NotebookLM 80/20 Mastery: 10x Learning

🧭 THIS WEEK AT AI SECOND ACT

Howdy, two things this week -

• GPT-5 launched

• A glimpse of the future of learning with Google Notebook LM

It’s been a slow few weeks at AI Second Act with summer, and I had some time offline. I’ve been pondering pivoting the content and positioning, actually, so let’s see. Growth is a bit slow :(. [please share :)].

Next week will be super challenging with our eldest going to a US college, so let’s see if I can still publish. Of course, feedback is welcome :).

🧰 AI NEWS + LEARNING

Here’s a few things I found recently:

• GPT-5 launched, so please see the intro here and the longer video here

  • Analysis of GPT-5 impact and coding competitors such as Anthropic Claude Code —> Competitors to OpenAI will need to move quickly

• And a rant - I’m tired this week of all the hype around the ‘failed GPT-5 launch’ and users weren’t happy about GPT-4o [I think!] dissappearing for some time, blah blah blah -

  • Get over it! / ffs.

  • Wasn’t it not that long ago we didn’t even have AI? Isn’t GPT-5 brilliant? [Yes, it is!]. Stop the whining! Seriously, armchair bandits are ridiculous.

    • btw, subject for a complete edition, but if you don’t like AI and think it’s the doom of all of us, stop using it and stop posting on fracking Facebook about it!

  • Quick first thoughts on GPT-5

    • It’s evolutionary, not revolutionary - that’s OK!

    • It will become very mainstream as it just makes use of AI easier, automatically figuring out if it can give a quick answer or needs more reasoning/thinking

    • Coding is better, rivaling Claude, it seems

      • pricing way under-cutting rivals

    • Health is better - A huge gamechanger for humanity with AI + Health!

    • Hallucinations are way better - much fewer cases of AI getting it wrong, to very wrong, to slightly wrong

    • Summary: It’s great, people will love it.

🗺️ FEATURED INSIGHT

There are a few obvious main threads of AI use as we get into it here -

• Productivity / General Business

  • including research

• Coding / SW development

• Health and consumer information/co-pilot

• Education / Learning

Let’s talk about education and learning.

The ability of AI to help with learning topics is, frankly, astonishing. Both OpenAI ChatGPT and now Google Gemini have ‘learning modes’ where the model is instructed not to give away the answer so quickly, but help nurture the user to learn. Try it out!

Any topic you want to learn, you can learn, simply prompt it, state that you want to learn XYZ, and build a 3-week or 3-month curriculum, and the AI will create exactly that. Makes you wonder what schools, colleges/universities will be like in 5 to 10 years.

Then there is Google NotebookLM.

NotebookLM is Google's AI that reads your documents and has conversations about them.

Upload your PDFs, meeting notes, or project files. It reads everything, remembers everything, and can spot patterns across dozens of documents that you'd miss.

Critically, it helps you learn from that material. Even better, it then lets you create audio or video explaining the key topics.

TL/DR —> Simply upload material you want to learn and start prompting and learning from that material, including creating video or podcasts.

The audio is incredible - it creates a ‘fake’ podcast between two presenters discussing and summarising the material!

The 80/20 Setup (5 Minutes)

Step 1: Create focused notebooks

• One notebook per project/topic

• Don't dump everything into one massive notebook

• Examples: "Q4 Strategy," "Client Onboarding Process," "AI Learning"

Step 2: Upload your best sources first

• Stuff you want to learn

• Meeting notes from the last 3 months

• Key project documents

• Industry reports you've saved but never read

• Your previous analysis or presentations

Example Use Cases

1. Learn!

‘‘Summarize the material here, what are the key points, how can I learn this in sequence?’’.

2. The "Executive Brief" Prompt

"Create a 2-page executive summary of the key themes, decisions, and action items across all these documents. Include what's working, what's not, and what needs immediate attention."

This alone will save you hours of context-switching.

3. The "Gap Analysis"

"What questions are these documents raising but not answering? What information seems to be missing for complete decision-making?"

Perfect for identifying what you need to research next.

4. The "Different Perspective" Query

"If you were [CFO/new team member/external consultant], what would concern you most about what you're seeing in these documents?"

Three Workflows That Work

Weekly Reviews: Upload meeting notes + project updates. Ask: "What are the top 3 priorities emerging this week?"

Stakeholder Prep: Upload previous meeting notes. Ask: "What questions will [specific person] likely have based on their previous input?"

Knowledge Transfer: Upload process docs. Ask: "Create an onboarding guide for someone taking over this work."

Experiment with NotebookLM, see how it can help you summarise, synthesise, and learn from multiple material sources.

[There are also now some curated notebooks for you to view and start/understand the tool with.]

Future-proof your career with AI
— Brett

👉 Hit “Reply” and share your experience — I read every one!

_{Picture by Susan Q Yin on Unsplash.}