The palest ink is better than the best memory

Your mind is for having ideas, not holding them.

It started with Obsidian. First time: "this is it, knowledge graphs, pure markdown, local storage, no subscription, I'm going to become the kind of person who manages their knowledge" — lasted three weeks. Second time, different folder structure, better this time, I even had a separate vault called "deep ideas" which is the most pretentious thing I've ever done and I stand by it — lasted less. Third time, plugins, Dataview queries, a template for every note type, a different accent color as if that was going to rewire my brain somehow — lasted long enough to realize the problem wasn't Obsidian.

Then Notion, which is basically Obsidian for people who need rounded corners and a vegetable emoji next to everything to feel like they're being productive, and that worked fine until I opened the mobile app and it looked like a completely different product designed by a different company in a different country, and also no native calendar, which at the time was a dealbreaker because I was in a phase where I believed organizing time and organizing knowledge were the same problem, which they are not, but that lesson cost me four months and a Notion Pro subscription.

Tried three or four paid apps, won't name them because they don't deserve the SEO, all of them promised "your ultimate second brain" on the landing page, all of them had a 45-step onboarding, all of them died quietly in my subscriptions list next to two VPNs I also don't use.

Then Anki, which is different because Anki doesn't try to be pretty, Anki is like that math teacher in high school who smelled like stale coffee and never smiled but actually taught you how to think, and I thought: this is for retention, this is the real problem, and so naturally the next logical step for someone like me was to write a Python parser that converts a CSV with columns front|back|tag into an importable Anki deck, automating the whole pipeline, because I had been accumulating for months — years, actually — across loose text files, forgotten voice memos, untitled screenshots, a notes app graveyard on my phone, conceptual napkins: linguistic expressions I liked, 3am ideas with no follow-up, vague plans with ambitious names, half-understood philosophical concepts, personal contradictions I never resolved but wanted to remember existed, feelings with full contextual metadata ("felt this on a Tuesday when it was raining and a stranger on the subway gave me their seat without making eye contact and for some reason that felt like the most human thing I'd witnessed in weeks"), opinions I'd changed and wanted to track, experiences that if I don't write down in the moment are gone forever like they never happened, and I thought Anki was the final destination for all of it — until I found out, obviously, that Anki cards work better when you make them yourself, not when you batch-import them from a script, because part of the learning happens in the act of formulating the question, of deciding what matters and in what form.

And that's when the crisis started — because if I learn better when I process information manually then what even is learning, what is retention, what is knowledge if I can't tell the difference between having lived something and having filed it correctly. That crisis produced more material which I started posting on Instagram for a while, which worked until I got bored of the format because Instagram is not designed for long thoughts or for people who want to search something they wrote eighteen months ago. Meanwhile the Drive database — which started as a journal Google Doc and became a folder with subfolders with subdocuments with sections referencing other documents in other subfolders — had grown so large that opening it felt like walking into a room you haven't cleaned in years: you know there's valuable stuff in there but the cognitive cost of finding it exceeds the benefit. And so now I need a new pipeline, again, from scratch, like always.

Part of this I brought from home. I mean literally: there's a pattern of accumulation, of not wanting to let go, of feeling like if I don't record something it's as if it never happened, and I'm currently in therapy specifically to figure out whether this compulsion to capture everything is a universal trait — that anxiety about time passing and erasing — or whether in my case it has deeper, more structural roots, because there's a real difference between "everyone feels this a little" and "this is costing you hours every week and you still don't have a system that works." Philosophy doesn't solve this but it does contextualize it: there's something deeply human about wanting to externalize memory, from cave paintings to Anne Frank's diary to an Obsidian vault with 847 unlinked notes, we're all trying to say I was here, I thought this, I felt that, and make sure it doesn't vanish when the brain decides it's no longer a priority. The issue is that forgetting is a feature, not a bug — the brain is selective by design — and we're trying to hack that with folder hierarchies.

If I could go back in time and find myself — and I'd find me in the bathroom, genuinely, that's where half my good ideas happen, standing there like an idiot holding a toothbrush — I wouldn't give some wise speech. I'd just say: open whatever system you find first that seems vaguely reasonable and start putting things in it. Right now. Today. Because the perfect system doesn't exist and it never will, and the messy, broken, half-indexed database I have today — with its broken references and notes that just say "look into this later" — is more valuable than any perfectly architected system that's empty. And if you restructure it all six months later — different hierarchy, different categories, different logic — that's not starting over. That's refining. A messy archive is a living archive.

We're entering a period where AI can do real things for you if it knows who you are. The difference between an AI that knows you and one that doesn't is the difference between a doctor who's treated you for ten years and an ER doctor seeing you for the first time — both competent, but one has context. The distance between useful AI and transformative AI is the amount of personal context you can feed it. When you give a model your notes, your accumulated thought patterns, it stops responding like you're the average of every user it's ever seen (which is always mid) and starts responding as if it knows you specifically. That's when the archive starts paying dividends. The competitive advantage of the next decade won't be what you know — it'll be what you have documented. All that messy material suddenly has a use you hadn't planned for.

My therapist keeps telling me to stop looking backward — present-focused, mindfulness, the whole framework — and she's not wrong: the past can become a trap, a place you go to replay things you can't change and confirm stories you already believe. But there's something we keep going back and forth on, which is that looking backward and ruminating backward aren't the same thing. Rumination is the loop you can't stop — involuntary, repetitive, 3am stuff. Consulting your past is something else: it's going to your own database and asking "last time I made a decision like this, how did it go?" and getting an answer based on what you actually wrote at the time, not what your current mood wants you to remember. That's a different kind of looking backward. That's useful.

So collect everything. Sort later. Sort never. It doesn't matter. Collect the expressions you like and the plans you'll never execute and the feelings with full context and the contradictions you don't understand and the opinions you haven't formed yet. Put it somewhere, anyhow, with the most imperfect system in the world. Because the moment you find something that works — even temporarily, even if you migrate it somewhere else in two years — you're going to want the material. And the material is you.