Hey guys, been working on a new version of my Second Brain and thought it would be interesting for Fastlane folks.
A few years ago I discovered the idea of building a Second Brain and started building one. The first year went into identifying the system, then I kept developing it. I loved the idea of how Leonardo da Vinci, Franklin, and Luhmann kept their logs and diaries. So my approach for the past years was something like atomic notes, essays, logs, and project notes.
Lately I've been thinking more and more about a concept where AI could be a great assistant, especially for people who don't love structuring but want to move fast. So I decided to bring this idea to life and build an approach where I'm not just creating a note archive I'll never open again, but something that becomes an asset that can evolve with me.
So, in just one day (or maybe actually over several years), I built my new Second Brain prototype, powered by Claude Code (for now).
The point is to build a system where AI acts as a thinking partner. Not an editor, not a summarizer, not a search engine.
I looked at Karpathy's wiki and other interpretations, and so far I've made a first prototype.
My Second Brain is built on the principle of asymmetry. I have raw/ folder for my atomic notes, essays, book excerpts, project notes. This is my territory. AI never writes there, never edits, never suggests deletions. Every word there is mine.
AI has wiki/, the synthesis layer. Topics (broad hubs like Learning or Note-Taking), concepts (specific synthesized statements with history), an index, a log, a contradiction map, a registry of open questions. AI builds and maintains this. I read it and rarely edit.
Between the two layers, a conversation. I write a note, we discuss it. AI shows me what I already have, where the new thought continues an old one, where it contradicts, which concepts it touches. It asks Socratic questions. Suggests connections. And it writes nothing into the Wiki until I explicitly say "write." Two-phase ingest: first the conversation, then the commit.
Another critically important point. What goes into raw/ is material I've already processed. It's not copy-paste from YouTube, courses, articles, or books. These are already my thoughts, in note form in my own words, or full walls of text. AI doesn't work with raw material from the internet. It works with what I've already filtered through myself.
Four modes of operation.
Discuss, discussion of a note. Two stages: conversation without recording, then commit. AI writes nothing on its own. It even formulates the Evolution entry and waits for my confirmation or a reformulation.
Consult, conversation with what's accumulated. "I'm thinking about a new business, what do I already have on this topic?" AI searches my base, answers first with my own knowledge with references to specific notes and concepts, and only then, explicitly separating it, adds the model's general knowledge. If the base is empty, it says it's empty. It doesn't mask a generic answer as a personalized one. And critically, consult is not logged. It's an ephemeral conversation. If a note is born from it, it's logged as a normal ingest.
Review, a conversation about patterns over the week or month. What was in focus, what got abandoned, which open questions are hanging, how thinking shifted.
Lint, an audit of Wiki health. Full check: stub pressure (concepts frequently mentioned but without a page), stale entities (pages without updates despite new notes on the topic), broken backlinks, cross-reference gaps, unresolved tensions, drift candidates. AI produces a report by category, I decide what to fix.
Plus a sub-mode, Drift: synthesis of slow thinking shifts. For concepts I've touched five or more times in a period, AI reads the entire Evolution and proposes a Drift summary, a single line describing the overall trajectory.
Versioning via git. Every AI action (except consult) is a separate commit. This gives full chronology and the ability to roll back any Wiki change. The vault is a git repository.
Another important detail. Contradictions. If AI notices a contradiction or repetition and I don't react, it comes back to it on the next turn. And the turn after. Because the whole point of the system is exactly this: to see what I would have missed on my own.
It's a lab where knowledge doesn't accumulate but compounds. AI does the work that slows me down: connections, indexes, cross-references, catching repetitions, tracking evolution. And I do what only I can: thinking, doing, observing, formulating, choosing.
If da Vinci had an assistant that didn't get in the way of creating but remembered every page of his notebooks, knew how to structure, highlight the essence, and wrap things up, this would probably be it.
The engine has started turning. I've already added two atomic essays. In the next part (if there is one) I think it'll be about what this system has transformed into after a few weeks, or whether it survived at all.
Why did I decide to share this here?
I actually think the idea of an enhanced Second Brain is one of the best potential augments for entrepreneurs. My thesis is that entrepreneurship is the ultimate self-awareness and self-discovery journey. And it's amazing when you can use AI where the frontier knowledge is your knowledge plus AI, not generic AI knowledge that you're trying to fix through personalized prompting.
The Second Brain idea can be used to search for new ideas, develop products, tinker with possibilities, stress-test your own thinking, spot patterns across months of work, and turn scattered insights into compounding leverage.
And, guys, a quick question.
Is it better to make this kind of material a separate thread (and post further developments here if there are any on this topic), better to write it in the journal I had going, or actually better as a single message in the off-topic thread?
For now I figured a separate thread makes sense.
Thank you for reading!
A few years ago I discovered the idea of building a Second Brain and started building one. The first year went into identifying the system, then I kept developing it. I loved the idea of how Leonardo da Vinci, Franklin, and Luhmann kept their logs and diaries. So my approach for the past years was something like atomic notes, essays, logs, and project notes.
Lately I've been thinking more and more about a concept where AI could be a great assistant, especially for people who don't love structuring but want to move fast. So I decided to bring this idea to life and build an approach where I'm not just creating a note archive I'll never open again, but something that becomes an asset that can evolve with me.
So, in just one day (or maybe actually over several years), I built my new Second Brain prototype, powered by Claude Code (for now).
The point is to build a system where AI acts as a thinking partner. Not an editor, not a summarizer, not a search engine.
I looked at Karpathy's wiki and other interpretations, and so far I've made a first prototype.
My Second Brain is built on the principle of asymmetry. I have raw/ folder for my atomic notes, essays, book excerpts, project notes. This is my territory. AI never writes there, never edits, never suggests deletions. Every word there is mine.
AI has wiki/, the synthesis layer. Topics (broad hubs like Learning or Note-Taking), concepts (specific synthesized statements with history), an index, a log, a contradiction map, a registry of open questions. AI builds and maintains this. I read it and rarely edit.
Between the two layers, a conversation. I write a note, we discuss it. AI shows me what I already have, where the new thought continues an old one, where it contradicts, which concepts it touches. It asks Socratic questions. Suggests connections. And it writes nothing into the Wiki until I explicitly say "write." Two-phase ingest: first the conversation, then the commit.
Another critically important point. What goes into raw/ is material I've already processed. It's not copy-paste from YouTube, courses, articles, or books. These are already my thoughts, in note form in my own words, or full walls of text. AI doesn't work with raw material from the internet. It works with what I've already filtered through myself.
Four modes of operation.
Discuss, discussion of a note. Two stages: conversation without recording, then commit. AI writes nothing on its own. It even formulates the Evolution entry and waits for my confirmation or a reformulation.
Consult, conversation with what's accumulated. "I'm thinking about a new business, what do I already have on this topic?" AI searches my base, answers first with my own knowledge with references to specific notes and concepts, and only then, explicitly separating it, adds the model's general knowledge. If the base is empty, it says it's empty. It doesn't mask a generic answer as a personalized one. And critically, consult is not logged. It's an ephemeral conversation. If a note is born from it, it's logged as a normal ingest.
Review, a conversation about patterns over the week or month. What was in focus, what got abandoned, which open questions are hanging, how thinking shifted.
Lint, an audit of Wiki health. Full check: stub pressure (concepts frequently mentioned but without a page), stale entities (pages without updates despite new notes on the topic), broken backlinks, cross-reference gaps, unresolved tensions, drift candidates. AI produces a report by category, I decide what to fix.
Plus a sub-mode, Drift: synthesis of slow thinking shifts. For concepts I've touched five or more times in a period, AI reads the entire Evolution and proposes a Drift summary, a single line describing the overall trajectory.
Versioning via git. Every AI action (except consult) is a separate commit. This gives full chronology and the ability to roll back any Wiki change. The vault is a git repository.
Another important detail. Contradictions. If AI notices a contradiction or repetition and I don't react, it comes back to it on the next turn. And the turn after. Because the whole point of the system is exactly this: to see what I would have missed on my own.
It's a lab where knowledge doesn't accumulate but compounds. AI does the work that slows me down: connections, indexes, cross-references, catching repetitions, tracking evolution. And I do what only I can: thinking, doing, observing, formulating, choosing.
If da Vinci had an assistant that didn't get in the way of creating but remembered every page of his notebooks, knew how to structure, highlight the essence, and wrap things up, this would probably be it.
The engine has started turning. I've already added two atomic essays. In the next part (if there is one) I think it'll be about what this system has transformed into after a few weeks, or whether it survived at all.
Why did I decide to share this here?
I actually think the idea of an enhanced Second Brain is one of the best potential augments for entrepreneurs. My thesis is that entrepreneurship is the ultimate self-awareness and self-discovery journey. And it's amazing when you can use AI where the frontier knowledge is your knowledge plus AI, not generic AI knowledge that you're trying to fix through personalized prompting.
The Second Brain idea can be used to search for new ideas, develop products, tinker with possibilities, stress-test your own thinking, spot patterns across months of work, and turn scattered insights into compounding leverage.
And, guys, a quick question.
Is it better to make this kind of material a separate thread (and post further developments here if there are any on this topic), better to write it in the journal I had going, or actually better as a single message in the off-topic thread?
For now I figured a separate thread makes sense.
Thank you for reading!
Dislike ads? Become a Fastlane member:
Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)
Access Restricted: Unlock 1,000,000+ Posts
Stop Peeking Through the Keyhole.
Open the Door.
You hit a wall because the best advice isn't free—it's earned. Since 2007, MJ DeMarco has been active here daily (99.9% active rate), building a war room for entrepreneurs who refuse to settle for mediocrity.
- Active Daily: Life-changing content posted dozens of times every day. No dead threads.
- MJ's Inner Circle: Direct access to the author of The Millionaire Fastlane.
- Vetted Network: Connect with founders scaling to 7 and 8 figures.
- Proven Roadmaps: Strategy over motivation. Execution over "hustle porn."
"SCALED TO 6-FIGURE MONTHS"
"Tips and advice here saved me from huge mistakes, others allowed me to scale my business to 6-figure months."— User MitchC
"INSANE QUALITY OF PEOPLE"
"The number of high quality people I've met from this forum has been insane. All of it predicated on building real value."— User Richard Peck
"You are the average of the five people you surround yourself with."
Are you surrounding yourself with success?
