Hacker News Reader

Comment by AlexCoventry | Hacker News Reader

Isn't it a bad time to invest in this? If an AI can help you learn the material, it can probably replace you pretty quickly.

View on HN

This book mostly teaches things that I don't think AI can...

But also -- seemingly-silly things can help build foundation for more human learning.

Multiplication tables seemed to have an important role for centuries -- they don't turn you into Einstein, but I bet he knew what 2x5 equaled pretty early in life.

Here's just one Amazon review on this:

"This is the best overall resource that I've seen for software engineering job searches and coding interviews. There's a lot of free online resources out there so I was on the fence about buying this, but this is the most high-quality and practically useful collection of tips, tricks, advice, and explanation of relevant CS data structures & algorithms concepts that I've seen. There's also a lot of practical advice on how to approach the job search. This, plus leetcode, would be my most highly recommended resources for coding interviews. I think the content of this book would be relevant and useful from the new grad level through the staff SWE level. It's also a totally different book than the original CTCI (which I also have), not just another iteration of the same thing. I never write reviews for anything, but I felt like I had to for this because I was surprised by how good this is."

I think that a lot of the benefit might be having a defined plan for approaching a job search. E.g. what specific topics do you cover and what topics don't you cover? For a given topic, how in depth do you go and what aspects do you cover? Etc.

I could see using AI to research a given topic or walk you through a given problem. However, at this point I don't think AI is up to packaging up a well rounded textbook level study on a topic.

AI can absolutely help you. A lot of the benefits of the book come from the reasoning, though, which AI cannot do. For instance, you might try a question, and the optimal answer uses a heap, but how did we intuit that? What repeatable reasoning steps could we walk through to do that for a different problem? That's what 500 pages in this book is about. XD

You are incorrect. ChatGPT will absolutely tell you why the optimal answer would use a heap. I've found that ChatGPT is much, much better than any book resource because it amalgamates answers from everywhere into a coherent answer and you can keep asking questions about it to gain deeper understanding, as opposed to a book where if you don't understand that's it.

Hey friend, totally okay if we don’t see eye to eye on this. No big deal. But just to share my perspective—if you ask ChatGPT to break down a genuinely hard, new problem (not something that’s been around forever with tons of tutorials and blog posts), the explanations tend to stay pretty surface-level. For example, you might get something like, “We need to use DFS because we need to search the graph.” It doesn’t really get into the deeper reasoning behind why that’s the right approach or what led to this decision when others were possible.

There’s actually some interesting data on this here: How hard is it to cheat with ChatGPT in technical interviews?

Even AI experts point out that parroting tutorials isn’t real reasoning, and that’s still a tough spot for AI.

All that said, even if AI improves, I still think this book offers a lot of value. It’s packed with new templates and practical strategies to help you get unstuck during interviews—stuff based on data from over 100,000 mock interviews. No pressure if it’s not your thing, but if you’re curious, you can check out some of the technical (and non-technical) chapters at the link below. They cover approaches you definitely won’t see ChatGPT come up with. :)

https://bctci.co/free-chapters

Reply on news.ycombinator.com