 2026, 4:45 PM UTC |
| GitHub | 7 | 5 | Nov 7, 2026, 1:15 PM UTC |
| GitHub Issues | 7 | 5 | Nov 7, 2026, 4:15 PM UTC |
| GitHub Releases | 7 | 3 | Nov 7, 2026, 12:45 PM UTC |
| GitHub | 7 | 5 | Nov 7, 2026, 2:45 PM UTC |
| GitHub Issues | 7 | 5 | Nov 7, 2026, 5:15 PM UTC |
| GitHub Releases | 7 | 5 | Nov 7, 2026, 5:15 PM UTC |
| GitHub | 7 | 5 | Nov 7, 2026, 2:45 PM UTC |
| GitHub Issues | 7 | 5 | Nov 7, 2026, 5:30 PM UTC |
| GitHub Releases | 7 | 5 | Nov 7, 2026, 5:15 PM UTC |
| GitHub Discovery | 2 | 0 | Nov 7, 2026, 4:45 PM UTC |
| GitHub Trending | 7 | 5 | Nov 7, 2026, 8:00 PM UTC |
| 7 | 4 | Nov 7, 2026, 8:45 AM UTC | |
| Hugging Face | 7 | 4 | Nov 7, 2026, 8:45 AM UTC |
| Hugging Face Spaces | 2 | 0 | Nov 7, 2026, 10:30 PM UTC |
| Hacker News | 7 | 5 | Nov 7, 2026, 8:00 PM UTC |
| Hacker News | 7 | 5 | Nov 7, 2026, 4:45 PM UTC |
| Hacker News | 7 | 5 | Nov 7, 2026, 4:45 PM UTC |
| Hacker News | 2 | 0 | Nov 7, 2026, 4:45 PM UTC |
| Show HN | 2 | 0 | Nov 7, 2026, 4:45 PM UTC |
| MIT News | 7 | 4 | Nov 7, 2026, 5:45 PM UTC |
| OpenAI | 7 | 5 | Nov 7, 2026, 11:45 AM UTC |
| Product Hunt | 7 | 5 | Nov 7, 2026, 8:00 PM UTC |
| 量子位 | 6 | 5 | Nov 6, 2026, 7:00 PM UTC |
| OpenRouter | 2 | 0 | Nov 7, 2026, 10:05 PM UTC |
| GitHub Trending | 7 | 7 | Nov 6, 2026, 1:15 PM UTC |
| Hugging Face Models | 2 | 0 | Nov 7, 2026, 10:05 PM UTC |
| HF Daily Papers | 2 | 0 | Nov 7, 2026, 10:05 PM UTC |
| OpenRouter | 2 | 0 | Nov 7, 2026, 10:05 PM UTC |
| Product Hunt | 7 | 7 | Nov 6, 2026, 1:15 PM UTC |
| SEC EDGAR | 7 | 5 | Nov 7, 2026, 10:45 AM UTC |
| Sequoia Capital | 7 | 4 | Nov 7, 2026, 6:45 AM UTC |
| Simon Willison | 2 | 0 | Nov 7, 2026, 3:30 PM UTC |
| TechCrunch | 7 | 5 | Nov 7, 2026, 9:45 PM UTC |
| Contracts Finder | 7 | 5 | Nov 7, 2026, 8:15 PM UTC |
| V2EX | 7 | 5 | Nov 7, 2026, 9:45 PM UTC |
- Counts describe monitored public sources, not the whole web or search demand.
Seven-day figures use the complete UTC window above. Priority is a research triage score, not search volume, market size or a probability of success.
那一步
最近给框架发版时总卡在 Maven Central 发布那一步：一次要推送 90 多个工件、加起来超过 300MB ，之前一直用的插件走 OSSRH 路径，在这个体积下反复超时，要么手动补发、要么整段重来，非常折腾。 按照 Maven Central 官方的 Publisher API 文档 ，我用 Java 实现了一个 Gradle 发布插件，把正式版收敛成 staging → bundle → 上传 Portal → 轮询部署状态一条命令；又补上了 SNAPSHOT 直接发布，根项目版本号带 -SNAPSHOT 后缀就自动切快照仓库，不用再在两个工具间来回切。 两个自认为对多模块项目有用的取舍： 根项目聚合发布 ：一次扫描所有子项目，把它们的 publication 合并进同一次部署，避免同一版本拆成多个部
First observed Oct 22, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Surfaces
First observed Oct 12, 2026 · Hacker News
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
本地
First observed Oct 10, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
重构
First observed Oct 13, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Warp Science
Wearing cameras and sensor gloves, the startup’s human chemists will train robots to perform various lab tasks.
First observed Oct 22, 2026 · TechCrunch
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Notes
First observed Oct 20, 2026 · Hugging Face
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
CultureLens
arXiv:2610.27202v1 Announce Type: new Abstract: Large Language Models (LLMs) increasingly operate across culturally diverse settings, yet existing evaluations largely measure cultural behavior at the output level. We introduce CultureLens, a reference-grounded benchmark and mechanistic study of how LLMs use cultural information. CultureLens contains 9,261 sc
First observed Oct 22, 2026 · arXiv
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Windows PCs
First observed Oct 12, 2026 · Hacker News
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
PrefDiag-Bench
arXiv:2610.30573v1 Announce Type: new Abstract: Unlike conventional GUI agents, personalized GUI agents require user preferences to guide sequential execution over changing GUI states. This raises two challenges: determining HOW each preference should alter execution strategy and WHEN it should take effect along the interaction trajectory. Existing benchmark
First observed Oct 23, 2026 · arXiv
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Koody
Hi HN, I'm Hallie, founder of Koody. Today, I'm launching Koody for Business: AI-powered accounting, bookkeeping, and tax prep software for sole proprietors and small businesses. For years, I used accounting software for my business. I went into it to create invoices, categorize and reconcile transactions, and work with my accountant to prepare our books. I
First observed Oct 20, 2026 · Hacker News
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
一人
First observed Oct 16, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
保护视力
《摄影学英语》闭眼拼写 功能已经上线 保护视力，无需查看屏幕 练习听力，不看屏幕磨耳朵 练习口语，用声音代替键盘输入 支持用耳机听，用领夹麦克风讲话，超级好的收音体验 本地识别，速度快，成本低，没网也能用。 记忆在于重复，碎片化时间学习，参加无聊的活动时，不看屏幕强化记忆 睡前，闭眼拼单词，拼一会儿就睡着了 《摄影学英语》 App Store 下载地址 https://apps.apple.com/cn/app/id6808490052 其它特色功能：无需打开 App 用桌面小组件也能背单词 《摄影学英语》其它特色功能： 做了款摄影学英语 App ，用老照片里的场景学英语 做了款肉鸽游戏背单词的 App 欢迎 V 友们品鉴！
First observed Oct 25, 2026 · V2EX
- 7d mentions
- 0 previous 1
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
时自行添加有权使用
这是我自己做的小产品 Dime Dime Baby AI ，主要面向想做 TikTok / Reels 热梗视频的海外用户，在推广节点介绍一下。 用法是上传一张有使用授权的成年人照片，选择清晰度和时长，生成后在 My Videos 下载，不用自己写提示词。现在支持英文、西语；提供 480p / 5 秒，以及 720p / 5 秒、10 秒。 收费是一次性积分包，无订阅。入门包 $9.90 / 10 积分，可以生成 1 条 720p / 5 秒，或 2 条 480p / 5 秒的视频。不是免费生成工具。 下载的是无声 MP4 ，音乐需要发布时自行添加有权使用的音源。AI 结果可能有面部细节、手部或动作偏差；仅使用本人或获得授权的普通成年人照片。 网站（带 V2EX 来源标记）： https://dimedimeb
First observed Oct 24, 2026 · V2EX
- 7d mentions
- 0 previous 1
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
修复公告
昨晚发布的 PHP 的 PHP2Shell 漏洞（ CVE-2026-109618 、CVE-2026-109619 ）让很多运维和安全部门睡不着觉了，属于严重的远程代码执行漏洞。如果服务器使用了受影响的 PHP 运行环境，可能存在较高风险。 宝塔官方已经发布修复公告，PHP 7.4 、8.0 、8.1 也有对应的宝塔修复版本，不一定非要立即迁移到 PHP 8.2 以上。 宝塔修复方法： 登录宝塔面板，进入「软件商店 → 已安装」。 点击右上角「更新软件列表」。 找到网站实际使用的 PHP 版本，检查并更新到对应的安全修复版本。 更新后确认版本，并重启相关 PHP 服务，检查网站运行是否正常。 官方公告及详细操作： https://mp.weixin.qq.com/s/pJ6mHOR01XEQlqLus2lwp
First observed Oct 21, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Bulin
First observed Oct 14, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Speculative Evolution
arXiv:2610.28706v1 Announce Type: new Abstract: LLM-driven evolutionary search is a promising paradigm for scientific discovery, but existing methods remain costly and slow, often applying expensive reasoning uniformly while underusing evidence from failed, redundant, and saturated trials. We introduce Speculative Evolution (SpecEvo), a multi-LLM framework t
First observed Oct 22, 2026 · arXiv
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
智能指数榜首
我认为 DeepSeek 、MiniMax 、Kimi 、GLM 这些 LLM 公司被严重高估了： 月活（ MAU ） ： 豆包 4.2 亿，千问 1.7 亿，DeepSeek 1.2 亿，智谱 2000 万，Kimi 1000 万； 商业化（ ARR ） ： 豆包 272 亿，智谱 68 亿、阿里 80 亿、DeepSeek 27-34 亿、Kimi 20 亿； 开源贡献 ： Qwen：开源了超过 460 个模型，衍生模型超 30 万个，全球下载量破 20 亿次； GLM：开源 50 余款模型，GLM-5.2 以 51 分位居全球开源智能指数榜首； DeepSeek：开源了 DeepSeek-V4 、DeepSeek-R1 、DeepSeek-VL 等模型。 豆包的战略和 Google 的 Gemini 是很
First observed Oct 24, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
送码第二波
Vonode：4G 模块插在家里一台常开的 Linux 小主机（ N100 、树莓派都行）上，卡留在家，iPhone 上收发短信、打电话、管号码。App 1.4.2 已上架，节点更新到 v1.4.3 。 现在能做什么 一台节点最多插 5 个 4G 模块，多个号码一个 App 管 收发短信，新短信推送到 iPhone ，验证码短信不会被误拦 用自己的号码打电话、接电话（走 Wi-Fi 通话，需运营商支持），拨号盘联想联系人 eUICC 卡上的 eSIM 号码远程切换 保号：显示每个号码还剩多少天，快到期提醒，支持定时自动发短信 / 打电话 骚扰短信过滤、发件人规则，来电黑白名单（任何国家号码），被拦原因一目了然 节点离线 / 恢复推送提醒，一键网络体检 App 备份：换新 iPhone 直接导入 中、英、日、韩界
First observed Oct 21, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
免费数据库客户端
First observed Oct 13, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
运维平台
大家好，8 月底在 V2EX 分享过自己开发的开源运维平台 Ops Admin 。距离上次发帖快两个月了，最近一直在结合日常运维中的需求补功能、修问题，来汇报一下进展。 项目地址： https://github.com/qishu321/ops-admin 项目网站： https://qishu321.github.io/ops-admin-website/ 上次的介绍： https://www.v2ex.com/t/1238337 简单介绍一下：Ops Admin 基于 Go + Vue 3 ，主要面向个人运维和中小型运维团队，把主机、数据库、Kubernetes 、批量执行、作业编排、监控告警和构建发布放到一个工作台里，也支持通过 SSH 网关访问内网资源。 项目截图 界面截图 下面是这次更新后的几张界面截
First observed Oct 22, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
这两年
这两年，用 AI 写文章的人越来越多了。 不同平台对内容风格的要求不同，用户对 AI 生成内容的接受程度也不一样。“去 AI 味”逐渐成了一种需求，甚至有人把相关方法封装成了各种 Skill 。 昨天看到 Artificial Analysis 发布的一张不同平台内容风格偏好的雷达图，觉得挺有意思。以前我们追求文章逻辑清晰、结构完整，现在这些特征反而可能成为被怀疑是 AI 写的理由。 于是我参考图中的 8 个维度，又整理了一张人类与 AI 写作特征的通用对比图。当然，这只是基于一些常见印象做的概念性比较，并不是什么严谨的研究结论。 画完图，忽然之间手就痒了，索性让 AI 帮忙做了一个「 AI 味」趣味鉴定小游戏，顺手开源到了 GitHub 。 在线体验： https://dare-creation-taste-
First observed Oct 21, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Agent4Sci
arXiv:2610.28763v1 Announce Type: new Abstract: Frontier coding agents are increasingly framed as end-to-end automated researchers, and ML for science offers a demanding test of that promise: improving a scientific model often requires domain insight beyond routine ML engineering. We introduce Agent4Sci, a long-horizon testbed for measuring how far today's a
First observed Oct 22, 2026 · arXiv
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Agent落地报告
First observed Oct 16, 2026 · 量子位
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
放到工作台里
大家好，8 月底在 V2EX 分享过自己开发的开源运维平台 Ops Admin 。距离上次发帖快两个月了，最近一直在结合日常运维中的需求补功能、修问题，来汇报一下进展。 项目地址： https://github.com/qishu321/ops-admin 项目网站： https://qishu321.github.io/ops-admin-website/ 上次的介绍： https://www.v2ex.com/t/1238337 简单介绍一下：Ops Admin 基于 Go + Vue 3 ，主要面向个人运维和中小型运维团队，把主机、数据库、Kubernetes 、批量执行、作业编排、监控告警和构建发布放到一个工作台里，也支持通过 SSH 网关访问内网资源。 项目截图 界面截图 下面是这次更新后的几张界面截
First observed Oct 22, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
MEDIATOR
arXiv:2610.27047v1 Announce Type: new Abstract: EHR foundation models promise scalable clinical prediction across health systems, yet their deployment is constrained by a tokenization bottleneck. The same clinical concept can appear as different ontology codes, local identifiers, synonyms, or free-text descriptions, which standard EHR tokenizers treat as unr
First observed Oct 22, 2026 · arXiv
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
自建邮局
First observed Oct 20, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Procedural Atoms
arXiv:2610.26926v1 Announce Type: new Abstract: Large language models (LLMs) possess broad factual knowledge and strong reasoning capabilities, but solving complex tasks also requires procedural knowledge about how to act. Existing approaches often represent procedural knowledge through trajectories or bundled skills, limiting flexible reuse across interacti
First observed Oct 22, 2026 · arXiv
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
Flint
Since we last built Flint earlier this year ( https://news.ycombinator.com/item?id=48834924 ), we have been wondering if interactive charts can also find they home in terms of simple high-level specifications to allow AI agents to produce them easily, reliably and with high-quality. I have always been amazed by NYT interactive stories back in school, as they
First observed Oct 21, 2026 · Hacker News / TechCrunch
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
国考选岗工具后
First observed Oct 19, 2026 · V2EX
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100
-
- 7d mentions
- 0 previous 0
- 7d sources
- 0
- 7d growth
- Not comparable
- Priority
- 8 / 100