I’m the first to admit I’m a Luddite. It’s an odd thing for a tech journalist to say, but it’s the truth.
I rely heavily on pen and paper for my notes and research, and I only recently discovered the magic of E Ink tablets for my scribbles.
I’m not opposed to changing things, I’m just set in my ways and think if it’s not broke, I don’t need to fix it.
However, I’m making an early New Year’s resolution to expand my horizons and try new ways of working. That led me to combine NotebookLM and ChatGPT.
I’ve used both separately before, but I wanted to see if there were ways to use them together to get even better results. Each has its own advantages, and they play off each other beautifully.
If you’re stuck in a rut, here’s how running NotebookLM and ChatGPT side-by-side can help you work more efficiently.
NotebookLM is excellent for research and limiting noise
I don’t need the entire internet
NotebookLM’s strength is its ability to focus.
ChatGPT isn’t ideal for research because it’s all over the place. It’ll pull in sources I don’t want and fill my screen with information I don’t need.
NotebookLM lets me compartmentalize my work.
I recently wrote an article about how stadiums affect your 5G signal. During my research, I had to read several technical white papers on various subjects. It took quite a bit of effort to pare down what I needed and bring it together into an article that made sense.
NotebookLM would’ve simplified that process. I could’ve given it the three PDF files I was using, and then gleaned the information I needed.
I wouldn’t have to worry about stuff sneaking in from other sites or reports because I could have confidence that it was only drawing from what I wanted.
I wanted to see if NotebookLM could have a cumulative effect on my work, so I started entering my own articles as sources, building up a notebook of previous reviews and features I’ve written.
I can’t remember what I did last week, let alone a year ago. Nothing erodes your short-term memory faster than having a toddler running around.
Building the knowledge base of my own works with NotebookLM makes following up easier.
Next time I’m reviewing a new device, I can reference all my work on the previous generation to see if I had any unanswered questions or if significant issues were addressed with the new model.
Instead of sifting through 10 or 20 articles, NotebookLM lays it all out for me, so I’m ready to get to work testing whatever new device is in front of me.
ChatGPT plays a role, but I keep it in check
It makes a wonderful sounding board
If NotebookLM is the research assistant of the operation, ChatGPT is the sounding board. I don’t rely on it for anything mission-critical, and I’d quadruple-check anything it outputs.
However, it’s still a useful tool for spitballing — just throwing out ideas and arguments to see if they make sense. I don’t always get gold when running ideas past ChatGPT, but you’d be surprised when a piece of information pops up that you didn’t think of.
It’s helpful for catching obvious errors or when something doesn’t make sense. It’s not ideal for scrubbing through your work with a fine-tooth comb, but at least you can know you’re not way off course.
NotebookLM helps me narrow down the information I’m looking at, while ChatGPT keeps me on the course I set. It’s important to have both.
I may get more use from the NotebookLM side of the equation, but ChatGPT is another avenue to share ideas, even if it’s AI talking back to you.
ChatGPT is also helpful at the beginning of the process. You can ask it for ideas on where to start researching a broad topic, especially one you’re unfamiliar with. Always make sure the sources are authoritative, but it’ll at least give you a place to start.
I don’t love using ChatGPT to actually structure my writing, and you have to be careful what bleeds over. But for general information and making sure an article makes sense, it has found its way into my workflow.
Nothing replaces the human touch
AI can’t formulate your ideas
I may have embraced some new additions to my workflow, but that doesn’t mean I’m still not apprehensive.
AI, and ChatGPT in particular, should never be used blindly for anything mission-critical. There are too many mistakes, and I even double-check what I get from NotebookLM.
It’s best to use AI to summarize what you already know or have some understanding of, so you can pull out what’s useful and discard the nonsense.
I’ve been waiting for AI to get better at cutting down on hallucinations, but I think I’ll be waiting a bit longer.
Until then, I’m keeping AI restricted to what I outlined — using it as a tool rather than a replacement for research and thought.
Don’t be afraid to try something new
I’m always the last person to watch a new show or try out a new way of doing something. I don’t consider it a negative, as being new doesn’t automatically translate to better or improved.
However, when it comes to using AI to research and dig through massive amounts of information, I’m glad I have some new tools at my disposal.
