Next steps and continued learning with Claude Code

Link copied

Attach video time

You now have a solid understanding of all the core fundamentals of using Claude Code. I'd strongly encourage you to move forward with AI-assisted development -- things are just getting started and you don't want to miss this wave.

I've built some complex things with Claude Code, but the most ambitious one is this exact course platform you're using right now. Nearly 100% of the code was written by Claude Code! But it took my own preexisting skills, judgment, and taste to shape it into what it is today.

Try using Claude Code for any new greenfield project, as it really does work fantastic in that area. But the AI has gotten good enough that you can also start using it within existing projects.

If you're hesitant, start slowly. Have it assist you writing a function here and there. As long as you review what it writes, understand the code, and push back on anything that's amiss, you should have zero issues having it write production-ready code.

What's next?

We've just scratched the surface. There's a lot more to explore:

• Skills -- specialized collections of instructions and resources that can dramatically improve the quality of Claude Code's output for specific tasks

• Hooks -- tap into before/after actions within Claude Code, similar to the Git Hooks process we covered earlier in this course

• MCP -- extend Claude Code's capabilities with external systems, or have it write an MCP server for your own API

• Claude Agent SDK -- build custom autonomous AI agents for your own applications or CI/CD pipelines with the same power as Claude Code

Since Claude Code is really an "everything" app, you don't even need to use it just for code. You can ask it to carry out tasks on your computer (freeing disk space, creating calendar invite files, etc.) or even make it into your personal agent.

If you'd like to keep going with me, I put together a follow-up course called Build Your Own AI Harness. It takes all the scattered prompts, skills, and configuration you've been collecting along the way and turns them into a single organized harness -- the kind of setup that actually holds up when you point it at a real production codebase.

Now start experimenting, read the docs when you get stuck, and if you have questions, drop them in the comments on the related lesson.

Good luck!