Ian Hsiao’s blog

AmpCode is cool

There are infinite work to be done and undone

近期隨想

“The world is no longer intelligence-constrained; it’s passion-constrained.”

亞洲問題

大不了多「失敗」幾次就有更多 track record 了 XD

Enabling Slack MCP in Codex and Claude Code (workaround)

Codex CLI and Claude Code both refuse to log into Slack's hosted MCP server because they only know how to do Dynamic Client Registration, and Slack doesn't support it. Here's the bearer-token workaround that actually works today, plus the GitHub issues to watch.

San Francisco

What a great city!

Revisiting Engelbart in 2026

A framework to understand the far-reaching effects of AI.

隨波逐流 Currents

你總是擁有選擇的權利

Closing the Agentic Coding Loop with Just Enough Human Feedback

I think I figured out how to steer the agents minimally but still have the desired result--theoretically; and I think it's very different from what people are talking about today, so it's worth noting.

Software as toys

Software as toys? Human as boot loaders? Will the next generation of intelligence kill us? How is coding changing again? Some thoughts.

《The Dream Machine》閱讀小心得

人類打造工具、工具反過來增強人類，這種往復的循環讓人振奮，也讓我深受感動。

Notes on “The Art of Doing Science and Engineering”

Luck favors the prepared mind.

The Art of Doing Science and Engineering 閱讀筆記

Luck favors the prepared mind.

Learn >> Build, Source

Learnings from an Ex-Amazon Engineer

There are often moments when I can’t tell if LLM over-engineered, and that’s when I know I have limited knowledge on the subject. I rely a lot more on seniors and talking to models to learn here.

The Three Most Important Things in My Life Right Now: Build, Learn, and Source

I just wanted to say that, as someone who just graduated and doesn’t have a golden handcuff, I am very excited about the future!

My Current Mindsets / Value

My understanding of risk and uncertainty before graduating college

Take the largest bet that you are ok with failing.

大學畢業前夕對風險與不確定性的理解

選擇那些你能承受失敗的最大賭注。

過去兩個月打造產品所學

相信自己，然後穩步前進

如何選擇大學

這篇文章是我為回答 17 歲時自己的疑問所寫，當時我在思考要讀哪一所大學。不過，我認為如果把「大學」換成其他你會待上幾年的組織或地方，其中的一些思考或問題依然適用。

在 startup 工作 6 個月

作為一個對做酷東西和做產品感興趣的人而言，在一家優秀的新創公司實習是一件很幸運的事。一方面是因為能與很強的人共事，可以學到他們的心法以及做事方法論，另一方面是因為新創公司速度快、人少的特性，讓我有機會接觸到一些我現在能力範圍外的任務 (punch above my weight)。這篇文章主要寫的是這幾個月對專注的新理解，以及在 AI 時代做技術工作的一感想。

與世界碰撞

每一次出遠門都很值得

和自己說說話

你知道嗎？你也可以 prompt engineer 自己哦

How to Pickpocket and How NOT to be Pickpocketed

Jun 3, 2024

Takeaways from How to Do Great Work

Doing great work is a depth-first search whose root node is the desire to. So if you don’t succeed, either try again, or backtrack and then try again. Corollary, never abandon the root node.

在 startup 工作 1 個月

最重要的是明確最重要的是什麼

不鼓勵、不反對

我成長背景的一隅。

休學的一些反思

這幾個月閱讀了很多人成長、自我實現與反思的文章，也有機會接觸到一些很尊敬且想成為的人。他們給我了帶來很大的啟發，也給了我一些往前的、探索的動力。

Hi, this is my new blog :D

敢於嘗試，機會是主動拿到的。

我從北大休學的原因

Ask Me Anything! (active)

A Dating Service with Text-embedding Model?

Social Vegan - a dating app proof of concept using OpenAI's text-embedding model and Pinecone vector database to match people based on their relationship values.

Markdown File Query

Query any markdown file (Notion/Obsidian) using vector embedding, Pinecone vector database and GPT - a semantic QA system for your personal knowledge base.

神经整合 - 如何设计生物神经元网络

探讨如何利用生物神经元实现逻辑门运算，包括与门、或门、非门和异或门的设计方法，以及人工神经网络的基本原理。