I passed the Google Generative AI Leader certification yesterday. Not my usual territory, but the way I studied for it is worth writing down.
How I normally do this
Gather the official docs, work through practice questions, take notes, do a lab for the product-related bits, repeat until it sticks. It works. It’s also slow, and a lot of what I read ends up flat in my head, in one eye and out the other by the following week.
This time I dumped the study materials and Google’s official documentation into NotebookLM and used it more like a study partner than a reference doc. I wasn’t expecting much from it, to be honest.
The part that actually helped
Not the chat, which was fine but unremarkable. It was the audio overviews. NotebookLM turns your uploaded material into a podcast-style conversation between two AI voices talking it through. I was skeptical the first time I played one, and then I listened to the whole thing on my commute without really trying to, which told me something.
Hearing a concept explained out loud, in a back-and-forth, landed differently than reading the same paragraph twice. A few topics I’d skimmed past on the page actually clicked once I heard them argued over.
The Q&A mode earned its keep too. I’d get it to quiz me, push back when an answer felt off, and ask it to walk through why a specific choice was right or wrong. Closer to how I’d study with a colleague than how I’d study alone with a PDF.
Where it falls short
It only knows what you feed it. A couple of times I noticed the summary was vaguer than what the source material actually said, and it didn’t flag that it was simplifying. If your source set has gaps, the tool won’t tell you, it’ll just answer from what it has.
It’s also not a substitute for reading the primary material at least once yourself. I used it to reinforce and stress-test what I’d already read, not to skip reading it.
Why bother, given this isn’t my field
Network security and reliability is where I actually work. I picked up this certification because generative AI is increasingly showing up in the infrastructure and security tooling I deal with day to day, and I’d rather have a working understanding of it than nod along in meetings.
The bigger surprise was how much I ended up thinking about the studying itself, not just the material. Turning a stack of documentation into something you can listen to and argue with is a genuinely different way to learn. I’ll use it again for the next one.