Zettelkasten, without the graveyard
Luhmann wrote seventy books from ninety thousand notes. Most people who copy his system build an archive nobody visits.
Luhmann wrote seventy books from ninety thousand notes. Most people who copy his system build an archive nobody visits.
The cheapest list in any productivity system is the one almost nobody maintains, and it is the one that makes you look organised to everybody else.
The moment you run a second agent, the problem stops being memory and becomes agreement. That is a layout question.
A to-do list has no opinion about how long anything takes. That is why yours has eleven things on it for a day with four hours in it.
The most valuable thirty minutes in any productivity method are the ones almost nobody still does by March. The fix is structural.
Do it now if it takes two minutes. True, useful, and quietly responsible for some of the worst days you have had.
An annual goal has fifty-two weeks of tomorrow in it. Twelve weeks has none, and that single change is the whole method.
Most projects fail at the start, quietly, by never being defined. Five bullets take four minutes and prevent the failure everybody blames on execution.
Four lines, written the week it ends. Three of them are history and the fourth is the only one that changes anything.
The most useful thing an agent did for this product was hand its founder six checkboxes. Delegation runs both ways now.
Experts do not fail from ignorance. Three rules from the operating room, running as a tree you can copy.
Most agent setups fail on permissions or on amnesia. One shared branch with three parts fixes both, and you can read it.
Most people cannot drop an idea because dropping it feels like deciding it was worthless. Parking it costs one line and settles the argument.
Highlights are the most collected and least used notes in the world. The fix is one rule about which line is the parent.
Kanban has two rules. Teams adopt the board and skip the second one, which is the one that does the work.
Filing by topic fails because most notes belong to three topics. Filing by how soon you will act on it has exactly one answer.
Most personal OKRs are a wish with three sub-wishes. The repair is a rule about numbers, and it is uncomfortable on purpose.
One agent read the same tree twice and spent 962,608 bytes. Asking what changed instead cost 1,060. Same knowledge, 900 times cheaper.
Nobody re-reads meeting notes. So write the four lines that need to leave the room, and let the rest be noise you never look at again.
The same session, written as prose, is 11,222 words. Written as an escalating outline, the top layer is five lines and 60 words, and nothing is missing.
Merlin Mann's phrase got attached to the wrong noun for twenty years, and the misreading is why people burn evenings not achieving it.
Streak apps are designed to make you feel good. A five-column table is designed to show you the week you missed four days.
An agent reading prose has to guess what the fields are. The same facts as a table cost fewer tokens and admit no guessing.
Two minutes of setup gives an agent a place where Tuesday's decisions survive until Friday, readable next to your own notes.
The fastest capture tool ever built is already on every device you own, and it can write straight into one branch of your outline.
GTD does not die because people stop caring. It dies at capture and at the weekly review. A tree with checkboxes fixes both.
Three of the four boxes fill themselves. The fourth decides your year, and it is empty in almost every matrix ever drawn.
Doing the hard thing first is good advice that fails at 08:45, because by then you are the person who has to pick it.
Eight changes shipped and deployed while the person who ordered them was out of the house. He typed one line. Nobody re-read a transcript.