A few months ago I was exploring Claude Code to get a feel for how it handled some of my projects. But I was also feeling a little limited…like I suspected there was so much more I could do, I just didn’t know what.

Then I took the short deepLearning.ai course on it and it made my experience SO much better (and I immediately switched away from Cursor).

I’ve since recommended it to everyone who uses AI for coding (or related tasks!) and is a little lost. It’s just about an hour long, free, and (as usual) Andrew Ng delivers. He doesn’t teach the course himself, but I credit how useful it is to his trademark clarity and high standards - he’s easily in my top 3 AI educators.

Here are 5 mini-takeaways

  1. Learn Claude’s shortcuts like /init, /clear, and /git, that help you manage memory, context, and version control without breaking flow. These are the basics.

  2. Start with Plan Mode (Shift + Tab twice). It lets Claude sketch out what it’s about to do before touching your code.

  3. Learn to hook up Claude Code (in VS Code) to the Playwright MCP server, so it can take screenshots of UI updates and keep tweaking until it’s right. (And BTW the course explains things like Playwright and why it’s powerful beautifully.)

  4. Use worktrees to have Claude Code implement things simultaneously. I don’t use this much because I like to review work closely and make small tweaks, but it’s good to know it exists!

  5. The course also shares a bunch of tips and quirks - like giving Claude better context to boost token efficiency, or asking it to “think a lot” when something’s tricky. A great reminder of how powerful (and limited) LLMs still are.

It’s one of those rare “hour in, massive amount of learnings out” moments. Let me know if you take the course, I’d love to hear what you think!