The Djed is an ancient Egyptian symbol of stability, a pillar that holds things up.
This account is about building that for your life: a personal AI system that knows your goals and helps you take one useful step a day.
Ideas from a book in progress.
@henruinfo Nice that you measured each step. I'd add one more source of aliases: every time a real question misses, save the words you actually used into that fact's header. Your own misses make the most honest test set.
A daily review can be short. Paste your notes into your AI assistant and ask: "Using only what I pasted, where do things stand? Suggest one step I can finish today within my limits. Label each line as fact or proposal."
About 3 to 5 minutes is enough.
@paulopierrondi I ask it about something I know changed recently and see which version it quotes. Dating every note helps too: when two disagree, the newest wins, and it has to tell me which one it used.
@mrr_nunya Plain files for me. Every rule carries the date and why it exists, and when a newer one replaces it, the old line gets marked superseded instead of quietly staying in. Forgetting has to be on purpose too.
Your morning starts with the sleep you actually had and the hours already spoken for.
Begin from those facts. The start of the day doesn't have to be a verdict on you.
@boscoshubin I split when two jobs need different notes. If one agent has to carry both the money side and the writing side, its memory gets muddy and it starts mixing them up.
@j0m1st I don't really reset, I hand off. Each new piece of work starts fresh with a short note: what was decided, why, and what's next. Everything else can go.
@rswebdesigner Not really, because the two never share a notes file. If something matters to both, I copy it across myself, on purpose, so nothing drifts over by accident.
You don't need to have it all worked out. You can begin with the situation in front of you.
Write down what is happening now, and one thing that is yours to do today. That is enough to start from.
@CodeTiwg One short file per repo, but only the things it can't work out from the code: decisions, the reason behind them, and what not to do again. The rest it can read fresh each time.
During a job search, every day can start to feel like a test.
Let today be smaller: one useful step, picked for the energy you actually have. Tomorrow can carry the rest.
Money in hand and money you're hoping for belong on separate lines.
An offer is not a payment. An invoice is not money received.
Writing down which is which can make a tight month feel calmer.
@KrrishKapoor8 I keep decisions with their reason and date, and let day-to-day status notes go after a week or so. When a newer decision replaces an old one, I mark the old one as replaced instead of keeping both. That's most of the cleanup.
@vansh_txt Not under a rock. I let it suggest options, but the choices that are hard to undo later, like the schema, I make myself and write down with the reason. Then it builds inside those.
@PrakharTandon29 The redundant step comes from the brand guide living inside ChatGPT's memory. Ask GPT once to write everything it knows about your brand into one plain file, check it yourself, then hand that same file to Gemini or Claude. The file becomes the source, not one app.
Found a mistake in your notes? Fix the line, then add a short dated note saying what was wrong and why. Later you can still see what changed, and when.
@UnaHub8 I draw it at anything that can't be undone. Drafts, research and sorting it can just do. Anything that sends, pays or deletes waits for my yes.
@adityaa_behera Same here, one shared notes file. My one rule: only one of them may change the decisions in it, the other reads and suggests. With two writers it drifts fast.