NotebookLM In 30 Minutes

I learned how to use Notebook LM for

you. So, here's the Cliffnotes version

to save you the hours and hours that

I've spent deep diving into this

extremely powerful tool for both working

and learning. It is honestly one of my

favorite tools, and it's crazy how much

value you get from the free version. So,

in this video, I'm going to show you the

core features of Notebook Outlook,

including things like distilling and

summarizing information and how to make

it even more powerful in combination

with other tools to even do things like

automating workflows and building AI

products. As per usual, it is not enough

for you just to listen to me talk about

stuff. So throughout this video, I will

have little assessments which if you can

answer these questions, then

congratulations. You would be educated

on using Notebook LM. Now, without

further ado, let's go. A portion of this

video is sponsored by Augen Code. Let's

start off with first answering the

question, what is Notebook LM for? Well,

if you go to the landing page for

Notebook LM, there in big massive

letters, it says understand anything.

The core kind of concept of notebook LM

is that in our information age, the

problem is not the lack of knowledge.

It's our ability to understand and make

sense of it. And that is what notebook

LM is here to do. It allows you to take

a lot of different types of information

from a lot of different sources,

synthesize them together, and help you

understand them deeply. I'll go through

each feature of Notebook LM in detail in

just a little bit, but first I want to

explain the overarching workflow for

Notebook Lab. When you first go into

Notebook LM and you click create a new

notebook, step number one is to upload

your sources. You can drag and drop all

sorts of different files and also link

from your Google Drive, different links

including websites and YouTube links and

copy text as well. You can also click to

discover sources tab and be able to

search other external sources too. You

can then chat with your sources in this

chat window asking you to do things like

summarizing or analyzing. But that's not

all. What really makes Notebook LM stand

out are these specialized sources here

under studio to help you dive deeper

into the content and to understand it

better, including things like generating

an audio overview, generating an AI

podcast.

>> Today, we're diving headirst into

something

>> which is what first made Notebook LM go

viral, and also do video overviews, mind

maps, and all sorts of reports. I'll be

showing you examples of these all in a

bit, and it is really, really cool. As

its name suggests, Notebook LM is also a

note-taking application. So, it allows

you to add a note and include your own

notes. And finally, there are sharing

features and analytics features which

are paid pro features. But like 90% of

notebookm is completely free. This

primary workflow is already extremely

powerful. But if you combine it together

with other applications like Google AI

studio or Firebase Studio, which are

also both free Google products, you can

make it even more powerful to build

entire workflows and apps. And you don't

even need to know how to code. All

right. Now you have a good general

overview of the setup for Notebook LM.

Let's now cover each of these features

in detail. Starting off by adding some

sources. I am interested in building a

vertical AI agent app or learning

languages. And I want notebook LM to

help me figure out what I should be

building, including things like

analyzing industry trends, different

pain points, brainstorming and

validating different ideas before coming

to a four set of features to build for

my MVP, my minimal viable product. This

is a report from OECD on the trends

shaping education 2025. Um, and this is

a very very reputable report about the

education space. So I'm going to add

that as a source. I'm also going to add

a couple YouTube videos. This is one

from Y Combinator about how to build AI

agents, how to validate startup ideas.

So notebook alum can help me with that

as well as an article which is very

popular for startup founders. Then I

will copy some text which is what I got

from a deep research over here from

Google Gemini that goes into a deep dive

of the competitor analysis of the

education space including the core SAS

competitors in the space. Later in the

video, I'll be going into more detail

about how to use other applications like

deep research in order to get really

good sources. But for now, I just want

to show you all the different types of

sources that notebook elm can take,

including this copied text information.

But that's not all. I also want to click

the discover button here to search for

additional sources that I think will be

useful in this task. To get a better

understanding of customer pain points, I

wrote look at a variety of UGC user

generated content sources including

Reddit Xora defined painoints and

language learning education space. Click

submit. And I'm going to add all these

Reddit sources, but not this one because

I'm more interested in user generated

content, not academic perspectives. So

I'll click import. Rename them by typing

user in the beginning. So I know that

these are about users pain points. I'll

name these previous ones evaluate

because it's for evaluating different

startup ideas. And we'll call this one

trends because it's about industry

trends. I also want to search and add a

few more sources on industry trends.

Always good to take a look as well on

what it comes up with. Okay, this looks

legit because it's a study. And we'll

add a couple more sources. Moving on to

the chat interface. It already gives you

a little overview over here. But I do

want to dive deeper. Starting off with a

better understanding of the industry

trends specifically. So, I'm actually

going to uncheck everything except for

the industry trends and ask it to

summarize the industry trends for

language learning. Click enter. And it

gives us a little summary talking about

the language learning app market is

experiencing rapid expansion influenced

by several key trends, drivers,

restraints and opportunities, key market

insights and drivers blah blah blah. It

also gives you the sources where it's

coming from. So you can actually um very

easily double check and click into this

as well. This is one of the main

functionalities of notebook LM because

it's so grounded in the sources that

it's being provided. This has much less

room for hallucinations. So yeah, key

market insights and drivers, market

segmentation by type, by application, by

language, okay, and by geography. So I

want to ask some questions like what

about AI first language learning apps

versus traditional SAS apps cuz I want

to build something that is an AI first

like vertical agent um application. Who

are the industry leaders here and where

are the gaps? So it says in AI language

learning app the market growth is

largely fueled by advancements in

digital education technology. So it's

showing that Dualingo for example is

introducing AIdriven custom lessons and

realtime voice recognition. And you can

see that it would actually just tell you

that the provided sources do not

explicitly differentiate between AI

versus traditional SAS apps. So it

really is grounded in the sources that

you provide. But it does show you all of

the different competitors that are in

the market right now and gaps in the

market where AI could be able to help

with. Very interesting. So before we

move forward, make sure that you do

click save to notes because if you don't

do this, all of this is actually going

to disappear because notebook has a

privacy policy where it doesn't actually

save the chat history and it doesn't

train on it as well. Okay, cool. Going

to do something similar for the user

pain points. Summarize the user pain

points and we can see some of the key

pain points including effectiveness and

fluency challenge. Many language

learning applications are perceived as

failing to lead to true fluency, depth

or mastery, motivation, engagement,

constant accessibility of traditional

methods, fear of mistakes, etc., etc.

Also, make sure to save to note here so

we don't end up losing that information.

So, you should continue doing this and

asking more questions and learning more

from these sources. But I want to draw

your attention now to the studio section

where there are also other features to

help you understand this content better.

This is one of my favorite ones called

the audio overview. So, let's click this

and it will generate an AI podcast

specifically for the user pain points

that we selected. That took a couple of

minutes. So, it is ready.

>> Let's unpack this. Today, we're diving

head first into something, well,

something many of us have tried or maybe

are trying right now

>> or thinking about trying. Yeah,

>> exactly. Learning a new language. So,

for this deep dive, we've gathered quite

a stack of sources. We've got stuff from

Reddit, uh, real learners, teachers.

>> There's some really candid stuff there

>> and a whole heap of opinions on

Dualingo. What was it from Tik Tok,

actually?

>> Ah, Dualingo always sparks a debate,

>> doesn't it? Just

>> All right, let's skip ahead a little

bit.

>> People repeatedly pointed out it lacks

explanations

>> on grammar, tenses.

>> Exactly. Tenses, grammar, and

importantly,

>> you can actually join this conversation

as well. King the interactive mode.

Hello. So I wanted to talk a little bit

more.

>> Well, hey there. Want to share your

thoughts on this?

>> Yeah. So I want to understand a little

bit more what exactly is the primary

pain point of using something like dual

lane go like what is the singular um

issue that people have with it?

>> That is such a great question to start

with and it actually really gets to the

core of a lot of the feedback we found.

>> Yeah, that's really smart to dig into

that right away. If we connect this to

the bigger picture, the primary pain

point with Duolingo based on what we've

seen isn't that it's bad per se,

>> right? It's more about what it isn't

designed to do.

>> Exactly. It excels at things like

drilling vocabulary and basic grammar,

giving you those consistent snackable

lessons,

>> which is great for getting started. But

the singular issue, the one that keeps

coming up, is that it doesn't really

build conversational fluency or deep

understanding of how

>> how would using something like AI be

able to address these pain points.

>> That is an excellent follow-up question

and it really gets to the heart of where

edtech is going.

>> Yeah, that's exactly where we wanted to

take the conversation.

>> So, if the pain point is the lack of

conversational fluency and deeper

understanding, AI has some fascinating

potential there.

>> The sources really highlighted a few key

ways. One of the biggest is enabling

natural conversation practice.

>> Imagine having an AI you can talk to

anytime.

>> A practice buddy, but one that's always

available.

>> Exactly. One that gives immediate

feedback on grammar, carings you gently

without judgment.

>> Basically, a patient conversation

partner available 24/7 who won't judge

you.

>> This directly addresses that fear of

looking.

>> Isn't that just so cool? The first time

that I heard this, I was freaking mindb

blown. This is two AI hosts. It's

insane, isn't it? So, you can do this. I

think it's a really great way of really

exploring and diving deeper into all of

this information, understanding

everything much better. And I want to

show you now my secret pro tip for how I

use audio generation to learn things

really, really quickly. It's how I keep

up with a lot of the AI trends and

things that are going on. And I'm going

to use another example for this cuz

let's learn about something new. But I

do want you to put a bookmark on this

specific example because we're not just

going to end it here. I'm actually going

to show you how to build out this AI

application for learning languages using

a combination of notebook and AI

assisted coding tools like Firebase

Studio, which is also a Google. But

first, let's cover the core features of

Notebook LM. All right, here's my pro

tip. Are you ready? This is another

notebook about social media's impact on

youth mental health. And I generated an

audio overview. We can take a listen to

this. Welcome to the deep dive. We're

here to sift through all the noise, pull

out the crucial insights, and well, get

you properly informed. Today we're

diving deep into something that touches

almost every family, social media and

etc etc which is really great if you

want to dive really into this um and

really enjoy this process. But sometimes

I just need to learn the information

really really really quickly and there's

like a lot of fluff during these

podcasts, right? So what I like to do is

actually come here and download this

audio and put it into Google AI Studio

which is also a free tool. Then I ask it

to get a transcript of this podcast and

then condense it by removing all of the

fluff and changing it to a single person

as opposed to two people talking. And

then I can download it again and play it

at 2x 3x speed and get through all of

that information super quick.

>> The provided text is the growing concern

about social media impact on mental

health particular.

>> And that my friends is my super secret

tip on how to learn things really really

quickly. Just listen to this while

you're like running around, you know,

going outside, hiking, working out,

whatever. And actually, did you know

that when you're moving around and

listening to audio, it helps you

concentrate better? Anyways, going back

to notebook, I want to show you guys

some of the other features that you can

use under the studio tab. Um, you can

also generate a video overview. So, we

can have a video over here. And let's

actually check it out. All right, let's

talk about something that's well, it's

everywhere, right? Social media, teen

mental health, you see the headlines all

the time. But we're going to go a little

deeper, get past those headlines, and

see what's really going on. And it's not

just talk. The numbers really back this

up. Check this out. The NYC Health

Department found that 27% of teens on

social media have an anxiety diagnosis.

Now, compare that to teens who aren't on

social media. That number plummets to

just 9%. We're talking three times

higher. That's a huge, huge gap. It's

not just a feeling. It's something

health crisis. So boom, the debate just

became a fullon legal battle. And look,

this whole thing didn't just happen

overnight. It's been building for years.

You can trace it back.

>> Need I say more? Crazy, right? Oh my

god, it's crazy. And the graphics that

are being generated, they are in fact

accurate to what is being represented in

the sources as well. Oh my god. Like

this is such a game changer. also

generate a mind map where it's a great

way to visualize this information under

social media and impact you know you can

look at this and then you can actually

click in to these different categories

as well. So impact on youth mental

health adolescent addressing the issue

and emotional regulation you can click

into all of these. So it's a really

great way of like visualizing how the

data flows together and then in the chat

interface it would explain. You can of

course also download the mind map looks

like this and you can know throw it on a

slide or something if you're making a

presentation about it. And speaking of

presentations you can also come here

under reports and generate something

like a briefing doc very official

looking briefing document that you can

send to people as an executive summary a

study guide. This is pretty similar to

the briefing doc in terms of the content

that's being displayed. But it also has

things like a quiz where it would ask

you to answer each of these questions in

two or three sentences. So this will

help you learn things more deeply and it

has the answer key as well. Some essay

format questions and some glossery of

key terms. Something else that you can

do is that say you actually want to

study this information, you can click

convert to source and you can ask a

specific questions for the study guide.

out like for example extract three quiz

questions in the study guide. Yeah. And

it says like here are three quiz

questions extracted from the study

guide. Uh say describe two key

characteristics of the adolescent brain

that makes it particularly susceptible

to the influence of social media. So um

I can say like validate if my answer is

correct with the answer key is the

answer for two key characteristics. I

actually don't know what it is. I'm just

going to like guess here. Uh

maybe it's still developing uh

decision-m capabilities to differentiate

what is real or false. Um is susceptible

to emotional

deregulation. I don't know. See that's

correct. Uh it does not fully allow the

specific key characteristic pride in

answer guide. Okay, my bad. Rapid

growing and is highly plastic. Uh, and

yeah, okay. Well, seems like I need to

study more on this subject. But yeah,

you can see how this can be extremely

powerful if you are actually learning

things and want to test yourself. Man,

if I had notebook when I was still in

college, I I really think I just could

have like cut my study time down by 80%

probably. So jealous of you guys who are

still in school. The timeline feature

under reports is also quite useful. If

for some reason you're interested in

learning about the complex world of

bananas, you can generate a timeline of

the banana development. Very

fascinating. Prehistory ancient time of

banana all the way to the modern banana.

Wow. Amazing. Another little pro tip,

you can go to Gemini,

click canvas, copy this detail timeline,

and write create an interactive

visualization.

Wow, you have an interactive web

application dashboard about the banana

timeline. Amazing. Gemini doesn't have

the best like formatting. It's not the

most visually appealing one. So, if you

do have Claude, copy this. Go to Claude

and type create and interactive of this

banana

timeline. Yay! Amazing. The great banana

timeline. Wow. From ancient

domestication to modern crisis. 9,000

years of banana history. Wow, so cool.

So many banana activities. So, this is a

notebook about my AI agents boot camp

that we run. The FAQs feature is also

quite handy. So, this is the Lonely

Octopus AI agent boot camp, um, which

looks like this. So, what is quite

useful is that you can generate some

FAQs that you can put onto your website.

We've already seen the add notes little

button show up a few times now, but I

really wanted to emphasize how powerful

this button actually is. So, here is a

midterm performance review for Bobby

Adams, who is a product manager. We're

going to paste in some other people's

feedback for Bobby as well. And we are

actually going to generate notes um as

we're going through some of this. So,

Laya is a note-taking app. So maybe you

know you're looking through Bobby's

report and then you're generating notes

like needs to improve communication

skills um is improving technical skill

set and completed goals for Q2 whatever

it is that you want to do like OKRs are

completed feedback from manager is

relative and consider for promotion. So

yeah, like as you're going through um

some of these sources and learning more

about it, you can actually add these

notes like evaluate, we can hear in the

chat as well, evaluate Bobby Adam's

performance regarding signal signal sync

B2. Uh so we're seeing ah like great,

you know, all of this is great as well.

So you can save this as a note. It's

generating more information. So what was

a key accomplishment? Save that to

notes. We can generate a timeline of

Bobby's. So you see that generate cast

characters, timeline of main events.

Save that to notes too. And then what

you can do is so you convert everything

to source. So now there's additional

sources that we can draw from. Then you

can ask it like should we consider Bobby

promotion based upon all that

information. I hope you can kind of tell

like this um loop between adding

different things like summarizing

different things together and then

different user input as well. Converting

that back into a source. There's like

this loop between adding notes and then

putting it back into source. um that can

really help you refine and iterate upon

whatever it is that you're trying to

understand and dive deeper into. Yes.

So, we should consider Bobby Adam

promotion.

If you're a developer and interested in

building apps 10 times faster, I highly

recommend that you check out Augment

Code. It's built for real engineering

work, including debugging, writing

tests, refactoring, and navigating

complex systems. Augments Context Engine

feeds models exactly what they need from

your codebase and can handle code bases

with 10 million plus lines of code and

over 10,000 files. So no matter how big

or old that your app is, you can still

ship incredibly fast with Augment. You

can even launch cloud agents to refactor

or fix tests while your laptop is

closed. It runs in your IDE including VS

Code, Jet Brains, Vim, and Cursor. And

it's secure. Augment is ISO and SOCK 2

certified. And there's no training on

customer code ever. You can try it out 7

days for free at the link in the

description. Thank you so much Augment

Code for sponsoring this portion of the

video. Now back to the video. All right,

before I show you how to combine

Notebook LM with other applications to

make it even more powerful, I just want

to briefly cover the additional features

that you can get if you do have the paid

version of Notebook LM. As I said

earlier, you really already get so much

just from the free version alone. But

there are like a few perks. So there's

three different levels. There's the free

tier, DC pro tier, and then there's the

enterprise tier. And some of the things

that you do get on the pro tier is that

here you get 300 sources. Uh while on

the free tier you only get 50 different

sources that you can add. On the pro

tier you can also have a configure chat

section where you can define your

conversational style like change that a

little bit and also change your response

length. You can customize these a little

bit more. You can also customize these

different outputs a little bit more as

well. like for example on the audio

overview defining what the length of the

um audio is and what things that the AI

host should focus on. Honestly, these

are nice to have, but I find that it

doesn't really make too big of a

difference, I would say. Um yeah, I I

don't think it really makes too big of a

difference. Anyways, you do also have

the share tab over here. So, this is

when you can add other people um into

this notebook and you can share them

with them. You can also add a little

welcome note like for example like hello

check out you know something like that

and you can give a different access like

between full notebook and chat only and

copy link and just paste the link to

them. You also get analytics. Um so here

I don't have any analytics because you

need to share with at least four other

users to see what the uh chat activity

is but you would be able to see

analytics across different users as

well. And finally, there is enterprise

level as well where you pretty much get

all these features, but you just get

more security like privacy and security

guarantees from it. So, in terms of the

actual pricing models, it seems like you

don't really purchase it um by itself.

You really just purchase it based upon

the package of which Google product that

you have. So, in my case, I get the pro

um version because I have Google

Workspace for business. Okay, let's now

talk about Notebook LM in combination

with other applications.

One of my favorite combinations is

between Notebook LM and deep research.

So using deep research, I would generate

something that is of deep dive into a

specific topic and then I would add that

as a source into Notebook LM. For

example, looking at very specific market

trends for electric vehicles in Malaysia

between a time of 2020 to 2025. I would

just add that as a source in addition to

all the other sources because after all

you know with all the different bells

and whistles that notebook LM has really

the key thing like everything that is

grounded with is the sources. So if you

are able to curate better sources for

notebook LM the output that you're going

to get is also going to be much much

better and deep research is the number

one way that I use in order to generate

highquality sources. We also really like

the combination of notebook LM plus cla

which I think I showed you already

previously. Um whenever I have something

and I want to change that into a

dashboard or a visualization I just

throw it into claude and have it

generate a dashboard. You can also do

this like using Gemini or chatbt as well

but I find that with claude it comes up

with the best dashboards. Notebook plus

manis is also a really powerful

combination because you can use notebook

to distill down information and then use

manis to actually implement it. For

example generating slide decks. Notebook

LM is really good at generating the

content of the slide deck and then you

can throw it into Madness to actually

create the slide decks yourself. Madness

is also great if you're implementing

something like a design guide that you

generated from Notebook LM and you want

to actually implement that into your

landing page. Now, finally, going back

to that bookmark, remember we were doing

the research for building an AI app for

language learning.

So what we want to do now after

exploring more into the content that's

in these sources is to distill

everything down into a set of features

to create something called a product

requirements document or what is common

these days called a product requirements

prompt that we could then put into an AI

assisted coding applications something

like firebase or lovable bold or like

cloud code to actually build out the

first version of that product.

You with me? All right. So what we're

going to do now is we're actually going

to distill down this information and

combine these sources together. So the

ones that are for evaluating a startup

for example, I'm going to type here

summarize the ways to evaluate what is

considered

a good AI app idea. So it has this and

what I'm going to do is save this to the

note and I'm going to convert this to a

source. Then I'm going to do the same

thing for the trends part to summarize

the trends specific to online learning

u for language learning.

Save that to note as well and we'll add

that as a source. And finally select all

of the user sources. Summarize the user

pain points for online language

learning.

Now I'm going to select these summarized

sources and write using

these selected sources outline the key

features for an AI agent app. So I do

want it to be specifically an AI agent.

And if you're not really sure what I

mean by the difference between that,

like what it actually means to be an

agent, um I'm not going to go into too

much detail here, but you can check out

this video over here where I explain

specifically what makes an AI agent. But

yeah, I do want this to be like a

vertical AI agent application

specifically for language learning. So

moment of truth here, we're going to see

what are the key features that would be

required. See, so a good AI agent app

for language learning should incorporate

several key features. All right. So,

real-time conversational practice,

role-play scenario, 247 availability,

safe environment, incident detail

feedback and correction, including

comprehensive error analysis,

explanation of mistakes, personalized

reporting hyperpersonalization and

adaptive learning path, dynamic

curriculum, blah blah blah. Okay,

there's a lot going on here. So what I'm

going to say is So I'm going to say I'm

only going to be building an MVP

minimable viable product two core

features that will be the most

competitive and suitable for an AI

agent.

So it says here it would be real-time

conversational practice and instant

detailed feedback and correction. That

seems to be what we're going to be

building. And it has all these sources

here as well. All right, cool. Next,

we are actually going to jump into chat

GPT or Gemini if you want. Doesn't

actually matter which one. And we're

going to input. This is a prompt that's

going to guide us through the process of

thinking even more deeply about the

exact application that we want to be

building. And we want to do this step

before we jump into Firebase Studio, for

example, to actually build out this

application. Because the more clearly

you think through what it is that you

want, the more details that you have and

like what text stack you, what the vibes

are that you're going for, what is the

customer experience, what is the user

journey feel like. The more you think

through these things, the better your

end result is going to be. So the first

question to ask is, what are you

building? Can you give me a pitch? And

this is where I am actually going to go

back to Notebook LM and copy what it is

that we just said we were going to

build.

Oh, and by the way, um I will be putting

this product requirements prompt in the

description below. So you can just copy

paste that if you want to build it

yourself. Anyways, so we're going to

answer the rest of these questions. So

what's it for? Who are your users?

So I want to say like beginner to

intermediate

languages

start with English. This is the easiest

there needs. If for example you don't

know like you want to have more detail

you can also keep going back to notebook

LM here and then ask it

the questions here.

All right.

It's you know there's a lot of

information here but it's okay. It's AI.

It can take it. So what's the main user

flow? Can ask that as well. What does

the main user flow look like?

So yeah, we can copy that. It's an art

and a science here. You know, sometimes

you want to get it from notebook.

Sometimes you just want to use your own

brain a little bit as well. So what are

the must have features in your MVP? So

here's a quick checklist.

So for now I'm just going to write user

O

voice input with speech recognition real

time AI conversation agent feedback

system

memory/personalization

and push notifications.

All right. Now we will copy this. Go

into Firebase Studio and

copy this

and actually build it.

I'm not going to show you the full

details of me actually building and

iterating through all of this entire

process because I actually made a full

video on Firebase Studio which I'll link

over here um that goes through the

process of building an entire app. which

is really similar to what I did in order

to end up with this application.

>> Hi there. Let's talk about food. What's

on your mind?

>> I like potato. Eating potato.

>> Potatoes are great. Do you have a

favorite way to prepare them?

>> I like to eat raw potato.

Raw potato? That's interesting. I've

never tried it that way. Is it crunchy?

Great effort. Keep practicing. It seems

like you repeated potato in your

sentence. To make it flow better, you

could say, "I like eating potatoes."

>> Yay. Isn't that really, really cool? We

literally just went from the very very

beginning to the ideation phase,

analysis, research, all these different

things to an actual functioning

prototype. Pretty cool, huh? Yeah. So,

in this example, I use Firebase Studio

because it's free. It's also because it

seems like we're sticking mostly to like

Google products since Notebook LM is a

Google product. But really, you can also

use other tools as well. Uh you can have

like you can use Levelable. I also made

a video that goes through how to build

um applications using Levelable, which I

will link over here. And if you are a

technical person and you're comfortable

with code, you can also start building

it out using claude code which I do have

a video about here as well. So yeah.

Okay. Thank you so much for watching

until the end of this video. I really

hope this was helpful for you in

learning more about Notebook LM and just

realizing how crazy powerful it is,

right? Uh so as promised, here is a

little assessment. Please answer these

questions and put them in the comments

so you can actually retain all the

information that we covered today. Thank

you so much and I will see you guys in

the next video or live stream.