Navigating Change: Staring into the Glare and Finding a Practical Use for AI
FM (Friday Morning) Reflection #13
When the bright light of change is shining directly in your eyes, it’s hard to keep the full picture in view. We may turn away, stare into the glare, or look down to focus on what's immediately in front of us.
If you’re driving and can't see because you're headed toward the sunrise or sunset—what do you do? Even if we adjust the sun visor to block most of the light, we likely will slow down, grip the wheel tightly, turn down the music, and get hyper-focused.
The same thing happens in a thunderstorm when the windshield is inundated with rain and the wipers can’t keep up. What do you do? The same—plus maybe turn on your hazard lights or find shelter under an overpass (though that’s not a good idea during a tornado).
It’s interesting that our behaviors are largely the same in these very different situations. Yet many of us will press on without clarity, moving forward on whatever path we've set.
If you were to stop for a moment—maybe squint a little or briefly pause for the heaviest rain to pass—you can often start to make out the outlines of what’s there and begin to better navigate the way forward.
Moving Past the Hype of Generative AI
In several of my recent reflections, I’ve brought Generative AI into the narrative. We’re at a point where there is a lot of excitement and debate, but finding clear examples of real value can be challenging so far. If you’re just coming to the tech (as many still are), you’re probably either on the upslope to super high expectations or have started the descent from that first peak into disillusionment.
When I look at media coverage of Generative AI, there’s a similar split. Experts are divided—some call it transformative, others are more skeptical, and few think it will be the end of civilization. My sense is that this often depends on your point of view—whether you’re a potential beneficiary or are concerned about an adverse impact on your livelihood.
Today, I’d like to move past the hype and talk about real value. To illustrate this, let’s look at NotebookLM, a Google Labs product that’s received a lot of attention lately. I don’t usually get into talking about specific products here, but I think this is one where it makes sense to go into depth.
NotebookLM: The Value in Simplifying Complexity
NotebookLM is still in its early stages. Billed as a "Research Assistant," it's a utility that can analyze documents, websites, videos, audio, and other sources. It ingests these materials to create a comprehensive summary, FAQs, timeline, and briefing document.
NotebookLM also does two other very useful things—first, it makes these documents available so that you can "chat" with them, just like you'd ask a chatbot a question, with answers based on the data you uploaded. Super helpful on its own.
And the second thing... it stopped me in my tracks when I first heard it. NotebookLM will also create a conversational "podcast" of the content, a spoken word Deep Dive, complete with banter between two AI-generated hosts.
That's right. The NotebookLM team has modeled the dynamics of human conversation between two people—in a format designed to facilitate knowledge transfer—and built a generative engine to make all that content enjoyable to listen to.
Much more than just an audio summarizer, the Deep Dive feature acts as a Generative AI translator, providing information in a way we've all wanted at some point: “Can you just give it to me in a way that I can understand?”
We’ve relied on the oral transfer of knowledge for millennia. Our brains process spoken words intuitively. We respond to tone, cadence, and back-and-forth exchanges that reinforce understanding—especially when we're brave enough to “ask the question that everyone else was thinking.” NotebookLM’s engine finds adjacent topics, metaphors, and supplemental information to illuminate the core subjects at hand, and in that way, also makes it relatable. You still need to review the content, and often I'll generate two or three versions to get one that has the mix of content and tone that I'm looking for in the output.
Experimenting with NotebookLM's Deep Dives
I’ve been experimenting with NotebookLM to see how I might use it. I've loaded my resume, several Friday Morning Reflections, and the HR benefits information from my employer to create many different Deep Dives. In each case, NotebookLM created useful and often engaging narratives with and around that content.
I tried to make it sad by loading the Pennsylvania Unemployment Compensation Handbook, but the narratives remained helpful, despite the complexity of the content. This highlights the effectiveness of the prompts engineered by the creators of NotebookLM—prompts we don't see behind the scenes. There are no strict controls but there is a new feature to inform where you want the focus of the narrative to be—“talk about this, don’t talk about that.”
The point is: the Deep Dives take a lot of information and make it digestible, along with the rest of NotebookLM's features to creates FAQs, provide summaries, and allow you to query the data with the built in chat-powered search.
For fun, I loaded several of my Friday Morning Reflections into NotebookLM, and I’ve included an example of the output.
NotebookLM-Produced Deep Dive:
On Eric Duell and Do Good by Doing Better
Note: I provided minor guidance to NotebookLM to generate this—asking the model to introduce me, my philosophy, and then to focus on an overview of FM (Friday Morning) Reflections and the podcast content posted on Do Good by Doing Better.
Why This Matters
Think about all the hard-to-digest, complicated topics you deal with in life—unemployment compensation, insurance coverage, the fine print on a bank loan. How many times have you just given in and signed the paperwork, only to find something in the fine print that made you regret the decision later? Imagine having someone (or in this case, something) to translate that into concepts that are clear and relevant to you. What if you could also "ask questions" of your data, find detailed answers in the original documents, dive deeper when needed, or generate materials to further study the topic? That's what NotebookLM does.
In the grand scheme of things, this isn’t going to fix climate change or bring about world peace. It’s a seemingly small change. But the benefits could be huge—a little bit of help, multiplied hundreds of thousands of times over, has the potential for a big impact.
The techniques to store documents and make them "chat-able" with an LLM front end have been around for a while. But what I like about NotebookLM is that it makes something technically complicated so simple. It shows clear utility value. It’s easy to see how this technology has the potential to empower you and me.
Hopefully, we’ll see more examples like this in the consumer market soon.
AI's Real Impact: Small Changes, Many Use Cases, Big Results
There are plenty of ads around right now about how AI can solve every problem, but honestly, many of those problems aren’t that pressing. I can get through the holidays without an AI-generated greeting card, and I don't have any ex-friends that I need to airbrush out of my selfies right now. But picking the right insurance policy or benefits for my family during open enrollment? That could mean thousands of dollars saved or lost.
This is an example of how, despite the bright lights of AI-this, AI-that, hot-new-exciting-different in the news all... the... time, the outlines of what’s to come have started to emerge. And this is one development that I’m encouraged about.
Let’s hear your take.
Have you used NotebookLM? What other “small” benefits of Generative AI have you discovered with the potential to make a big impact? What have I not considered here?
Let’s talk about it! Please share your thoughts and reactions in the comments.
And have a great weekend!
Photo Bonus
Took this earlier in the week from the Little Island Park on the Hudson River, adjacent to Pier 57 in Manhattan. I used a new experimental film, Green 600, just released by Polaroid. Despite the glare, with the right framing and exposure, you can see the outlines of the buildings on the other bank of the river.