John Lin

Lead SRE at LINE Taiwan. Crafting observability stack. Prev: Infra Eng @tencentcloud. I play tennis. 🎾

[antirez @antirez]

This is accurate, and you can make it more extreme in most cases today as: *build* your narrow-scope-dependencies, unless it is really a complicated stuff. In the second case, internalize it and specialize it.

[Barret李靖 @Barret_China]

大多数程序员会慢慢出现两类症状。

一是不想写代码。因为写的没有 AI 快，也写的没有 AI 好，编程将沦为普通人的玩具（架构和工程目前还不是），程序员容易找不到成就感。

二是不想指挥 AI 写代码。由于一的存在，加上指挥一个没有感情的机器，去做自己不认可的事情，会比自己亲自去做更难受。因为躬身事中，疲于处理一个又一个的问题，会忘记自己的存在；而看着 AI 在自己不认可的方向，持续接近目标，会无数次巩固自己的存在。“我是谁？我在干啥？”

程序员应该做的，就是摆脱自己会写代码的标签，慢慢朝着用户价值层做迁移，去做自己认可的、喜欢的事情，或者用自己认可的、喜欢的方式做事情。😃

[John Lin @johnlin__]

@plateau 有閒可以來玩 https://t.co/hlZOgQPbE2

[Guido van Rossum @gvanrossum]

Everybody is adding a feature where you can manage your agents from your phone. Don't use it. You'll just get even more addicted, and will burn out even quicker.

[Theo - t3.gg @theo]

Lmao. Kind of insane that Azure wasn't monitoring this internally but I'm thankful they listened and fixed it finally.

[John Lin @johnlin__]

有夠慢，我也這麼覺得

[Theo - t3.gg @theo]

Azure customers hosting OpenAI models should now be seeing a 10x improvement in latency and throughput

You're welcome :)

[John Lin @johnlin__]

@badlogicgames @FireworksAI_HQ Is this related to the PR 3678?
https://t.co/AKwOAumsUp

[John Lin @johnlin__]

https://t.co/trbA3LfWeS

[John Lin @johnlin__]

@dhh What plugin you’re using?

[John Lin @johnlin__]

@plateau 買max後，我覺得我是所有開發環節的瓶頸了

[John Lin @johnlin__]

買max了 token 用不完！

[John Lin @johnlin__]

@plateau Boris 7.7B!

[John Lin @johnlin__]

It seems that many large OSS projects have AI contribution policies, and you can see that their overall direction is consistent.

CMake
- Contributors are welcome to use AI...
- AI tools and services are not considered authors. They should not be listed as co-authors

Kubernetes Community
- Using AI tools to help write your PR is acceptable
- Listing AI tooling as a co-author, co-signing commits using an AI tool, or using the assisted-by, co-developed or similar commit trailer is not allowed.

https://t.co/bZoiJyPYf7

https://t.co/Bv6Seh6FJE

[John Lin @johnlin__]

@thdxr @steipete How to collect metrics from OpenCode ?

[John Lin @johnlin__]

@plateau 這風格不錯 太讚了

[John Lin @johnlin__]

我不禁在想，難道這就是AI帶來的未來嗎？一個專案看起來上線，但其實不能，可能你的老闆看到他功能很棒覺得很讚要趕快推動到市場，但其實維運是一坨屎。那最後就是你能力不夠，人家應用寫那麼好demo都不是問題，為什麼不能生產上線佈署，為什麼rate limit修不好，為什麼2FA不可用？

[John Lin @johnlin__]

剛跟幾位前同事在研究一個有趣的專案，旅程規劃的服務，可以100%自建，沒有訂閱費，你擁有你的資料，非常適合與親友協作規劃旅程及回顧

[John Lin @johnlin__]

隨著研究投入的時間變長，開始發現一些問題，因為專案重度的AI 生成（vibe coding) 一些奇怪的問題開始發生，例如莫名的rate limit 機制、登入綁定不可用的2FA 等問題，開始覺得是不是踩了個大雷，整體來說專案不是不可用而是太多小問題，而這些問題可能都需要透過源碼修改才能完成。

[Andrej Karpathy @karpathy]

Wow, this tweet went very viral!

I wanted share a possibly slightly improved version of the tweet in an "idea file". The idea of the idea file is that in this era of LLM agents, there is less of a point/need of sharing the specific code/app, you just share the idea, then the other person's agent customizes & builds it for your specific needs.

So here's the idea in a gist format: https://t.co/NlAfEJjtJV

You can give this to your agent and it can build you your own LLM wiki and guide you on how to use it etc. It's intentionally kept a little bit abstract/vague because there are so many directions to take this in. And ofc, people can adjust the idea or contribute their own in the Discussion which is cool.