𝚅𝚒𝚝𝚊𝚕𝚢𝚁

Learning mathematics

[Thariq @trq212]

https://t.co/R6exTuF7P8

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

@thsottiaux Please unban my codex account. I believe I didn't violate the usage policy, but you suddenly banned me today. The account has conversation history and important data, and I can't access them now, which is really disappointing and unacceptable. I have appealed, but it doesn't help

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

@0xMulight 不如gh

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

@thsottiaux Thanks for doing this. Just to share some feedback: since the last reset was 7 days ago, many weekly Codex limits were about to reset automatically in a few hours anyway.

A reset 2–3 days into the cycle would feel much more impactful and would probably be appreciated even more.

[Tibo @thsottiaux]

I don't know what they are doing over there, but Codex will continue to be available both in the FREE and PLUS ($20) plans. We have the compute and efficient models to support it. For important changes, we will engage with the community well ahead of making them.

Transparency and trust are two principles we will not break, even if it means momentarily earning less. A reminder that you vote with your subscription for the values you want to see in this world.

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

codex又在逆天而行了，要把自定义slash command干掉，全转为skill，问题是这两个机制并不相同，不能互相替代啊

[Frank @frank_uid]

刚刚翻了一下 @philipkiely 的Inference Engineering，从模型到推理引擎、GPU和系统设计都有涉及到，大部份是llm相关的，少部份是diffusion，很适合扫盲

https://t.co/uXKqzvlWjI

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

@btc66c @grok 新年快乐，祝老板永远不死喵

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

蛙趣，用 https://t.co/7xdDA3Po1m 改键不成功，gpt 5.2帮我改好了

[Elliot Arledge @elliotarledge]

my friend @neuralkian just dropped a pipeline parallelism course for FREE! this is exactly what frontier labs would hire you to work on at scale in order to speed up training and inference on large models. you'll start with a simple example of overlapping computation on a small MLP, and work up from there!

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

@laofeiyyds 那还是上十万吧

[Ilya Sergey @ilyasergey]

My research lab is launching a new blog, where we will share thoughts and tutorials on formal methods, mechanised proofs, PL, and more.

https://t.co/DdaoMpHN0O

First post: verifying imperative programs in Lean 4 with Velvet, using symbolic automation and AI-assisted proving.

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

codex 已支持 subagents

[泊舟 @bozhou_ai]

我也分享一下我自己用的Skills

1.Anthropic官方Skills，这里面我只安装了 skill-creator
https://t.co/nWpEtHBcK6

2.UI UX Pro Max，用来写前端的UI，非常好用
https://t.co/bsmrS0WM5T

3.seo-review，我用来做AI网站SEO审查的
https://t.co/JWtSAjO1fc

4.content-creator 这是根据关键词，然后进行创建博客文章的，也是用于SEO
https://t.co/JtEyQbrLYk

5.ai-image-generator 这是自己写的，调用api然后生成图片，接着压缩为webp格式，最后上传到R2，非常实用，用于博客还有示例图创建

6.skill-prompt-generator 黄总的提示词生成，主要搭配生成图片的，也是强烈推荐
https://t.co/IRFGgx4BWG

7.Planning-with-files 参考Manus的Agent方法写的Skill。很适合多步骤任务。 https://t.co/FiCGOO9u6a

Superpowers 这个很好，但是属实有点太磨叽，然后我想法模糊一般都是直接在网页里面聊，所以移除这个

没有用NotebookLM是因为习惯了在网页里面使用，现在还没有想到用Skills能干点啥

最后我强烈建议各位，一定要多多搭建自己的Skills，从日常的一些工作流开始。

比如说我们有一些流程，我们可以先用claude code来实现一次，当实现完了，我们让cc总结成一个Skill，然后基于第一性原理进行打分，迭代调优

如果觉得麻烦，可以直接先用官方的skill-creator进行沟通，来创建一个自己的Skill，要知道，自己沉淀Skill才是最好的

[Yanhua @yanhua1010]

推荐一些我目前安装的Claude Skills:

1. Anthropic官方Skills
https://t.co/Xfgu1Thw69

2. Superpowers
1.6万Star的Skill精选，从脑暴、写需求文档、开发、测试全包含，口碑相当好。
https://t.co/D2ToCBGnbN

3. Planning-with-files
参考Manus的Agent方法写的Skill。很适合多步骤任务，用这个Skill指导其他Skill工作也挺好。
https://t.co/EKBZQmr03F

4. X-article-publisher-skill
王树义老师写的X文章发布Skill， 这个值得研究下，后续看看能不能实现其他平台的自动化
https://t.co/rjDPGC4gtg

5. NotebookLM skill
自动上传PDF、Youtube链接到NotebookLM，很适合NotebookLM内容的自动化处理
https://t.co/CX0vPril27

还有一些是自己平时总结的内容创作skills，后续继续分享具体使用场景案例！

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

@waveking1314 @grok @grok choose me and I will tell you an interesting story

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

非常神奇，田春的冰河年鉴依然在更新 https://t.co/sz6U4cxaLI

[𝚅𝚒𝚝𝚊𝚕𝚢𝚁 @der_emacs]

@tradermige 报名，1w还不够把我砸成废人

[恋 @hgds0]

来抽奖啦，今年第一条抽奖，给大家抽一条西太后项链，感觉饰品的实用性和颜值都挺高而且男生女生都可以戴～奖品可折现（1415元），按项链当日售价发红包！大家关注我 并评论 转发 此条推文就行啦，下个月（1月31日）开，拒绝黑幕我们到时候随机数抽奖，希望大家喜欢💕