Speaker A
What you're looking at right over here are a bunch of my YouTube videos being ingested into an LLM wiki. This LLM wiki, as you can see if I zoom in, are different YouTube videos, and what's connecting them are different relations.
Nate Herk demonstrates how to build and use LLM wikis to organize YouTube videos and other data into a smart, interconnected AI knowledge base.
Key Takeaways
- LLM wikis can automatically organize and link large amounts of video and meeting data into an intuitive knowledge base.
- Fable’s LLM shows better contextual and emotional understanding compared to other tools like Opus 4.8.
- Integrating multiple data sources into an AIOS enables powerful insights and automated content creation.
- Obsidian is a practical front end for managing LLM-based personal knowledge bases.
- Continuous data ingestion and proper routing improve the AI’s understanding and usefulness over time.
Summary
- Nate Herk shows how his YouTube videos are ingested into an LLM wiki that creates a second brain by linking concepts automatically.
- The LLM wiki allows easy navigation through related videos, concepts, and backlinks, making complex data simple and user-friendly.
- He compares Fable’s capabilities with Opus 4.8, highlighting Fable’s superior understanding and emotional prompting for beginners.
- Nate explains how multiple LLM wikis power his AI Operating System (AIOS), including one for YouTube transcripts and another for meeting recordings.
- The AIOS can generate visual summaries and insights about business performance, showing the power of integrated data.
- He references Andre Karpathy’s approach to building personal knowledge bases using LLMs and Obsidian as a front end.
- The video includes a step-by-step tutorial on setting up Obsidian, creating vaults, and ingesting data into the LLM wiki.
- Nate emphasizes the importance of routing data correctly and continuously evolving the knowledge base for better results.
- He demonstrates how to ingest multiple data sources, organize concepts, and use backlinks to create a layered, interconnected knowledge system.
- The video promotes a free course available in Nate’s community for building your own AIOS using these techniques.
Chapters
- 00:00Introduction to LLM Wiki and YouTube Video Ingestion
- 01:00Navigating the Mind Map and Backlinks
- 02:02Fable vs Opus 4.8: User-Friendly Interfaces
- 02:57Multiple LLM Wikis in AIOS and Use Cases
- 03:58Business Insights Generated by AIOS
- 04:50Setting Up Obsidian and Creating Vaults
- 06:15Running Claude and Initial Schema Setup
- 07:29Organizing Data and Indexing Concepts
- 09:17Ingesting Data and Cross-Linking Wiki Pages
- 11:13Adding Safeguards and Expanding Data Sources
Full Transcript — Download SRT & Markdown
Speaker A
So, we're starting to see this actual kind of like second brain of all of my YouTube videos and how they relate to each other, and all of this knowledge makes my AI OS so much smarter. And the coolest part about this is I didn't have
Speaker A
to connect these concepts at all. I was able to just say, "Hey, Claude code, go grab my YouTube videos and then ingest them into this wiki." And this thing continuously grows and grows. If I zoom in a little bit, let's open up one of
Speaker A
these videos. So, right here I've got Nano Banana two websites. When I open this up, we can see some information up here, but then as we scroll down, we can see, you know, summary, key takeaways, and other tools and things that are
Speaker A
mentioned and other techniques that have been discussed. And I can follow all of these links around. Let's say I'm interested in GitHub. I can click on the GitHub. I can see what this is about, and then I can see other times that we've
Speaker A
referenced GitHub. Here's some information that connects GitHub to Vercel. Why don't I click into that and learn some more about Vercel? And then Vercel can take me back to Claude code where I once again can follow all of
Speaker A
these backlinks until I get to where I need to go. And so, as this whole mind map of these YouTube videos starts to grow, we're able to see all of this come to life. And today I'm going to show you
Speaker A
guys exactly how you can get up and running with something just like this in about 5 minutes. It's so much simpler than you may think. Now, what's impressive about this isn't the fact that Fable was able to ingest all of it,
Speaker A
it's what Fable can do once you've given it the power of all of this data, 'cause we all know that data is king, context is king. Here's a cool example. I asked Fable in one prompt, I said, "Hey, I
Speaker A
want you to basically turn this messy blob of YouTube transcripts connections into something that people could actually look at and understand. I want this to be a simple resource that's not overwhelming, but shows my audience how these tools and techniques and ideas
Speaker A
connect to each other." And now we have this super cool HTML which I can click into, and I can see different ideas up top, agentic workflows and what it connects to. It connects to routines.
Speaker A
Routines connects to deterministic versus agentic automation, which connects back to N&N and Claude code and all of this kind of stuff, and it's just amazing. In my mind, something like this is a much more user-friendly interface than something like this. And what I
Speaker A
think is awesome about this is that I was able to prompt it in an emotional way. I said things like, "In a way that a beginner could understand and could click through and it wouldn't overwhelm them." And something like Opus 4.8 just
Speaker A
doesn't understand what that means as well as Fable. To show you what I mean by that, this is something that I worked on with Opus 4 almost a full day. We went back and forth, we built this thing
Speaker A
out, and I just didn't like it enough to share it with people because it felt overwhelming. It felt confusing. And the database on the back end that powers this is the exact same one. So anyways, what we're looking for here is same
Speaker A
thing. You can search through tools, techniques, videos. There's kind of like a layer framework that we discussed with the orchestrator, the models, the inputs, all of this kind of stuff. And it has all the same data, and I can
Speaker A
still click into these things, and I can follow the backlinks, and, you know, it's kind of the same idea, but once again, this version is just so much simpler to me, and I like it more. You can see as we click on a concept, we're
Speaker A
able to see on this right-hand side videos that it pulled this data from. We can read a little bit more about it, and we can see what else it's connected to.
Speaker A
So that's just one very small example. If you guys have been following me for a while, you know that in my AIOS, I have a few different LLM wikis. This is my YouTube transcript one. I've also got like my Herc brain one, which is pretty
Speaker A
much where I put all of my meeting recordings. So all of my meetings, whether they're internal or external, I store them here, and that's how I'm able to see how the different concepts that I'm talking about with people have
Speaker A
evolved and how they are going to continue to evolve. And when I'm scripting community posts, LinkedIn posts, writing emails, it takes all of this stuff into account because it knows everything about me and my business. So much so that right before this video, I
Speaker A
said, "Hey, Mr. Fable, I want you to go ahead and just tell me a story about the past 6 months." So, you know, we're halfway through 2026. Build me a visual journey of what we've done so far in
Speaker A
2026. And this is what it gave me in one shot. It was able to pull this picture of me. It pulled our logo, and you can see that this thing even feels... And you can see that the branding of this even
Speaker A
feels like AIOS. It's kind of dark mode, blue graph colors, and this is what it gave me. It pulled actual stats like how many subscribers I gained. Um, it has our highest revenue month, which I'm going to blur out, but it was able to look at
Speaker A
all this data and just pull it for me. This was a big pivot I made this year. I went from pretty much doing only end-to-end content to doing a lot of cloud code content. And you can see how
Speaker A
this was able to pay off if we look at my average views and our revenue and how the business has grown since I made that pivot. Then we look at some other things like how our churn has changed, how our
Speaker A
conversion has changed, other things about our revenue. But look at this. This is pretty funny. It pulled this different picture of me. If you guys remember, the one up at the front was a smiling one. This is one of me thinking.
Speaker A
And so it's able to just crawl through so much of the data and the resources that it has available inside of my Herc 2 project. It shows the whole funnel of the business, which proves that it understands how people enter our
Speaker A
ecosystem and all the decisions they can make inside of our funnel and where we try to push them to. Anyways, the point I'm trying to make there is the more data you give your projects, the better.
Speaker A
But specifically, making sure that you route them in the right way. And that's what the LLM wikis are really good at.
Speaker A
This is my Herc 2 project. You guys know that this is my AIOS, and we have so much information in here. We have different wikis, different projects, everything that I've worked on, and that's what you guys are trying to
Speaker A
build. By the way, if you want to go through a full free course where I show you how to do that, in my free school community, link is in the description, I've got a full build your own AIOS
Speaker A
course in there, completely free. So, link is in the description for that. Okay. So, here's where all of this started. Andre Karpathy said, "LLM knowledge bases. Something I'm finding very useful lately is using LLMs to build personal knowledge bases for
Speaker A
various topics of research interest." He indexes the sources, which I'm going to show you how to do, and then he uses something like Obsidian as the front end, which is what you guys just saw.
Speaker A
So, the first thing you want to do is go to obsidian.md and then install this for whatever operating system you're on. So, in my case, I installed this for Windows, run the wizard, get it set up, and open up the app. When you open up
Speaker A
the app, it will look like this, except for you won't have this stuff, and then you're going to go down here and click on Manage Vaults. It might just pop up like this. And then you're going to go
Speaker A
ahead and create a new vault. So, this one I'm just going to call AI test, and then you're going to choose a location for this. So, it could be on your desktop or what I typically like to do.
Speaker A
is I put my vaults inside of my Herc 2 project. So, my Herc 2 is able to look at a ton of these different little LLM wikis that I have split up and separated by topic essentially. But, for this
Speaker A
example, I'm just going to put this one on my desktop. So, I'm going to go ahead and create that wiki. You can see here or sorry, not wiki, vault. We're going to turn this into a wiki. So, right now
Speaker A
this is what we have and then what I'm going to want you to do is you're going to go into Claude code or wherever you use Claude code. So, in this case, I'm using VS Code and you're going to open
Speaker A
up that vault in something like this. All right. So, I've just opened up the vault. You can see that we have a dot Obsidian folder. We have a welcome.md.
Speaker A
That's just going to be, you know, the default when you open up an Obsidian vault. And then what we're going to do is open up Claude code. So, however you like to use it, in VS Code I like to use
Speaker A
it in the terminal. So, I'm going to go ahead and run Claude to open this up and then we're going to go ahead and get started. Now, one thing to call out here is we see that it says, "Okay, Fable,
Speaker A
you only have it until July 7th on your limit, otherwise it will be usage credits." I did however see this tweet from Thor that said, "Yeah, that's true, but we do plan to bring it back to part of your subscription as soon as
Speaker A
possible." As mentioned in our original blog post, which if you see at the bottom of this blog post, it does actually mention that. So, it doesn't say when, but hopefully they will be able to extend the window and bring it
Speaker A
back as a standard part of your subscription plan. Anyways, what you're going to do after that is you're going to go to this page. I will have this link in the description. It is Karpathy's LLM wiki gist and I'm
Speaker A
literally just going to copy this entire thing. If you want to stop and read it, feel free, but I'm going to copy this entire thing and what we're going to do is take that back into our Claude and
Speaker A
paste that in there. And then what you're going to do is just go ahead and take a screenshot of this so you can paste it in. I said, "You are now my LLM wiki agent. Implement this exact idea
Speaker A
file as my complete second brain. Guide me step by step. Create the Claude.md schema with my full rules. Set up the index, the log, define folder conventions, and show me the first ingest example. From now on, every interaction follows the schema." So, I'm
Speaker A
going to go ahead and send that off. We are using Fable here. Like I said, you probably don't need Fable. Fable's probably overkill for this. It's about what Fable does after you have all that data in there. So, if you want to switch
Speaker A
this back to Opus, run the ingest, and ingest future documents with Opus, that's probably a better call, honestly.
Speaker A
I'm just going to be showing you Fable in this video. Now, what's really cool is as you start to put different stuff in here, it's going to sort of dynamically change the structure. So, let me show you what I mean by that. If
Speaker A
I open up this wiki, you can see that in the wiki I have comparisons, I have concepts, I have sources, because this one is about my YouTube videos. So, it's read through them and it's analyzed that. We've got techniques and we've got
Speaker A
tools. But, if I switch into something like my Herc brain, which is the one that's more so around like my meeting transcripts, you can see that this is pretty much a flat structure. It basically just has all of my meeting
Speaker A
recordings right in here, and it didn't want to organize them yet. And maybe at some point if we run some, you know, sweeps through, it will find some different folders to organize them in.
Speaker A
But, sometimes keeping this flat is actually better. And by flat what I mean is basically just having everything in the wiki, rather than having it drill down to even more folders. The reason being, you want to make sure that your AI can
Speaker A
easily search through all this stuff. We have the raw, which is where you put stuff. Then the AI will read everything in the raw and ingest it into the wiki, and that's where it might take one source and split it up into like five or
Speaker A
six, or maybe even 10 little wiki pages. Then we also have the index, which is like a table of contents. We have the log, and then the dot md files are all of the other wiki files. And this exact
Speaker A
structure is how this Herc brain one is set up. As you can see, it's very flat.
Speaker A
But, if we go back into my YouTube transcript one, this one's not flat, right? This one has all of these other subfolders, like we just talked about.
Speaker A
And to show you what that looks like in this example, here is my index. You can see all the tools, you can see all of the uh techniques, everything that has been mapped out here with all the backlinks. And then the log, you can see
Speaker A
that I did a few batch ingests here. I've done one there, and then every time that I've ingested another YouTube video or ingested other data sources, it will show a log of that there. The whole point of this is that the AI can
Speaker A
incrementally build and maintain this wiki. So, it needs to be able to look at things like the index and the logs and all the backlinks to actually crawl around and find the data that you're looking for. All right, so now you can
Speaker A
see that this is done. Our project is set up. We have our index, which is blank. We have our log, which is pretty much blank. And then we have our raw folder and our wiki folder. So, what you
Speaker A
can see in here already is that in the raw, it processed this LLM wiki idea.
Speaker A
This file is basically the gist that Karpathy wrote up. So, it decided to ingest that. And then in the wiki, it's already planned out to have concepts, entities, and sources. So, that is what we have to start with. Now, what I want
Speaker A
to do is we're going to ingest two different things and I'm going to show you different ways you can do it. So, the first thing that we're going to ingest is the Claude Fable 5 and Mythos 5 system card. So, I'm going to go ahead
Speaker A
and download this as a PDF. What I'm going to do is take this PDF and I'm going to drag it into the raw. So, now that PDF lives right there in the raw.
Speaker A
And then what I'm going to do is we're going to take this OpenAI previewing GPT-5.6 soul and we're going to just do this as a URL instead. So, what I'm going to do is I'm going to paste in the
Speaker A
URL and I'm going to say, "Hey Claude, read this article and then ingest that into our wiki here. And then also, I dropped in a PDF in the raw called Claude Fable 5 and I want you to also
Speaker A
ingest that." And that is all that I'm going to tell this model. Once again, it should be understanding how this is set up. We should see a new record in the index as well as a new record in the log
Speaker A
and some new sources in the wiki. So, whether Fable decides to turn this PDF into one or five or maybe even 50 wiki pages because of how big that PDF was, same thing with the URL, I will let you
Speaker A
guys know when this finishes up. Okay, so that finished up. It took about 10 to 12 minutes and you can see here that out of those two sources, it created 20 wiki pages and they're fully cross-linked. Now, look at this. The
Speaker A
connection that made this worth having as a wiki instead of two separate summaries, the two sources reference each other, and Frontier Model Cybersecurity is where that lives.
Speaker A
OpenAI benchmarked GPT-5.6 Soul against Mythos Preview, and I flagged the thing easy to miss reading them separately.
Speaker A
OpenAI compared to the April predecessor, not to Mythos 5, and the two labs used different harnesses, so the numbers don't line up directly. So, anyways, let's go ahead and pull open this wiki full screen and take a look.
Speaker A
Okay, so this is what it looks like. You can see we've got OpenAI down here, and these are the ones that it relates to, like the Claude code said, it referenced Claude Mythos 5 in the article. So, that's
Speaker A
pretty cool. And we can see sort of the distribution here. We can see how much we've got things like government-coordinated model releases.
Speaker A
We've got layered safeguards, competitive use safeguards. And if we go over here to the wiki, we can see that we have concepts, we have entities, we have sources, and then we have topics.
Speaker A
So, the entities is cool because in here we have models. We've got Fable, Mythos, Mythos Preview, Opus 4.8, GPT-5.6. We've got Tropic and OpenAI. And then if we go to the log, you can see this was the initial setup, and then we had the
Speaker A
OpenAI article and the Claude Fable 5 system card. So, the lesson here is that we now have this system where we have Claude code that looks at a bunch of our data sources, right? It looks at the wiki, and it looks through potentially
Speaker A
multiple wikis. And inside the wiki, what happens is there are routing rules set up so that our agents are able to figure out where to look for what specific thing, like the data that it's looking for, because it has to crawl
Speaker A
through basically all of this in an efficient way, so it's not wasting our time and our tokens to find the right answer. And that theory is basically what Claude code is. It's basically figuring out how can my Claude at MD
Speaker A
work as a router to be able to look through my past projects, to be able to look through my business context, and find the right spot. And so, from here, once you've already started to get the structure figured out, you're just going
Speaker A
to start adding more data sources, and you're going to watch how this evolves, and you're going to constantly check and see if it all makes sense. Because if you do a batch ingest, and you don't like the way that it organized some of
Speaker A
these folders and files, then maybe you go ahead and change it up a little bit.
Speaker A
You know, you start to open up these pages like competitive use safeguards, read about it, click through, and see if it all still makes sense. And if you don't like how things are happening, then update the rules in the way that
Speaker A
you ingest. Like I said, every LLM wiki that I've set up, they have different structures and a little bit different rules because of the type of data that's in there. Whether that's meeting transcripts or, you know, personal data
Speaker A
or, you know, proposals, whatever it is that you're ingesting here, make it make sense. Not only to the AI, but make it make sense to you. The whole point is that you could also go through this and follow the chain and find what you're
Speaker A
looking for. And the greatest part about all this is once you realize, "Oh, look, everything in this wiki, it's just a markdown file. It's just markdown files with routing." That means you're not locked down to using this only in cloud
Speaker A
code. You can connect your Hermes agent to this. You can connect Codex to this.
Speaker A
You can connect whatever you want to this because it's just markdown files. And if you guys want to learn more about the whole idea of like building a second brain, then check out this video right here where I go over every level of
Speaker A
building a second brain, and how you know if you actually need to like move up a little bit or move down a little bit, and figure out what's right for you and your system. So, anyways, thanks for making it to the end of the video, and
Speaker A
I'll see you guys in the next one. Thanks guys.
Topics:LLM wikiFable AIClaude codeObsidianAI Operating Systemknowledge baseYouTube transcriptsAI automationAndre Karpathypersonal knowledge management
