Preparing for my dissertation proposal defense, I decided to rethink my entire approach using NotebookLM and Google Docs.
I uploaded my dissertation notice into NotebookLM and used Deep Dive mode, where two AI agents analyze and discuss the material. After listening to those conversations multiple times, I realized I wasn’t satisfied with my draft.
Three issues became clear:
• One of my core tasks was simply too difficult, even for someone with a background in data analytics and Python
• The study lacked grounding in learning theory to support how participants would actually succeed
• Most importantly, the focus had drifted from how people learn to something closer to training data engineers
That last realization forced a reset.
I went back through the learning theories I’ve studied over the past 18 months at Central Michigan University and selected only those that truly support the study. I built a structured study guide in Google Docs, then used voice playback to repeatedly engage with the material, refining and removing anything that didn’t align.
What stood out to me is how effective “listen-and-do” learning can be.
For me, combining auditory input with immediate application is far more effective than passive reading, especially in technical domains. You’re not just absorbing information. You’re building pathways through repetition, feedback, and action.
Leaning into that approach isn’t just a preference. It’s a strategy for reducing friction and improving learning outcomes.