Iron Brew Works

I have a confession: I collect information the way some people collect junk mail. I don’t do it on purpose. I just see something interesting and think, I’ll need this later, and then it disappears into a folder or a doc or a tab I never close, never to surface again.

I used to do this with physical stuff. Newspaper clippings, printouts, torn magazine pages. Eventually I had a tub of them. The tub didn’t contain information – it contained the ghost of information I once thought was important.

The digital version is harder to see, which makes it easier to accumulate. Google Docs is the primary offender. I have hundreds of them. Some have titles. Most don’t. A document called “Notes – Research – Q3” tells you nothing six months later except that at some point I was optimistic.

For years I’ve managed my actual work with a loose version of David Allen’s Getting Things Done. The system is simple: if a task takes less than ten minutes, do it now. If it takes roughly an hour, put it on the calendar. Everything else goes into a folder called Someday/Maybe. This works. I’m not going to oversell it – it’s basically a holding pen for good intentions – but it got me out of the habit of letting things pile up on my mental desktop.

Then I started using AI research tools. Claude can synthesize a topic in twenty minutes that would have taken me two hours of reading. So I started asking more questions, exploring more rabbit holes, which meant the Someday/Maybe folder got very, very large, very quickly.

A pile you can’t surface is just hoarding. I had known this about the tub of clippings. I was slower to admit it about the docs.

So I built something. I’m hesitant to call it a system because systems imply I’ve solved the problem, and I’m not sure I have. But I sorted the pile by one variable: how certain I am about what something is for.

If I know the purpose – a client project, an article I’m committed to, a piece of work going on my private dev server – it goes to Notion. Notion connects information to other information and works well with agents. Future article ideas go into a Notion database tagged as “idea.” That’s a surface I actually revisit, not a drawer I open twice a year and feel bad about.

If I don’t know the purpose yet, but I want to stay current – blogs, newsletters, the daily reading I try to do for about an hour every morning – that goes to Reader, from Readwise. I set up a Gmail filter so newsletters skip my inbox and go directly to Reader. Reader is a beautiful reading surface on my phone and iPad. I can highlight something if it matters, but I don’t have to. The whole article saves either way.

The third bucket is the strange one: pure curiosity research. Something Claude turned up that I found genuinely interesting and have absolutely no plan for. I built a “save to Reader” skill that drops it there tagged against a growing list. Six months later, I might want to write about something and I can’t remember the article – but I remember the idea. I query Reader and find it.

That’s the magic. Not the tools. The tools are fine. The magic is that the pile is finally findable.

A Someday/Maybe folder you can query is no longer hoarding. It’s a library. The goal was never to stop collecting – I’m a learner, collecting is what I do – it was to stop drowning in the things I’d collected.

The tub of clippings is gone. Some things you only understand you owned once you finally go looking for them.