AI Era: My Simple Research Workflow
You find a huge list of Markdown files from some awesome repo.
There are dozens of links inside those Markdown files. You’ve timeboxed a research topic.
You have two options:
- Slow down time so you can carefully read all the links.
- Delegate the task to NotebookLM.
But there’s one problem:
- You first need to extract the URL list and possibly exclude some (like GitHub repos).
- Or, you might only be interested in YouTube videos—and that’s fine.
Here’s a CyberChef recipe (which I learned from my wife) to solve this problem:
Extract_URLs(false,false,false)
Filter('Line feed','github.com',true)
Filter('Line feed','youtube.com',true)
Sort('Line feed',false,'Alphabetical (case sensitive)')
Unique('Line feed',false)
The recipe:
- Extracts all links from Markdown.
- Excludes GitHub and YouTube links.
- Sorts the links alphabetically.
- Removes duplicates.
Now you’re ready to give these links to NotebookLM to boost your research process.
Homework: Since some links may have disappeared, develop a simple shell script to exclude broken links and keep your notebook clean.
At the time of writing, there’s no built-in way to remove broken links from a NotebookLM notebook.
Direct CyberCHEF links: