My second brain: built on Raindrop and Google Drive, gets more valuable every week
I used to save everything and go back to none of it. Instagram saves, Twitter bookmarks, TikTok favourites — a graveyard of things I found interesting once and never touched again. The save was always the end of the road. Here's how I fixed that, what the system actually looks like, and what six months of consistent use produces.
The graveyard problem
Every platform has a save feature. And every save feature creates the same illusion: that you've done something with the thing you found. You haven't. You've just moved it from the feed to a slightly less visible feed that you check even less often.
I had this everywhere. Instagram saves full of reels I'd never rewatch. Twitter bookmarks I'd open once, feel overwhelmed by, and close. A TikTok favourites tab that was basically a black hole. I was curating, just not for myself — I was curating for a version of me that never showed up.
every platform has a save feature. none of them have a "use it later" feature.
The RAM insight
The thing that changed my thinking was Andrej Karpathy's framing of a second brain — not as a filing cabinet, but as external memory. The distinction matters.
Your brain is a processor, not a hard drive. The more you ask it to remember, the less bandwidth it has to think. Every time you hold something in your head — a link, a stat, a half-formed idea — you're using working memory that could be doing something more useful. A second brain doesn't replace your thinking. It frees it up.
I don't want to run a fresh Google search every time. I want to start from where I already am — from the last version of that thought.
That framing — start from where I already am — is the thing that made it click. The goal isn't to save more. It's to not have to rediscover things you've already thought about. Every time you Google something you've already looked up, you're paying a tax on your own knowledge.
your brain handles thinking. everything else lives outside it.
What I actually built
Two tools. No exotic stack.
Anything I find on X, Instagram, newsletters, or podcasts — one tap. It goes into a collection. I don't think about it in the moment. The point is to not lose it. Collections: AI & Tech, Marketing & Growth, Startup, FinTech. Unsorted gets cleared weekly.
Every week, I open a doc and write down what I saved and why it mattered. Not summaries — my reaction to it. What it connects to. What it contradicts. One sentence of context per item. That's the thing that makes it useful instead of archived.
The Raindrop save without the Drive synthesis is just a fancier bookmark graveyard. And the synthesis without the capture is just journalling in a vacuum. The two layers are what make it a system rather than a habit.
the full loop — capture to synthesis to use. the 30-minute weekly review is where most of the value actually happens.
Things are connected — and the map shows you what
One thing I didn't expect: doing this consistently shows you what you actually care about, separate from what you think you care about.
I thought I was interested in AI tools broadly. Six weeks in, I could see that almost everything I saved was about AI workflows for marketing and sales specifically — outbound, knowledge systems, content at scale. Not AI in general. A very specific slice of it. That's not something I would have told you upfront, but the pattern was obvious in the data.
The map that emerges from your saves is more honest than your LinkedIn bio. It reflects what you kept returning to when nobody was watching, when there was no agenda. That signal is valuable — for understanding where your thinking is actually heading, and for knowing what you have genuine things to say about.
What people get wrong about the effort
Most people look at a second brain and think: project. Six weeks of setup, tags and folders and hierarchies, then maintenance. That framing kills it before it starts.
The value isn't in the setup. It's in the compounding.
Month 1 feels like nothing. Month 6 feels like a competitive advantage.
At month one, it feels thin. You have 40 bookmarks and a few Drive notes and you're not sure what the point is. At month three, patterns start showing up — you can see what topics you keep returning to, what ideas are in tension with each other, where your thinking has moved. At month six, you have something that nobody can replicate quickly: a proprietary index of your own perspective on a domain, built up one save at a time.
That's not something you can Google your way into. It's yours. And in the age of AI, that's the interesting part.
What the age of AI changes about this
A second brain used to just be personal. A way to not forget things. That's still true. But now the data inside it is actually deployable.
Run your curated knowledge through a model and you can generate content briefs in minutes. Surface the three posts that relate to a topic a client asked about. Pull the pattern from six months of saves and turn it into a guide. Your second brain becomes a context layer — not just for you, but for your team, your clients, your workflows.
I'm already using it this way. Every blog post I write starts from something I saved, annotated with why I found it interesting. The writing isn't research — it's synthesis of things I've already thought about. That's a completely different starting point.
The longer-term play is transferability. Build your second brain, then make it accessible to your team. The curation you've done — the POV embedded in what you chose to save and what you wrote about it — that's training data for how your team thinks. You can give someone context that took you six months to develop, in a format they can actually query.
In the age of AI, proprietary data is the moat. Most people are building it accidentally. The ones who are intentional about it are six months ahead before the conversation even starts.
Where I'm taking it next
Right now my system is two layers: Raindrop for capture, Drive for synthesis. The next layer I'm building is retrieval — a way to query my own second brain in plain language and surface the relevant context instantly. Not a search bar. A conversation with six months of my own thinking.
That's a knowledge base workflow — one of the services I build for clients. I'm just building it for myself first, because that's the only way to know if it actually works.
If you're already saving things everywhere and going back to none of it — the fix isn't a better tool. It's a weekly 30 minutes to write down why you saved it. Start there. The system grows around that habit.
The Workflow — Newsletter
Operator notes on AI workflows, GTM, and what I'm actually building. Sourced directly from this second brain, weekly.
Subscribe on Substack →