Before You Share Your AI Game Plan in the Comments: Protect Your System and Learn With NotebookLM

You are watching a video. You like it. Near the end the creator says something you have heard a hundred times. Tell me how you are using AI in your business. What are you doing to level up? Put it in the comments.

Your thumb is already over the keyboard. You have a good answer. You have spent years on it.

Stop for ten seconds. This article is about those ten seconds, and about what to do with them. It covers why the system you built is worth more than a comment, what you can safely share, and a smarter way to learn from the creators you watch: using NotebookLM to study a whole channel instead of one video at a time.

The question sounds friendly. Look at what it asks for.

"Tell me how you do it" is an easy sentence to say. It is a hard thing to answer well.

If you answer fully, you are giving away a process, the order of the steps, the tools and the thinking. Years of trial and error, flattened into three lines of text, for anyone to copy.

I am not saying creators who ask are doing something wrong. Many are curious. Many want to learn. Comments help a channel, and a good conversation helps everyone. I am saying the person answering should know what is being asked for.

What you actually have

Think about what is inside your AI practice if you have been at it a while.

• The prompts that took you fifty tries to get right.
• The sequence you use, which steps come first, and why.
• The tools you tested and dropped, and the reasons you dropped them.
• The mistakes you made that taught you what not to do.
• The history of your own conversations with these systems, which has shaped what they give you back.

That last one is the part people miss.

AI builds you a system tuned to your history

When you work with an AI system for a long time, it ends up with a lot of context about you. What you do. How you think. What you have tried. What you will and will not do. What you are good at. If you have kept a substantial record with these systems, what they build for you will look different from what they build for someone else.

Take two people who ask an AI the same question: build me a system to grow my business. The first has a long history. The AI knows their skills, their schedule, their audience, their past results, the way they like to work. What it builds is shaped by that. The second has none. The AI gives a general answer. Even with the same words, the outputs differ, and the first one is more likely to fit the person who asked. It lines up with their real capabilities.

That has a direct consequence for sharing. Suppose you paste your result into a comment. The reader tries it. It might not fit them at all. It was built for your history, not theirs. What looks like a plan may be a poor fit for the person copying it. Or they may ask their AI to adapt it, and what they get back is still not tuned to them.

There is a quieter point too. Unless you told the AI to ignore everything it knows about you, and build the biggest wealth machine it can, over and over, its answer reflects you. That is the useful kind of answer. It is also the kind that is hardest to hand to someone else.

The playbook you earned has value

Here is the idea I most want you to hold on to. The playbook you earned has value. You earned it. I do not mean you read about it. I mean you did it.

If you started in the early days, the history of how you got here is part of the asset. Maybe you were reading research papers before most people had heard of the technology, trying to understand what the architecture actually meant. Maybe you decided those papers were describing something real and started digging in. Maybe you spent hours every day for years going through content, running experiments, throwing away what failed. ChatGPT's public launch in late November 2022 pushed this into the mainstream, but plenty of people were already working before then. Whatever your path, you paid for it with time, money and attention. A comment is not an even trade for that.

I know this from the other side. I have given away too much in places where it mostly ended up in other people's pockets. I did a lot of the giving and not enough of the building. I have learned from it. When someone asks how I am doing this, what I am using it for and exactly how I put it into play, I ask myself a simple question first: do I have time to teach the full lesson, and is this the right place?

What to protect, what to share

This is not a case for hiding everything. Generosity builds trust. Teaching is a great way to learn. So draw a line, and draw it on purpose.

Safe to share freely

• General principles. "I use AI to summarize long documents before I read them."
• Categories of tools, without your exact configuration.
• Lessons about mistakes. "I stopped trusting any number the AI gave me without a source."
• Your opinion on what is overhyped.
• Questions you are still working on.

Share thoughtfully, in the right setting

• Your step by step workflow.
• The structure of your prompts.
• How your tools connect to each other.
• How you organize your data and your notes.

The right setting might be a paid class, a consulting conversation, a private community, or a partnership where there is a fair exchange. Not a public comment thread read by thousands of strangers, some of whom compete with you.

Do not share at all

• Client names or client information.
• Private data, credentials or account details.
• Anything covered by an agreement.
• Anything you would be uncomfortable seeing used against you.

For a closer look at the privacy side of working with AI, I wrote about owning your memory instead of training AI for free, and about setting up an AI assistant safely. On the same theme as this article, there is a piece called should you send your business idea to an influencer.

A better move: learn from them, do not just answer them

Here is the flip side. The same impulse that makes you want to comment is the impulse to learn. So channel it.

If you are watching a creator and wondering how they think, you do not have to ask in the comments and hope for a reply. You can study their body of work directly. That is where NotebookLM comes in.

What NotebookLM does

NotebookLM is a Google tool that lets you build a notebook from sources you choose. You add documents, web pages and videos. Then you ask questions, and it answers from those sources instead of from the open internet. It can summarize, compare, find where a source says something, and point back to the passage it used.

That last part is why it works for learning. A general chatbot answers from everything it has seen. NotebookLM is meant to answer from what you gave it. For research, that is a more reliable shape. I wrote about the general problem in why your AI keeps making things up.

Using it on a creator's channel

Here is the method. It turns a pile of videos into something you can question.

1. Pick one creator whose thinking you want to understand. Not ten. One. Choose someone whose work is public and whose videos you already watch.
2. Add their videos as sources. NotebookLM lets you add video links, and it works from the transcript. The number of sources per notebook has a limit, and Google changes the limits over time, so check what the current cap is. If a channel has more videos than your limit, choose the ones that matter most, or split it into several notebooks.
3. Add a short note for yourself describing what you want to learn. "I want to understand how this person approaches building systems with AI."
4. Ask structured questions. I will give you a list below.
5. Save the answers that matter into your own notes, in your own words, and check them against the videos.

Questions that pull out the thinking

• What are the five ideas this creator returns to most often?
• How has their advice changed over time? Compare the earliest videos with the latest.
• What tools do they mention, and what do they say about each?
• What mistakes do they warn about?
• What do they say they would do first if starting over?
• Where do they disagree with common advice?
• What do they say they do not know yet?
• Give me the order of steps they describe for a typical project, and cite the videos.

The questions about how the thinking changed, what they would do first and where they disagree with common advice are the useful ones. Anyone can summarize a video. What you want is the train of thought: how the person reasons, what they weigh, what they apply to AI and why.

A tip on keeping it clean

Ask for citations every time. If the answer cannot point to a source, treat it as a guess. Watch the clip it cites. A transcript can contain errors, and a summary can stretch a point. The notebook is a map. The video is the territory.

Do it ethically

A few rules keep this clean.

• Use public videos that you are allowed to watch.
• Use the tool's own features to add sources. Do not try to bypass a platform's rules.
• Study to learn. Do not copy a creator's work and publish it as your own.
• If you use an idea, build your own version from your own history and give credit when it is due.
• Do not upload private, paid or confidential material you do not have the right to use.

The aim is to understand how someone thinks, then apply it to your own situation. That is how learning has always worked.

Turn what you learn into your own system

Studying a channel is only useful if it changes what you build. So close the loop.

1. Pick one idea from the notebook that you could actually use this week.
2. Write it in your own words, including why it fits your situation and what could go wrong.
3. Ask your AI to adapt it using what it knows about your skills and your schedule. This is where the history I mentioned earlier pays off.
4. Test it small. One week. One measurable outcome.
5. Keep what works. Add it to your own playbook. That playbook is yours, so you decide where it goes.

This is the part most people skip. They collect ideas and never test them. A notebook full of insights does nothing until one of them runs in your life.

A hypothetical walk through

Here is a made up example to show how the pieces fit together. It is not a real person.

Sam runs a small landscaping company. Sam watches a creator who talks about using AI to run a business, and keeps seeing the line about leveling up. Sam wants to learn how this creator thinks about scheduling and customer follow up.

Instead of commenting, Sam builds a notebook with about thirty of the creator's videos that touch on small business operations. Sam asks: how does this creator describe the first three systems a small service business should automate? The notebook answers and cites four videos. Sam watches those four clips and finds that two of the notebook's claims were a stretch. One clip actually said something narrower. Sam corrects the note.

Next Sam asks: what does the creator say they would never automate? That answer turns out to be more useful than the first one, because it marks the limits of the whole approach.

Then Sam picks one idea, a simple follow up message after each completed job, and asks an AI assistant to adapt it. The assistant knows Sam's service area, the usual job size and the tone Sam likes. What it drafts is not what it would draft for a plumber in another state. Sam tests it for a week on real jobs and measures how many customers reply.

Later, a different creator asks viewers in the comments how they are using AI. Sam leaves a short note about the lesson on what not to automate, and says nothing about the follow up workflow, the tools or the wording. That comment helps the conversation. It does not hand over the system.

Keep your own notebook, and keep it private

Here is a move that pays off quietly. Build a private notebook of your own work. Your best prompts. Your notes on what worked. Your experiment log. The decisions you made and why you made them.

Three things happen when you do this.

1. You can ask questions of your own history. "What did I try last spring that failed?" is a question most people cannot answer, because their experience lives in scattered chats and half remembered tabs.
2. You see your own pattern. Reading your notes back, you notice what you keep returning to. That is your real playbook, and it is often different from the one you think you have.
3. You decide what is shareable. When someone asks for your game plan, you have the material organized. You can pull out a safe principle in a minute, and you know exactly which parts to keep back.

Keep private material private. Do not put client data into a tool whose privacy settings you have not read. Check the settings of any AI service before you upload anything sensitive, and keep the most sensitive material out of it entirely.

Why the process matters more than the answer

When you copy a result, you get a result. When you study how someone arrived at it, you get a method. A method travels. It applies to the next problem, and the one after that.

That is also why a full answer in a comment disappoints the person who gets it. They receive the result of your thinking without the thinking. They cannot adapt it when something changes, because they never saw why each step was there. The people who learn the most from your answers are the ones who ask follow up questions, and those conversations happen best in a place where you can have a real back and forth.

What if the comments really do contain good ideas?

This cuts both ways, and it is worth saying. A comment section is a pile of public ideas. Some are good. If you want to learn from the comments on a video, the same tool logic applies. You can collect public comments you are allowed to read and ask your AI to look for patterns and the ideas that repeat. Then ask it to use a promising idea to build you something similar, or better, fitted to your situation. The result will differ from what it would build for anyone else, because it is shaped by what you have already done.

That is a good use of the technology. It is also a reminder that other people may be doing the same to your comments. Posting a full playbook publicly makes it raw material for everyone's AI, not only for the friendly reader you imagined.

A simple decision rule for the next time a creator asks

Before you type, ask four questions.

1. Who is reading this? A friendly audience, or anyone with a keyboard?
2. What am I giving away? A principle, or a process?
3. What do I get back? Trade, visibility, a relationship, or nothing?
4. Would I be fine seeing this copied? If not, do not post it.

If the answers say share, share. A short, generous, specific tip builds trust and costs you nothing. If the answers say hold, you can still engage. Say what you learned from the video. Ask a question back. Offer to talk privately if there is a real fit.

Some comments that protect you and still help

• "The biggest shift for me was treating AI as a research assistant first, then a writer."
• "I got the most value when I stopped asking for answers and started asking for what it did not know."
• "Happy to compare notes. What problem are you trying to solve?"

These are real contributions. None of them hands over your system.

Where the real wealth in AI comes from

I want to end on the bigger picture, because it matters more than any single comment.

People want to know how others are using AI to build wealth. Fair. But no one's exact process makes you rich by itself. What builds value over time is a combination of things: your own skills, your own history with the tools, your own judgment about what to trust, your own audience or customers, and the discipline to test ideas and keep what works. Those pieces are personal. You cannot copy them from a comment. What you can do is learn how others think, build your own version, and protect the parts that took you years to earn.

If you want more on this thread, see AI for everyone, not just the wealthy, which AI you should use and why to run two models against each other, and the habit of always pushing back on an AI answer.

Limits of this advice

I am not a lawyer, and nothing here is legal advice. Tool features, source limits and terms of service change, so check the current rules of any tool or platform before relying on them. Results vary from person to person. No system guarantees income. This is a way of thinking about protecting your work and learning from others, not a promise of a result.

Frequently asked questions

Should I never answer when a creator asks how I use AI?

No. Answer, but decide how much to give. Share principles and lessons freely. Keep your detailed workflow for a setting where there is a fair exchange.

Why would my AI system be different from someone else's?

If you have a long history with an AI system, it has more context about your skills, goals and habits, so what it builds is shaped by that. Someone with no history gets a more general result.

What is NotebookLM?

It is a Google tool that answers questions from the sources you add, such as documents, web pages and videos, and can show where in those sources an answer came from.

Can NotebookLM learn from a creator's whole channel?

You can add many of a creator's videos as sources, up to the current limit on sources per notebook. If a channel is larger than the limit, pick the most important videos or split them across several notebooks.

Is it okay to study a creator's videos this way?

For personal learning from public videos, generally yes, if you use the tool's own features and respect platform rules. Do not copy their work and publish it as your own, and do not upload private or paid material you have no right to use.

How do I know the answers are accurate?

Ask for citations, then watch the cited clip. Treat uncited claims as guesses.

What is safe to put in a public comment?

General principles, lessons from mistakes and opinions are generally safe. Client information, private data, credentials and anything under an agreement should never be posted.

What should I do with what I learn?

Pick one idea, rewrite it in your own words, adapt it to your situation, test it small for a week, and keep only what works.

Can AI really use the comments on a video?

You can collect public comments you are allowed to read and ask AI to find repeated ideas. Remember that others can do the same with your comments.

Where do I go from here?

Choose one creator, build one notebook, ask five good questions, and test one idea this week. Then come back and tell me what you found.

I am Connor. Thanks for reading.