# AI Dungeon Run | Andrew Ng's AI For Everyone – Chapter 2: Building AI Projects

https://ijuhsu.com/en/posts/ai-for-everyone-ch2 · 2025-08-21
Category: ai · Tags: ai, Andrew Ng, AI For Everyone, 讀書心得

---

Chapter 2 gets into the mechanics of AI projects, with plenty of concrete examples.

Machine Learning, for instance, requires collecting data, training a model, then testing it in the real world. A Data Scientist collects data, analyzes it, and turns hypotheses into actions. Either way: **data is everything.**

When starting a project, it's recommended to pair an AI Expert with a Domain Expert, because:

*   **AI Expert:** Knows what AI can and can't do technically
*   **Domain Expert:** Knows what actually matters to the business

But the key question throughout the process is: how can AI automate **tasks**, not **jobs**?

This hit home for me — I'm currently building a Research AI tool at work. AI can't replace a Researcher, but it absolutely can help a Researcher finish their daily tasks faster.

So if you find AI can't complete a Researcher's entire job and conclude it's "dumb" — that's a misplaced expectation. We should expect task-based automation, not job-based automation.

* * *

# **Thanks for reading :D**

If you enjoyed this post, **feel free to click the coffee button in the lower right to support us and give Lottery a can 🐾**