Telegram RegisterThe public register of Telegram

Channel

Laisky's Notes

@laiskynotes

On this record: Growth · Engagement · Reactions · Posts · Citations · Cite this entry

3,291subscribers

+3 since we began measuring on 7 August 2026

Risers and fallers across the register · movement among entries of 3,162–10,000.

Register entry

Telegram ID-1001579892928
TypeChannel
Username@laiskynotes
CreatedBetween 1 August 2021 and 31 January 2023— estimated from Telegram’s id allocation, not measured. How this range is calculated.
First recorded7 August 2026
Last confirmed live11 August 2026
Measurements held4
Confirmed unchanged1 time, most recently 11 August 2026
On Telegramt.me/laiskynotes

Growth

3,2883,2913,289.57 August 2026 — 3,288 subscribers7 August 2026 — 3,288 subscribers8 August 2026 — 3,290 subscribers11 August 2026 — 3,291 subscribers7 August 202611 August 2026
4 measurements spanning 4 days, net +3. Dots are measurements; the straight line between them is drawn to join them, not to claim we know the path taken in between — snapshots are recorded only when a count changes, so gaps mean “no change observed”, never “interpolated”. The vertical axis spans 3,288–3,291 and does not start at zero.
Measurement log — every subscribers count we have recorded
Measured (UTC)SubscribersChange
11 Aug 2026, 03:343,291+1
8 Aug 2026, 07:133,290+2
7 Aug 2026, 14:183,288no change
7 Aug 2026, 14:073,288first reading

Engagement

19 posts held, back to 17 October 2025the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 2 pagesof Telegram’s post history, 20 posts per page.

Nothing published in the last 30 days. ERR and ER are rolling 30-day measures, so there is nothing to compute — we hold 19 posts for this entry, the most recent from 5 June 2026. An engagement rate over an empty window would be a number about nothing.

Reaction mix

385 reactions across 17 posts, in 10 distinct kinds. The most used accounts for 41.6% of them.

Every reaction kind recorded on the sample, most used first
ReactionCountShareShare, drawn
👍16041.6%
13434.8%
😢4511.7%
👏225.71%
🔥143.64%
😨51.30%
👎20.519%
🐳10.26%
🤔10.26%
🤯10.26%

No sentiment is inferred, and none should be read in. This table is ordered by count and by nothing else. Emoji do not carry stable meaning across languages or communities — 🙏 is thanks in one channel and mourning in another — so we publish which ones were pressed and how often, and pass no judgement on what an audience meant by them.

Precision. Telegram publishes reaction counts per emoji and short-forms each one — 4.34K, 1.2M — so any single kind at or above 1,000 reaches us at three significant figures, and only counts below 1,000 are exact. The shares above are ratios of those figures and carry the same error. This is also why the total here can differ slightly from a reaction total printed elsewhere on the page: both are sums of the same rounded parts, taken over samples with different edges.

Coverage. Reactions were read on 17 of the 19 sampled posts in this sample. Summed by Telegram’s own count on each post — not by adding up the per-emoji breakdown above — those same posts carry 385reactions in total: the kind of figure the paragraph above means by “a reaction total printed elsewhere on the page”.

Measured over the 19 most recent posts we hold, published 17 October 2025 to 5 June 2026, using the newest reading held for each. Telegram Stars are excluded: they are a payment, not a reaction, and they have their own section.

Recent posts

5 Jun 2026, 15:10 UTC≈1,650 views22 reactionsread 7 August 2026
Photo

读完 Mario Puzo 的《教父 The Godfather》三部曲小说,才意识到自己当年年轻时根本完全就没看懂电影《教父》。然后又去把电影重温了一遍,才发现原来电影里有很多细节和暗示是我之前完全没有注意到的,而小说中的情节实际上也比电影更为丰富和复杂。 首先说小说原著,实际上小说三部曲互相之间并没有前后联系,更像是三本相同世界观背景下的三个独立故事。唐·科里昂家族的故事仅出现在第一部,而电影也是将这第一本小说的内容改编成了三部电影,电影第一部和第二部选择了原著第一本的部分情节,第三部电影则是自行创作延续了小说第一部的故事线,讲述了迈克尔·科里昂在成为家族领袖后的故事。 所以我们在谈论教父时,主要关注的是教父小说第一部的内容。这本书讲述了唐·科里昂如何从一名贫困的意大利裔美国移民,逐步崛起建立起自己的黑手党家族,后来因为和新崛起的毒品势力发生冲突而遭受重创,最后家族在第二代家主迈克尔的带领下重新崛起,却又在巅峰期激流勇退

19👍3

22 May 2026, 03:16 UTC≈7,470 views63 reactionsread 7 August 2026

记录一下最近遇到的一个糟心事儿。 最近两个月开始又开始找工作机会,总的来说很不顺利,托朋友联系了两个内推也是石沉大海。(各位要是有远程工作的机会也请联系 job@laisky.com ) 昨天在 Linkedin 上有个 web3 背景的猎头联系我说有个不错的 contract 工作,就是需要先完成一个 code assessment。这个测试放在 notion 上,指向一个 github repo。但实际上,只要你在本地运行 npm start,就会启动恶意脚本,偷偷下载一个 js 脚本并执行。 这个攻击手法相当低劣,低级且恶劣,低级到了随手搜一下 eval 就能发现。 各位都小心一点,不要在本地运行任何来历不明的代码。即使源码里没有搜出可疑的内容也并不可靠,完全可以把攻击代码放在第三方依赖里。vscode 用户可以使用 devcontainer 来执行这些临时项目,能极大的缓解威胁。

😢45👍11😨52

14 May 2026, 21:32 UTC≈3,800 views66 reactionsread 7 August 2026

作为一个从业十多年的程序员,我已经半年没手写过一行代码了。上个月大概提交了 20 万行,这个月大概 4 万行。 昨天认真考虑了一下这个现象,我个人认为,虽然 AI 仍然存在显然的不足与缺陷,在非主流领域的代码水平不高甚至有些低劣,但是程序员确实不应该再执着于手写代码和 review 了,自动化、工业化的星辰大海就在眼前,有太多值得探索和尝试的领域。 • 不要再自视优越地吐槽 AI 代码质量差,而是应该探索如何生成高质量的代码。 • 不要再抱怨 AI 代码太多 review 不过来,而是应该探索工业规模的测试和 review 方案。 我认为从 opus-4.7/gpt-5.4 开始,AI 的能力到达了一个分水岭,它能够在严格的约束下,完成它本身并不擅长和理解的工作。可以称之为 harness,不过我不喜欢使用这些转瞬即逝的 fancy words。 软件工程师,作为一个工程师,关注的应该是在限定的资源和时间内,找出和解决问

👍4617👎2🤔1

5 May 2026, 21:45 UTC≈2,700 views5 reactionsread 7 August 2026

https://www.notion.so/laisky/GitHub-RCE-Vulnerability-CVE-2026-3854-Breakdown-Wiz-Blog-352ba4011a8681c3afdfe340d1a76fd5?source=copy_link 上周看到这个号称仅用 git push 就实现了对 GitHub 服务器的远程控制。这篇文章介绍了攻击原理,没想到实际上这么草台。 网关服务 babeld 接收用户请求,进行身份验证后,就会设置 X-Stat header 作为内部服务通信的关键权限信息。然而,X-Stat 作为重要的内部属性,居然允许直接拼接外部的用户输入,而且在拼接时没有进行 sanitize 处理,导致攻击者可以简单的使用 ; 来实现任意注入。 这么重要的内部权限字段,居然接受明文的用户输入。这种低级错误应该连 GitHub 免费提供的 CodeQL 都能扫出来,太不应该了。

5

29 Apr 2026, 13:58 UTC≈2,440 views16 reactionsread 7 August 2026

https://laisky.notion.site/VMware-e1000-vmxnet3-34cba4011a8680ea9242e1f93d990b17?source=copy_link 一个小笔记。我家里用一台 windows 11 主机,通过 vmware workstation pro 运行了一些虚拟机,然后通过 tailscale 和我的其他 VPS 打通。 近期某些虚拟机在网络流量较大时会出现网络中断,表现为一个 CPU 完全跑满,外网延迟增加到 15s 以上,重启 tailscale 后能够恢复。 让 codex app 去帮我修复,它将 e1000 网卡修改为了 vmxnet3,目前已经用了一周左右,很稳定,没有再出现问题。整个故障的排查、修复和记笔记都是 codex app 完成的,不得不说 codex app 完美解决了我不懂也不想折腾 windows 的局限😂。

🔥142

13 Mar 2026, 04:04 UTC≈5,810 views70 reactionsread 7 August 2026

简单分享下我是怎么花最少的钱,但是又非常重度、高频的依赖 AI coding 的。 我目前的订阅是,最便宜的 copilot 和 codex。这两个的特点在于:codex 是按量付费的,而 copilot 是按次付费的。这个区别是最关键的,如果你更喜欢 cc,那么你可以把本文中的 codex 替换为 cc,一样的逻辑。 而作为 coding,其实工作可以拆分为两大类型,我相信其他大部分文字工作也可以这么拆分: 1. 计划阶段:此时会涉及非常高频交流,但是每次交流的内容都不多,你可以使用 codex 的最高思考模式,写下最核心的关键逻辑 2. 实施阶段:这是广义的实施,包括收集资料、撰写文档、coding、testing、debug 反复调试 都可以视为实施。 copilot 的按次付费,使它成为实施阶段的最佳性价比选择。发送一次 prompt 算一次费用,只要 agent loop 没有结束,期间一切的 tools 调用

49👍21

9 Mar 2026, 13:38 UTC≈2,770 views13 reactionsread 7 August 2026

https://developer.1password.com/docs/service-accounts/use-with-1password-cli/ 最近折腾的一件事就是把服务器部署的密钥全部用 1Password 管理了。 以前的做法比较粗糙,我有一个 private github repo 专门存放各种密钥,每台主机上都在固定的位置 clone 这个 repo。 现在我在 1Password 上创建了一个 vault 专门存放这些密钥,然后为每一台主机创建一个 service account,给它访问这个 vault 的权限。每台主机上安装 1Password CLI,并使用对应的 service account 登录,就可以用下面的形式定义密钥了。 SESSION_SECRET=op://configs/oneapi/SESSION_SECRET SQL_DSN=op://configs/oneapi/SQL

👍93🤯1

18 Feb 2026, 14:24 UTC≈3,650 views11 reactionsread 7 August 2026

https://laisky.notion.site/Effective-context-engineering-for-AI-agents-Anthropic-2ddba4011a868170ac2ddd9017afadde?source=copy_link Anthropic 介绍 context engineering 的文章。 - prompt engineering 主要关注如何撰写 system prompt。 - context engineering 则关注 agent 相关的所有工程问题。 注意力是稀缺的,context 过长,反而会导致模型能力下降(context rot)。 所谓 Agent,就是在 loop 内自动调用工具的 LLM。 context 注入的两种方式: 1. 在推理前,利用 RAG 检索和注入 context(延迟低) 2. 在 Agent 运行的过程中,通过 tools 自动

8👍3

18 Feb 2026, 14:16 UTC≈2,070 views6 reactionsread 7 August 2026

https://laisky.notion.site/Building-Effective-AI-Agents-Anthropic-2ddba4011a868146b90bf176d265cf81?source=copy_link 虽然人们喜欢说 agent 取代了 workflow,但是 agent 本身其实也是由一系列的 workflow 最小功能单元组成的。 区别在于,workflow 是由固定的逻辑和工具组成的。而 agent 则是由 LLM 动态选择工具和生成逻辑组成的。 本文介绍了用来构建 agent 的 workflow 功能单元: - chain: 最常见的串行流水线 - routing: 动态判断任务类型,传递给不同的下游 - parallel: 并行处理多个任务 - orchestrating: 类似于 parallel,但是下游的任务是动态生成的,而不是预定义的并行流水线 - evaluator-o

👍51

12 Feb 2026, 03:56 UTC≈2,910 views22 reactionsread 7 August 2026
Photo

最近都在重度使用 https://jules.google.com/,Google 出品的自动代码优化工具。我主要用来给我的项目找 BUG,效果非常显著。 免费用户可以并行 2 个任务(声称是 3 个,实际上只能运行 2 个😓)。 选择好项目、branch 后,点击下面预设的 Performance、Design、Security 生成 prompt。然后建议把执行计划改为 Start(单次执行),而不是默认的 Scheduled,防止占用有限的任务配额。(如果你是尊贵的付费用户那请忽略) 免费用户只能选用 gemini-3-flash。我个人的使用体验是,它找 BUG 的能力非常强,但是修复的能力比较弱。我目前的使用方法是,把 BUG REPORT 粘贴给其他更强力的模型来修复,最近每天都能找出很多陈年老 BUG,每天都在打击我身为老程序员的自尊😭。 不过换个思路,这也是一个去刷开源项目 PR 的机会。

👏175

13 Jan 2026, 18:59 UTC≈3,790 views24 reactionsread 7 August 2026
Photo

https://laisky.notion.site/Agent-Skills-Comprehensive-Guide-2e4ba4011a868055b2b0e2e128da1538?source=copy_link 学习了一下 Anthropic 的 Agent Skills 设计。通过在文件系统中定义一套文件结构,让 Agent 可以按需加载(Progressive Disclosure)相关的技能描述和脚本。 Skills 通过 metadata + instructions + scripts 定义一个领域技能。Agent 在启动时全量加载 metadatas,然后按需加载 instructions,而 instructions 会指导 agent 调用 scripts。scripts 的内容不会被加载,只有 output 会被加载进入 context。 通过一个三层的 lazy-loading 设计,相当于 R

👍24

24 Dec 2025, 16:42 UTC≈2,420 viewsread 7 August 2026

https://laisky.notion.site/msanft-CVE-2025-55182-Explanation-and-full-RCE-PoC-for-CVE-2025-55182-2c2ba4011a8681b89b60cf02827a6276?source=copy_link 之前提到的 next/react 严重服务端 RCE 漏洞 CVE-2025-55182 攻击者传递两个 chunk。chunk 0 作为攻击载体,chunk 1 作为正常 chunk,然后: 1. 利用自定义 then 将 chunk.prototype.then 挂载到 chunk 0,使得 chunk 0 变成 thenable,成为可被执行的 Promise 2. chunk 1 中,通过 $@0,将 chunk 0 作为 decodeReplyFromBusboy 的返回值,被 await 执行(thenable 对象都会被

Showing the 12 most recent of 19 posts we hold for @laiskynotes. View and reaction counts are the latest single reading for each post, not a live figure, and a recent post is still accumulating both. A view count marked was rounded by Telegram before we ever saw it — t.me prints views in full below 1,000 and to three significant figures above, so ≈1,200,000 means somewhere between 1,150,000 and 1,249,999. Unmarked counts are exact. Text is reproduced from the public post preview and truncated for length.

Citation-graph rank

Citation-graph rank — 845,762 of 1,189,255entries in the measured graph. A weighted position computed from the forward and mention edges below — republished posts weigh more than named mentions — and recomputed periodically, over the whole graph. Published only as this ordinal position, never as a score: a position is a fact, and a score printed beside one channel’s name would read as a verdict this register does not make. The two counts beneath stay separate for the same reason mentions are never summed with forwards anywhere else on this page — a named-by count costs nothing to manufacture. The top 100 by this measure, or how it is computed.

Forward network

Republished by

Channels on the register that have forwarded this channel's posts into their own feed.

Built only from forwarded posts we have actually read, on both sides. Coverage is early and deliberately incomplete: a missing link means we have not read the post that would prove it, never that the relationship does not exist. Counts are distinct forwarded posts observed, so they only ever go up as we read more.

Cite this entry

A live page changes as we take new readings, so a citation should name the measurement it is based on, not just the URL. The line below cites the subscriber count as measured 11 August 2026 — this entry's latest reading, not the date you are reading this.

“Laisky's Notes” (@laiskynotes), 3,291 subscribers as measured 11 August 2026. Telegram Register, tgregister.com/channel/laiskynotes.

Full measurement history, CC BY 4.0. Every reading this register holds for this entry, not just the latest one, as a dated, downloadable record: CSV · JSON. Free to use with attribution to tgregister.com. Each file carries its own generation timestamp, which is the figure to cite for exactly when the data was retrieved.