There’s a few high performance, super-powered, motivated beings who manage to keep an organised list of tasks and work through it every day. Same people who clear their mail and hit inbox zero daily.

Most people, I would wager, aren’t like that. Most of us are messy, lazy, procrastination loving sloths with motivated bursts of productivity few and far between.

I’m one of the sloths.

I’ve started and stopped task management systems way too often. The list starts out small and manageable, then grows into an untameable monstrosity, and at some point I stop looking at it, because it hurts too much to see what’s still open.

And when I do have the energy to look, I realise I’ve completed most of it already. Through my work sessions.

I’m not the only one. I asked my Obsidian community how they manage their tasks, and almost everyone described the same thing: the system works for a few weeks, maybe months... but eventually it turns into a mess. The tasks just keep sliding further down a never-ending pipeline.

Every task system I’ve tried had the same hidden requirement: me.

My discipline to open the list.
My willpower to look at it.
My energy to clean it up.

The list never did any of the work. It just sat there and waited for me to show up.

The task list or the coach?

If you had the choice between managing your own task list and having a personal coach who manages it for you, which would you pick?

I’d pick the coach.

It becomes the coach’s responsibility to make sure the important things get done. The coach checks in on me every day, makes sure the list is moving, and tells me off when it isn’t. I’m effectively offloading my willpower onto someone (or something) else.

Say doing my taxes is on the list. On my own, I look at it, feel lazy, and skip it. Tomorrow it’s still there, a little heavier, and I’m a little less likely to open the list at all.

With a coach, hiding gets harder. The coach says: do your taxes today. I say: not today, I’m feeling lazy. The coach’s job is to ask why. Is it my energy? Is the task too vague? Is it just hard? Once the real reason is squeezed out, the coach can help me approach the same task differently tomorrow.

Maybe the coach breaks it down to a smaller step that feels manageable. Maybe it just tells me to stop being a lazy little fuck and do it.

I like that coach. Other people might want a gentler one. Either way, the persona is yours to pick.

The task list depends on your discipline. The coach lets you outsource it.

Billionaires have this for every area of their life. A personal trainer for their body. A nutritionist for their diet. And a PA for their tasks. The task list still exists. It just isn’t theirs to maintain.

That is what I want from a task system. Not another burden. A (smart) assistant.

And for the first time, you don’t need fuck-you money to have one of your own. You can build one with the power of ✨AI✨. Obsidian holds the knowledge, something like Claude Code sits on top of it and does the remembering.

I’m not saying it’s effortless to set up. An assistant needs two things from you before they can do any work.

First, a map of your life: which areas, responsibilities and projects exist, so they know what they’re managing.

Second, a place to look: the work session, where you leave a record of what you did and decided, so they know where you left off.

Give the assistant a map

Picture a real assistant on their first day. You hand them a pile of sticky notes: “set up call with accountant”, “buy muesli”, “call Jeffrey”, “get income statement”. They don’t have context yet, so they probably won’t do much of a good job. They’ll call the wrong Jeffrey, and suddenly you’re on a list you really don’t want to be on.

But over time, they get to know you, your projects, your life, and before you realise it they feel like a sword to a knight. An extension of you.

What a good assistant needs is a map. Which parts of your life exist, what’s ongoing, and what has an end.

What I’ve come to discover is most of our lives are broken up into “Areas”. And within those areas you might have some ongoing “Hubs” of responsibilities with “Projects” which have an end or completion condition.

I start with 7 Areas: Health, Wealth, Career, Social, Leisure, Personal Development, and Environment. The Life Mapping Workbook I use with clients is free inside my community, The Construct, once you sign up.

Hubs fall into areas. They’re the ongoing responsibilities of your life, the ones without a completion condition or end date. Taxes (Hub) and Savings (Hub) lie in Wealth (Area). Guitar (Hub) and Piano (Hub) lie in Leisure (Area). Running (Hub) and Gym (Hub) fall into Health (Area).

Then it’s as easy as sitting down with each one and asking yourself what projects you have in each of these buckets.

You can do this exercise from a top-down or a bottom-up approach. Some like to think about the projects first. Other clients prefer sitting down with the 7 area buckets and unravelling the hubs and projects from there.

To give the most boring example of this, take something each of us has to do eventually in our lives. Taxes.

• Wealth (Area)

  • Taxes (Hub)

    • Project - Taxes 2026

For the project you have a “completion” condition: submitting your 2026 tax application to the government. That’s the first thing an assistant wants to know: what does done look like?

How do we move “Project - Taxes 2026” forward to that completion condition?

Certain tasks immediately pop up. Here are some that come to mind:

• Find an accountant

• Get government links

• Retrieve expenses that are tax deductible

• Get income statement

• Submit documents

Now the sticky notes have a home. Every one of these belongs to a project, the project belongs to a hub, the hub belongs to an area. The assistant can double click into it all to make sense of your task list.

Normally this is where the manual work starts. You write each of those as a task, you review the list, you decide what’s next, you clean out the stale ones. Some people get a real dopamine rush from this. The item gets checked, it disappears, they feel organised. I have a video on exactly that setup: Obsidian Task Management - Dataview vs TaskNotes (Full Setup).

With the map in place, that still works. Every task routes to its project, so the list lives where the work lives. And because the assistant understands the structure, they can see those tasks too. You and your assistant are both looking at the same list.

So you get to choose. Go check things off yourself if you want the rush, and the assistant picks up that it’s done. Or hand the full maintenance over and never open the list again. And if you really need the confetti and cake, we can have your screen light up with “FUCK YEAH, ANOTHER ONE DONE” and a crowd screaming and cheering in the background. Absurd, but doable.

Either way, the list is no longer the thing you have to maintain. You keep a record of the work. The assistant keeps the list.

The work session way

A work session is a note you open every time you sit down to work on a project. You create your first one the day you start: “2026-09-24 Work session Braindump on 2026 taxes”.

You word vomit everything out on paper. All the things you need to do. You have a conversation with your favourite AI, and then you write down the rough list of tasks in the order you need to do them. This is the briefing. It’s the conversation you’d have with an assistant on the first day of a new project.

The things you want to capture here are your thinking process and the actual decisions you made. Your thinking process is more important than you realise.

We think we just want the end result. The cake. But we neglect to write down the recipe we followed to get it, the trial and error, and that’s what we’re losing out on. The work session is where the recipe gets written down, so the next cake is a little bit easier to make.

Why do I prefer the work session way of working?

Because you think in projects, and the assistant does the remembering.

You don’t need to create a specific task. You don’t need it to pop up anywhere. You don’t need to review a tasks page.

Next time you sit down to work on the project, you open up a new work session, “2026-09-25 Work session on 2026 taxes”, and the assistant (Claude, or your AI companion of choice) goes to the project it’s attached to, reads the previous work sessions, sees the tasks you highlighted to do in order, sees where you left off, then tells you exactly the goal of this work session.

This is where tasks become easier. You just ask the assistant what the highest priority is. It goes through your projects, work sessions and task lists. It knows what else you’re busy with outside those projects. It can check your calendar to see how much time you have, and read your morning reflection to see what mental state you’re starting the day in and what you’re stressed about. It has the context of yesterday’s work. Based on all of that, it pinpoints the most valuable task for today.

With a normal task list, according to grandaddy OG David Allen, you need to do a weekly review to keep the list up to date and clean. Now your weekly review is not a date with your task list, but a date with your assistant. This is where you correct and teach it. The first couple of weeks will require some input from you, just as it would with a real person, but eventually the assistant will get to know you and start running perfectly in the background.

One honest caveat, because I get asked this a lot: can you trust the AI to read five old work sessions and not get confused by what you used to think?

Only if you write down what you are busy with in the work sessions. If the work session is blank, the assistant has nothing to go on. If it shows the goals for the next 30 minutes, the process you followed, links you used, artefacts created, what you completed, and what you need to do next, that is when the assistant won’t get confused.

Don’t expect it to read your mind.

When you make a decision, write it down. When the project changes direction, note it. The assistant weighs the recent sessions heavier than the old ones, and if it’s unsure, it asks. But if the change of mind stays in your head, the assistant is reading a stale file. That’s not an AI problem. That’s the same problem a human assistant would have.

No task list to maintain, no weekly clean-up, no having to think what’s next. Just an uninterrupted trail of work sessions till the project’s status turns from “in-progress” to “done”. Exactly how the billionaire does it. They never maintain the list. They have people for that.

Before you hire one

Most people skip the setup phase (Areas, Hubs, Projects) and wonder why their system feels like a mess. They get so excited by the prospect of “AI” running their life for them that they skip the structuring completely. They hire the assistant before they have anything to hand over, and then blame the assistant.

So here are the steps:

1. Map out your life: areas, then hubs, then projects.

2. Put the map into your knowledge management tool of choice.

3. Work in projects and work sessions until it feels natural. No AI yet.

4. Hire the assistant for maintenance: it reads your work sessions, tells you where you left off, and keeps your projects current.

5. Hire the assistant for automation: the recurring stuff that runs without you.

Steps 4 and 5 are less work than they sound. Once your knowledge system is set up the right way, onboarding the assistant is as easy as telling Claude what kind of assistant you want: the persona, the check-ins, how hard it should push.

And if you already work in projects and work sessions, you have an assistant available today. Open Claude Code in your vault and ask:

“I’m working on Project - Taxes 2026. What’s next?”

It reads the project, goes through your work sessions, and tells you where you left off and what to do next. No task list needed.

None of this turned me into one of those inbox-zero people. I’m still one of the sloths. I still skip the taxes on a lazy day. But the list no longer sits there waiting for me to abandon it.

Instead of yet another burden, I now have an assistant on the payroll.

This article is the theory part, to get your buy-in. The next one is on how to actually “hire” the assistant.

Make sure to subscribe to see the follow-ups.

Subscribe now

Join the community

If you want the Life Mapping Workbook, and a place to ask anything Obsidian and Claude Code related, join The Construct. It’s free.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter | Free Community

That was the problem a recent client came to me with. CEO of a small tech startup, liked the Obsidian way of thinking, drowning in volume. Meetings, articles, email, ideas: all piling up quicker than he could process them.

Subscribe now

Here is what we set up. Every step is something you can start on your own this week, so treat it as a menu.

1. Think in projects before you touch any tools

We spent two full sessions getting everything in his career area out on paper, then building the structure in Obsidian: areas, hubs, projects. Nothing else until the containers existed.

For you: write down every open loop in one life area. Group them. Each group with an end state is a project, each one without is a hub. Filing this year’s taxes is a project. Health is a hub.

The Life Mapping Workbook walks you through this properly. It is free inside The Construct, my community, so join and grab it.

2. Practice input to output

Once the containers were there we spent a session doing nothing but running information through the system by hand: daily note, log entry, project note, work session or meeting note, artefacts, outputs.

And then we repeat that MULTIPLE times... Repetition is how the habit gets built.

This is the step most people feel they don’t need a whole session for. It’s also where people waste the most time (ironically, trying to save time and money), because they skip it and then wonder why the system breaks down.

A couple of weeks later I ask to see the daily note and it’s as empty as my willpower bank when there are Oreos in the cupboard.

For you: track something on your daily note every day for the next week. Pick one:

• A gym log with a workout note

• The day rating and a one-line summary

• 3 pictures from the day

• A meeting log with its meeting note

• What you ate and the recipes you find on insta or the You of Tubes

The goal is a tool you actually come back to.

3. Automate the first stream, meetings

He has a lot of meetings, so that was the first automation. We got him to start recording and pasting the transcript into a meeting note. Always start with the simplest solution and automate more later. Then I gave him a meeting processing skill in Claude Code and adapted it to his way of working: it pulls out the takeaways, action items, and drafts the follow-up.

For you: record your next meeting (I use Loom, others use Granola), paste the transcript into a note, and ask Claude to summarise it and list the actions. Knowing you don’t have to write down every detail whilst in a meeting lets you be present. Once you’ve connected your mail, the follow-up emails can be automated too.

4. Automate the second stream, sources

He processes a lot of articles and videos, so we built a source processing skill and set up the knowledge bases and hubs for the topics he keeps coming back to. Paste a URL, Claude fetches the content, writes the summary and takeaways, and files it under the right hub.

For you: next time you save an article, have Claude create the source note and route it to a hub. For example, I’m into magic... Think vanishing coins and card tricks. When I come across a trick I want to learn, I paste the URL of the video and Claude files it. It pops up in my “Magic (Knowledge Base)” - Tadaaaaaaaaaa! ✨🪄.

5. Automate the third stream, email

Like any founder, he has a lot of email back and forth. We connected his Gmail to Claude with a script that sorts and summarises his inbox every morning. He reads one summary, dictates his replies with WisprFlow, and tells it what to put in his calendar.

For you: this one needs a bit more setup, but even a manual “summarise my inbox from today” in Claude Code with your Gmail connected is a good first step.

6. A walkie talkie to your vault

The final step was a computer that stays on 24/7 at home. Easiest (at least for me and my clients) is a Mac Mini, but any old laptop works. Obsidian Sync, a Telegram bot, and Claude Code on it, and he now has a walkie talkie to an active Claude chat that can do everything he can do from his desk:

• Make changes in Obsidian (daily note, projects, people, meetings)

• Email, calendar, meeting invites, summaries

• Source notes

• Quick notes and stray ideas from the phone

• Ask it to go do some interwebs research (”hey, what movies are playing tonight?”, “what is the weather tomorrow?”, “for Project X, create a new work session and compare Company X and Company Y”)

For you, the skeleton is simple: a spare machine that never sleeps, Obsidian Sync so the vault is on both, Claude Code running in a terminal there, and a Telegram bot that pipes your messages into that session and the replies back. The devil is in keeping it alive through updates, reboots, and lost Wi-Fi. That part took me weeks, and it’s what the guide covers.

Stray ideas from the phone are where it pays off most. Ideas don’t need complicated automations, they need zero friction at the moment they show up. I dictate most of mine with WisprFlow, which works on the phone too.

The order matters more than the tools. Containers first, then the inputs by hand, then automate the friction points, then put a remote on all of it.

Which of these are you already doing, and what else would you add to your system?

Join the community

If you want the Life Mapping Workbook, and a place to ask anything Obsidian and Claude Code related, join The Construct. It’s free. The step-by-step setup guides for meetings, sources, email, and the Telegram walkie talkie live in the paid tier.

Subscribe now

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter | Free Community

Then suddenly it snaps into place. You know exactly where you are going and which turns to take. Your foot gets heavy now that you know where you are accelerating to.

That is your brain doing a background task. Give it a destination, let it simmer for a moment, and it hands you the map.

But that only works for the tasks your brain is allowed to simmer on. The painful ones never get that far. Your brain pulls them out of view before the problem-solving can even start.

That’s not a flaw of yours. It’s the same brain doing a different job. One moment it’s loading the map for you, the next it’s shielding you from something that hurts. And the hurt is closer to literal than you’d think: fMRI studies found that in people who dread math, just anticipating a math task lights up the same brain regions as physical pain. Not doing the math. Just thinking about doing it.

Starting hurts more than doing. So your brain protects you the only way it knows how: it keeps you from starting.

Subscribe now

Piers Steel, a guy who studied motivation for a living, boiled this down to a neat formula with four variables: Expectancy times Value, divided by Impulsiveness times Delay.

Unsure you can pull it off? Expectancy drops. No enjoyment in it? Value drops. Have a TV on in the background? Suddenly Judge Judy is the most exciting show ever. Deadline far away? Delay eats the whole fraction.

Your brain runs this math before you’re even aware of it, and when the number comes out low, the hide and seek starts.

Here is my solution to getting you and your brain on the same side: thinking in projects. It attacks every variable in that equation at once - I’ve broken down exactly how in a previous video - but this article is more about the mechanism than the theory.

When you create a project note in Obsidian, you’re generating the map for the questions you’re about to ask. Same as that car ride. The vague, painful blob becomes a destination.

Take something we all need to do at some point in our lives: taxes. The moment you have a project called “Project - 2026 Taxes”, and you say almost out loud that you’re working on it, your brain starts pulling in all the steps you need to take.

So we have the broad project. Now we add a couple of layers.

We log it on our daily note. That gives us the chronological reference of when and how we moved this project forward.

This comes in the form of a daily log: a bullet that says “I’m working on Project - 2026 Taxes. I have 30 minutes in which I’m getting all the links I need.”

Then you create a dedicated work session note for the work you’re attempting.

You go in and start brain dumping all the loose and stray thoughts in your head. Thoughts your brain has already been working on in the background, just because you created this note. You prompted it.

In this work session you write what you did and how: the links you used, the steps you took.

The log and the work session get rolled up into the project automatically (via ✨Obsidian magic✨).

Let’s take a step back and see what we have:

1. A project note that acts as the destination, where you can see all your past efforts and every resource related to it.

2. A daily log and work session note showing your goal, what you did, and more importantly, how you did it. This is where the re-usable information gets generated. Think “valuable tax links Netherlands”, the kind of knowledge that gets re-used in the future.

3. And the part you’ll only feel in a year: future you on their knees praising you, because “Project - 2026 Taxes” just made “Project - 2027 Taxes” easy peasy.

Most people do the equivalent of taking a stack of papers, doing the work on it until the project is done, then proceeding to force feed their bin with all their hard work.

Next year they start from zero... again.

That seems very estúpido. At least to me. (I’m writing this from an island in Spain, the vocabulary is rubbing off.)

Why do you hate your future self so much that you’d inflict the same pain on them?

Are we expecting them to be masochists?

To bring this article home, let’s go back to the loading screen your brain sometimes gets when starting a task.

It’s not always 2 minutes. For the difficult things in your life, the loading takes longer, and the more painful a project feels, the more effective your brain is at hiding it from you instead of solving it. Your brain needs a place to anchor before the simmering can start.

That anchor doesn’t have to be impressive. The smallest step is opening a work session on a project and writing down what you want to achieve in the next 25 minutes. Every variable in the equation gets a nudge in the right direction.

Get moving. Your brain will catch up. Pinky promise.

If you’d rather skip the building part

If you want this structure set up for you, projects, work sessions and daily logs already wired together, that is exactly what the Evergreen Vault is.

And if you want something more, a system built around how you actually work, book a call: https://www.constructbydee.com

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter | Free Community

It doesn’t matter how pretty your dashboard is, if you never open it up it is useless.

My entire finance system is two files.

One is an Excel table where each row is a single expense or income event. The other is an HTML file that reads it and turns it into a full dashboard: monthly views, charts, cash flow, drill-downs - whatever my twisted mind wants to see.

No internet, no apps, no subscription, no account, and nothing deciding for you how you should look at your own money.

What to expect

The thesis of this article is simple.

We need to simplify in order to do something about our finances. Get all your finances into one place and keep track of it on a weekly basis.

We’ll go through why manual is sometimes better, how to set up the two files, how to mould the dashboard into your perfect finance app by talking to Claude, and how to automate it later without losing the awareness.

My goal for this article is for you to take your finances by the choke hold and do something about it.

Download

It is a free download, it is up and running in a minute, and you can reshape any part of it by pointing Claude at it and verbalising your genius instructions towards your perfect finance app.

Download it here: https://www.constructbydee.com/Finance-Tracker-Download

Open the HTML and point it to the XLSX file.

How an ice-cream date paid me 510.50 EUR

We went out for pistachio and hazelnut gelato on one of the hottest days here in Amsterdam, and before we’d left the counter she was logging the expense on her phone.

Manually.
In the age of AI.

She does this for every transaction, and has for years.

I started questioning her.
“Don’t you know there are automated solutions?”
“Doesn’t your bank have expense tracking features?”
“How long have you been frozen before waking up in 2026?”

She said that the fancy banking features didn’t work for her. Her money is spread over multiple accounts, and no app gave her the simplistic holistic view she wanted.

Somewhere between the questions I realised she probably understands her money faaaar better than I ever did with my automated setups.

So I implemented her solution for myself - and my gawd did I immediately get my money back.

After two weeks of logging my finances daily:

• A jiujitsu gym membership I signed up for as a trial class and forgot about for 6 months. One email later they refunded everything - 360 EUR

• A Strava subscription about to renew that I never used - 12.50 EUR

• A Zoom subscription I barely touched, cancelled - 18 EUR

• Meals I’d fronted for friends at restaurants, payment requests finally sent - 120 EUR

Total savings: 510.50 EUR

The manual step is not the weakness of her system. It is the whole point. Logging the expense is the moment of awareness, and awareness is the entire product. Automation removes the one behaviour that actually fixes your finances.

So the system I built starts manual, stays mouldable, and automates only the parts that don’t cost you that awareness.

What about other apps that can do it for me?

We aren’t technophobes.

It’s not like we are scared of fancy apps and AI finance trackers that auto assemble our finances across all our different bank and investment accounts. Your bank probably has some advanced dashboards too: trends, cool graphs, spider diagrams, AI chat functionality on top of it.

Here are two questions I want you to answer:

When last did you look at it?
When last did you change your behaviour because of it?

Me? I briefly look at my expenses maybe once every 6 months, or when I’m desperately low on cash and looking for subscriptions to decapitate.

One statistic I would never want to see is how much money I’ve pissed into the wind because of me just not being aware what was happening on a weekly basis in my finances.

The problem isn’t the tools. It’s two design decisions baked into all of them.

The first we’ve already seen: automation removes us - the decision maker - from the process, which results in no change in behaviour.

The second: they are designed by people who have put a lot of thought into them, for sure. But they are designed for the masses. For the average consumer.

Designing for average

There is a story about this we can learn from: the 1950s U.S. aircraft crash crisis.

In the late 1940s and early 1950s, the U.S. Air Force suffered an absurdly high crash rate, at its peak 17 non-combat crashes in a single day. Technology kept improving across those years, and still the crashes kept happening.

It took years to find the cause (I can’t help but think the people they most wanted to interview weren’t picking up the phone anymore). They ruled out mechanical failure, pilot error, poor training.

The cockpits were built around “average pilot” dimensions from the 1920s, and the leading hypothesis was that pilots simply drank more of their milkies and were bigger now than when the data were collected. A lieutenant named Gilbert Daniels measured 4,000 airmen to check.

Growth was not the culprit. What he found instead: not a single man fit the definition of an average pilot.

The cockpit was designed for the average man, and the average man did not exist. Designing for the average means designing for nobody.

The fix came in the form of adapting the cockpit to the pilot: adjustable seats, adjustable pedals, adjustable helmets. Not the other way around.

So why are we accepting static cockpits for everything else in our lives?

Your finance tracking should work the same way. It should adapt to how you want to work with it, not be dictated by the view your bank built for the average man.

You are not average.

You are ✨special✨.

So how do we optimise for opening your finance tracker?

Counterintuitively we start by including a little bit of friction. Remove some of the automation (not forever - just for the beginning). And most importantly, build something we can change.

How my table (finance database) is structured

It is one table with a couple of key columns. Every row an expense.

The minimum you need for this to work is Date, Amount, Counterparty, Description, and Tag.

I add Bank (because I have multiple accounts), Area (because I like splitting my finances by my life areas), and Notes (for anything I want to remember about a particular line).

I’m not spending too much time on this, because I keep it as simple as I can.

Are there better ways to structure this data? For sure. Do we need them? Probably not.

What is in my HTML dashboard?

Here are a few pictures to give you ideas, or go watch the YouTube video to see it in action:

Describing my graphs is not the point of this article. Go play around if you are interested, use it, change what you don’t like, and make it your own.

Updating the app is easier than you might expect

Here is what makes this different from every tracker you have downloaded and abandoned. The dashboard is one HTML file somewhere on your computer. Point Claude at that file and you can update it... however you want babbby.

Want a pink Barbie filter? You got it queen.
Want a spider diagram of your consumption across months? Yes Peter sir.
Want drill-down capabilities? Hell yeah sergeant shawty!

A real example of something that irritated me: after stacking a few filters there was no quick way back to the default view. So I opened Claude, pointed it at the HTML file, and said I wanted a clear-all-filters button on the transactions page. It read the file, made the change, I reloaded, and the button was there. I asked for it to stand out more. It highlighted it in red. Total time: 2 minutes.

You don’t even need to know what you want. Ask Claude what it recommends. It has seen more interfaces than any of us, so use that. On a whim I asked for a third theme next to light and dark, something like a calming forest green. Now my horrible expense months come wrapped in a colour that keeps me calm.

The product here is not my dashboard. It is that the tracker moulds to you. Found a feature you like in another app? Describe it, give Claude a screenshot, and command it to update yours.

Do I need to capture everything manually?

No.

My friend likes to do that. I don’t.

I generally take a screenshot of my bank app at the end of the day and ask Claude to categorise it into my Excel table, and to ask me if it doesn’t know how to tag something. This allows for the awareness of where my money is going to linger a little bit longer.

How often do I need to do this?

First off, you don’t need to do anything. So instead, see it as an “I get to”. 😘

Now that we’ve manifested abundance, the actual answer:

What matters is the cadence: we do daily (at most weekly) tracking. Compiling your finances at the end of the month isn’t much better than the bank dashboards. By then you’ve forgotten the weird once-off payment, the spontaneous sauna session you’ve booked, and that bill you should have split with your friend for the dinner out.

For the first couple of weeks I’ll be doing and checking everything manually, until I can hand over more to AI. Think connecting straight to my bank accounts, and a Telegram message at the end of the day asking me for a short voice note on what happened: who I had coffee with, what tag an expense should get.

The last part will be my daily reflection tool as well. I’ll have everything coming in automated, but it still needs me to have a 10 second talk about my expenses for the day. It then takes those notes and adds them to the expense lines.

Grab it

The free download, the Excel with demo data and the HTML app, is here: https://www.constructbydee.com/Finance-Tracker-Download

You’ll have it set up in a minute... pinky promise.

Part two is coming: automated ingestion, where Claude reads the bank statements, adds the rows, and learns your tagging. At the end of the day it asks one thing: “Picked up 4 expenses today. Tagged correctly?”

Stay tuned here or on YouTube.

If this landed, my channel is where I show these systems running live: YouTube. You can also sign up to my weekly newsletter where I share more of these type of thoughts.

And if you want assistance, here’s the link to the my consulting services to help you out with any Obsidian or Claude Code project you are working on.

Subscribe now

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

I’ve thought for the longest time that convincing people to build a system for managing their life is a useless task. That it is an unteachable lesson. That people only start once whatever system they are currently using fails them so miserably that it cannot be ignored.

Something happens. Overwhelm at work, too many years heading in the wrong direction, failed studies, dropped balls in their personal life.

That is the tipping point when people start to search for something that can help solve these problems. What they find is a whole corner of the internet preaching one fix: Personal Knowledge Management, or PKM for short. My tool of choice for it is Obsidian, a free note-taking app whose one superpower is linking notes to each other. So when I say PKM, system, or Obsidian in this article, I mean the same thing.

But a person who is doing just fine at work, at their studies, in life generally, they never stop to assess if there’s a better way.

They end up asking “Do I really need one?”

And my answer is always no. You don’t NEED one.

Just like you don’t need to brush your teeth.
You don’t need to eat healthy or exercise.
You don’t need to do your taxes.

Those are easy to convince people of, because the cost shows up on its own. Your teeth rot, you get fat, you get thrown in jail.

The cost of not having a proper system for your life never shows up on its own like that.

To see it, you’d need to split into a parallel universe, live both versions of your life, and fast forward to the end to compare. The cost is invisible because you can’t run the counterfactual. Life only shows you the cost when you reach the tipping point.

And by then you already gave up your weekends, carried the low-grade stress for years, and walked a long way down a path you never actually chose.

There is one group of people who can see the cost, though. The ones who crossed over. Switch to a proper system and you become your own counterfactual: you finally have a before and an after to compare, and the invisible cost of all those years shows up in bulk thanks to the power of hindsight. Ask anyone who made the switch and you get the same sentence back: “I wish I had started years ago.” Nobody ever says the reverse.

So this article is me attempting the unteachable lesson: convincing you not to wait for something bad to happen. The tipping point is optional. Consider this the check-up you were never scheduled for, before the rot sets in.

We all run some kind of system already. A physical notebook. A Google Drive folder with “turtles all the way down” in the form of sub-folders. A calendar and a to-do app. They’re all ways of fighting entropy. And yes, a system you stick with beats a perfect one you don’t. But most of them are a bad fit. You can do better. I know because this story starts with my memory failing me, and it ends with a software upgrade for my brain.

Let’s go through the costs of omission. What you’re giving up, item by item, by not having a knowledge management system run your life alongside you.

Subscribe now

Your memory is magnificent. And it lies to you.

Your memory has three jobs. It holds things, it hands them back when you ask, and it makes connections between them.

It is unreliable at all three, and worse, it never tells you when it has failed. It doesn’t even respect you enough to show an error message. You get false confidence instead.

Everything below is downstream of that.

1. You keep re-learning things you already knew

Imagine you are thrown into a deserted forest.

The forest has a lot of holes, and as you start to explore the surrounding landscape you fall into one.

With a hurt ankle you manage to climb out. Because pain is the fastest feedback loop, you commit it to memory as “WATCH OUT FOR THIS HOLE! Hole = ouchie”.

A couple of weeks later you fall into another. And another. And another.

Treacherous terrain indeed.

You might be able to remember some of the holes, but without mapping them you are bound to relearn the same mistake more than once.

This is what happens when most of us try to learn something. The moment it lands, your mind puts up a sign next to it: “This is important. Pay attention.” What nobody tells you is that the sign is written in non-permanent marker. You walk away certain it will hold, and the rain is already on its way.

A map is different. You circle the hole once, in ink, and it stays circled. And before exploring in any direction, you can study the path without setting foot outside your tent.

Our brains are fallible.

Our memory is shit.

2. You are judging your life through your emotions

Psychologists discovered something called mood-congruent memory. In simple terms, put someone in a bad mood and the memories they surface are bad ones. Put someone in a positive mood, and they nitpick the happy memories.

Your mood is a filter of your (already fallible) archive. You don’t notice the filter. You see what the search returned with the “happy mood filter” on and assume that is the untouched, pristine archive of what actually happened.

Imagine it’s the first of January and you sit down to reflect on the past year. A week ago you got broken up with by the person you thought was the love of your life. The glasses you have on are switched to the “morose” tinted filter, and you see the past year as the most disastrous of your entire life.

Do the same review after you had a good week and you get a different year. And run it once more in a few years, when the same relationship you’re grieving has turned into something you’re ecstatically thankful for, and you get a third year. Same twelve months, three different verdicts.

None of them is a lie. But each one holds only a piece of the true story. And you have nothing to fact-check them against.

The inputs are wrong, which makes your conclusions about how to approach the upcoming year questionable. So you go blindly frolicking in the wrong direction again, only to wake up one day close to your death and wonder why you wasted your life.

3. You are tired from things you haven’t even worked on yet

This one is not your memory being wrong. It is your memory running background processes that slow everything else down.

Imagine you have a deadline at work. You also need to do your taxes.

You’ve reserved some time this evening for the taxes. And the whole time, some part of you is running the work deadline in the background. So the taxes take twice as long as they should, and the time they steal comes out of the deadline you were worrying about.

The worrying about the work made the work worse.

This is not a vague feeling. Two researchers, Masicampo and Baumeister, measured it in 2011.

They took two groups. Everyone wrote down two things they needed to get done that week and hadn’t started. Study for an exam, work on a thesis, plan a trip.

The second group did one extra thing. They picked a time and a place for when they’ll work on each of their tasks.

Then both groups were given a chapter of a novel to read and some word puzzles. Nothing to do with their personal tasks.

The undecided group lost the thread of the story and scored worse on the puzzles. They knew it, too. Afterwards they reported being distracted the whole way through by the tasks they hadn’t started. The group who had picked a time and place showed none of it.

Both groups still had undone tasks. Both were equally sure they would get them done.

The only thing that changed was that one group knew when.

So the cost was never undone tasks. The cost is that an undecided task uses up resources you need for the work in front of you, and it stays there until you tell it when.

An open loop is any task your brain has decided it is responsible for and has not been told when to do. Water the plants. Renew the credit card that expires next month. Reply to the colleague whose mail you read on your phone and never opened again. Take the jeans back before the return window closes. Book the restaurant.

Your brain doesn’t store these by weight or size either. They all take up the same resources. The small ones are the worst offenders. Too trivial to sit down and schedule, too irritating to properly forget, so they circle like vultures around a carcass.

Take a quick break from this article and get a piece of paper and a pen. Write down every open loop that comes to mind.

How many did you get?

4. You think you understand things you do not understand

You talk to people. You read articles. You watch videos. It feels like understanding is being built.

It isn’t. Most of us are building recognition, not knowledge.

Recognition is your brain going “seen it”. Knowing is being able to build the thing from nothing, first principles, out loud, to a beginner, without any notes.

Feynman got there first. If you can’t explain a hard idea in simple words, you don’t understand it yet.

Psychologist Rebecca Lawson demonstrated the gap better than I can explain it. She asked people how well they understood how a bicycle works. Most rated themselves highly. Then she handed them a pen and asked them to draw one. Where the chain goes. How the pedals connect to the wheels.

Most of them couldn’t do it. Chains drawn round the front wheel. Frames that would fold the moment someone sat on them. And the regular cyclists were no better than anyone else.

They knew what a bicycle looked like. They had never known how it worked. And thirty years of looking at bicycles had never once told them there was a difference.

That is what consumption does to a subject you enjoy. It builds a very detailed picture of what the thing looks like, but not how it works.

Saving does the same thing, which is the trap people in the knowledge management space fall into so easily. Filing an article gives you the same easy hit of familiarity (and dopamine) that reading it did, so a full folder ends up feeling like a full head.

This is the Collector’s Fallacy: the belief that collecting information is the same as acquiring knowledge.

If it were that easy I’d be the best jiujitsu athlete alive, the world’s most yummy producing chef, and the leading expert in drawing feet.

Don’t mistake the feeling of consumption for the reality of retention.

Subscribe now

What it looked like in my life

I paid all four of these taxes for years before I did anything about it.

Study.

I did fine at university. That was the problem. The marks came, so nothing ever told me anything was wrong, and I paid the price for “good enough”.

Studying meant re-reading. Open a chapter I had definitely read before, recognise every heading, understand none of it, read it again.

Falling into the same holes every semester. Nothing from the last one had been kept anywhere except in a head that trashed it as soon as the module was passed.

I was doing to my degree what Lawson’s cyclists did to bicycles. A very detailed picture of what the subject looked like, and nothing about how it worked.

And it was joyless. I never once sat down to study because I wanted to. It was something to survive, and I assumed that was just what studying was.

It took me another ten years to realise the subject matter was never what makes studying interesting. How you study is. I can now pick up the nichest corner of the driest field and find it fascinating, because my system can’t help but squeeze insights out of it and feed them into the rest of my life.

Work.

Then I tried to climb the corporate ladder. Management and banking consulting, first in South Africa, then in Amsterdam. Long hours. Work seeping into every part of my life.

I remember one project where two of my co-consultants had to take themselves to hospital over the weekend with anxiety attacks. They were back at their desks on Monday. It happened more than once. South African corporate culture did not have the concept of burnout, so we weren’t aware that was what was happening to us.

The work ran in projects of about four months. New project, new bank, new systems, new tech stack, and about three weeks to become someone who could talk about all of it credibly. Then the cycle completed all over again.

Seven years of that and I definitely got better at consulting. But re-learning concepts I knew I had mastered before, every four months, made me feel like I was going insane. Falling into the same holes, on a schedule.

And good god, I carried all those open loops from my career into every dinner and every weekend and called it commitment.

For the longest time I thought that was just what a busy, successful life felt like. That feeling scattered was evidence of doing important things.

It isn’t.

I did try to fix it. Notebooks. Complicated folder structures. Whatever note-taking or to-do app was being recommended that year. None of it was suited for the task.

Then I found Obsidian and started using it in my career, and I told nobody about it. The answer that used to be buried fifty meetings back was suddenly thirty seconds away. I could do my work in a tenth of the time. The rest of the time went into building out this side business of mine. A tenth of the time, without any of the stress of letting something slip through the cracks, because I finally had a system I trusted more than myself.

And the part that scares me now is what I almost didn’t see. The managers a couple of years ahead of me, the prize for all of this. Balding from stress. Always anxious. Never seeming to be happy. Trading their families for a slightly more glorified cubicle at a bank.

Direction.

I was stuck in a job I hated for seven years. And in a relationship I knew wasn’t going to work out for about the same time.

Obsidian had fixed my work. I didn’t expect it to fix anything else.

Nothing clever about how I used it either. I took daily notes, and at the end of the day I’d write a short summary. What I did, what I was feeling. Sometimes just one sentence.

On Sundays I’d sit down and read the week back.

A similar thread. Again. And again. And again.

“What the fuck am I doing with my life.”
“Imagine being in this job for another 10 years.”
“This relationship is making me unhappy.”

It’s easy to ignore a feeling. It’s a little more difficult to ignore a thousand past versions of yourself screaming at you.

Remember the mood filter, the one that hands you a different year depending on the week you’re having? It can only rewrite what lives in memory. A random Tuesday note saying “this relationship is making me unhappy” was written inside the moment, before any filter got to it. It wasn’t trying to prove anything to anyone. Nothing bad even happened that day. I was just tired, and I wrote down what I felt.

And months later, it doesn’t matter what mood I’m in when re-reading these notes. Memory bends to the mood you ask it in. Paper doesn’t.

Your life doesn’t change from one rushed hour at the end of the year, trying to reconstruct twelve months you can barely see. It changes from catching the signal at smaller timescales, weekly, monthly, quarterly, while there is still time to act on it. It was never one January that moved me. It was one too many Sundays.

My situation didn’t change overnight. It took another six months to break things off with my partner. Another two years to quit the job. Obsidian is not a tool that will magically change your life in a day. Be suspicious of anyone telling you otherwise.

But each reflection brings you closer to action. And without a system that let me reflect accurately, I would have quit the job thinking the problem was the company. Found a similar one. Been just as unhappy inside a year, and blamed them for it too.

There is a name for that difference. Double loop learning. Single loop is fixing the action: go get a new job. Double loop is going after the assumption underneath the action: maybe consulting itself is the thing that doesn’t fit me. You can’t do the second one from memory. You need to see the same complaint show up in fifteen different weeks before you stop believing it’s about this project, this manager, this particular month.

Obsidian was never going to be the oracle that tells me what I want from life. It would be a bad sign if it did.

What it gave me was the ability to hold up a mirror to myself.

Obsidian left me nowhere to hide.

It wasn’t only my memory that got fixed

So that’s my story. Which, as evidence goes, is worth almost nothing. An n of 1 who says his life got better. A guy telling you that memory is fallible and offering a solution to it.

But there is something buried in it I did not expect, and it is something I only realised after four years of using this system.

Everything I have complained about so far is a memory problem. It holds badly. It hands things back wrong. It doesn’t connect them up.

So I expected a system to fix my memory. It did. The holes stay circled, the open loops have a place to close, and the Tuesdays are still on file, unedited.

What I did not expect was that it would change the way I think when the system isn’t even open.

Bobby

A hammer doesn’t keep hammering when you put it down.

Obsidian kept operating even after I closed the laptop.

I’m at a birthday party and I know maybe three people in the room. Someone introduces me to a guy. We end up talking for half an hour, mostly about food, and he is good company. Let’s call him Bobby.

While talking to him I already know he deserves a note in my system.

That evening when I get home I open my Obsidian and I make him one. The easy stuff goes down first. Where he’s from. What he does for a living. How many dogs and what they’re called.

Then the part that actually piqued my interest. He’s into cooking, and I host cooking evenings, so that is a link. He trained as a chef, so that is another one. He would get along with my friend Abby, who also cooks, so now those two notes point at each other. And he mentioned a short cooking school in Amsterdam, which lands on the pile of things I want to do.

But the useful part already happened, hours earlier, back at the party, with no app open anywhere near me.

I wasn’t rehearsing. I wasn’t running “Bobby, chef, two dogs” in the back of my head while he was telling me something else. My brain had already sorted him, live, into a structure it knows by heart. Who he reminds me of. What he touches. Where he goes.

So I could just listen to him.

My brain picked up the way my system is structured, the way I link notes and ideas, and internalised it.

What ends up happening is any new information that enters my life goes through this filtering process where, if I pay attention, I can see how my brain is making all of these connections in EXACTLY the same way my Obsidian does.

My memory turned into something superhuman. I remember names, facts, conversations better than I ever could before. I remember them for longer.

And it all allows me to be fully present in the conversation I’m having right in front of me.

I thought that your brain is your brain. You are dealt the cards and that is that.

But Bobby proved the contrary.

It is not a tool you use

There is a thought experiment philosophers use for this. Clark and Chalmers, 1998.

Inga hears about an exhibition, remembers the museum is on 53rd Street, and walks there. Otto has Alzheimer’s, carries a notebook everywhere, looks up the same address, and walks there.

So why do we happily say Inga believed the museum was on 53rd Street, but say Otto only checked a note?

Functionally they did the same thing. The information was there. Both of them trusted it. Both acted on it without stopping to think about it. The only difference is that one of them stored it in neurons and the other stored it on paper.

Their answer is that the notebook is not a tool Otto thinks with. It is part of the thinking. They called it the extended mind, and cognitive science now calls the broader idea extended cognition: your mind doesn’t stop at your skull.

Otto sounds like a special case until you check your own pockets. You don’t know your best friend’s phone number. Your phone knows it. You never sat down and decided to outsource that, and nobody thinks it makes you a worse friend.

It’s why leaving the house without your phone feels different. A forgotten umbrella is an inconvenience. A forgotten phone makes you feel a little naked. That difference is the line between a tool and an extension of you.

A system is the same thing, pointed at your whole life.

That is what it turns into. Not a place you put things. Part of the machine that does the thinking, sitting alongside the wet bit in your skull. The knowing stops living in one place. It lives in the system.

There is software on your computer and there is software between your ears. Run this long enough and you stop being able to point at where one ends and the other starts.

The framework you end up building by accident

Because the thing you actually built was never the folders. It was the mental model your brain now has to stick information to.

Psychology calls that a schema. A framework you slot new information into.

Everybody is running one of these. Almost nobody chose theirs. Yours got assembled out of your schooling, your job, your family, and whatever you happened to be exposed to. And it is quietly deciding what you notice and what slides straight past you.

Build a PKM and you end up designing yours on purpose. In a folder you can open and look at.

And the reason it moves into your head is that nothing connects itself. Every note you file, you decide what it touches. Who it is like. Where it goes. Do that a few thousand times and you are not maintaining it in an app on your computer anymore. You are running the same sorting routine on everything that walks past you.

Remember recognition versus knowing? This is the same gap. Consuming builds recognition. Filing and connecting, deciding where each thing goes, is what builds the knowing. The system forces you to do the second one.

Which is why you can’t short-cut yourself into a system that works for you

Structure comes from toil and repetition.

Someone hands you their perfect template and you get a folder tree, not a mental model, because you skipped the part where the building happened.

You can copy my folders in about four seconds. You cannot copy the two years of deciding where things go, and that is the only part that changed anything.

Stay with a tool like Obsidian for long enough and your brain can’t help but sync up to it.

Every minute you spend in it is like giving your brain a software update where it is installing a mental model for how to interpret, store, and use information that comes at you.

The “easy and quick” benefits you get from Obsidian are the memory fixes.

The hard earned benefits come from becoming intimate with a new mental model.

Subscribe now

So how do I start

There are three places this lives. Pick one and start tonight.

Your days

Create a daily note for every day. Capture what happened. A short sentence or two, and a rating for the day. Extra points for adding one to three photos (because a photo can throw you back into how you felt that day in a way that words sometimes can’t). This is the one that beats the mood filter. You are not writing it for tonight. You are writing it for the version of you in the future trying to work out what actually happened.

Then once a week, read the week back and write a short summary of the week.

At the end of the month, you can now look at your four weekly notes, and write a summary for the month.

See how small the yearly review just became?

Your work

Every project gets a page. Your taxes. A client. The bathroom renovation. A module you are studying for. This is the one that closes the loops and keeps the holes circled.

That page is the home for everything that touches it. Your work sessions. Your meetings. The people involved. The ideas you had in the shower. The artefacts that came out of this project that might help the next one.

The moment a project task lands on you, it goes on the page with a when, and your head lets go of it. When you sit down to work on the project, open the page first. When you stop, write one line about where you left off, so future you starts warm instead of cold.

If it relates to the project, you should be able to find it there. Not in your inbox. Not spread across three apps. Not in your head.

Your curiosity

Life is not only work, productivity and efficiency. A lot of it is just figuring out what you actually enjoy.

So when something pulls your attention, save it. Then write down why it grabbed you. This is the one that turns recognition into knowing. Saving is recognition. Writing down why is knowing.

If you like cooking, save the yummy inducing recipe, but then actually cook it. Invite friends over and brag about it beforehand so you have no way out of making it.

Follow the thing. It is usually pointing somewhere.

The tipping point is optional. Start with one of these three tonight. Future you will be in your debt.

If this landed, my channel is where I show these systems running live: YouTube. You can also sign up to my weekly newsletter where I share more of these type of thoughts.

And if you want a head start, here’s the link to the demo vault. Fair warning from the article you just read: buying my setup gets you a skeleton, not the mental model. The two years of deciding where things go is still yours to do. The skeleton just means you do the deciding in a structure that already works. It is still on you to become familiar with it and make the system uniquely yours before you’ll experience the benefits beyond the memory fixes.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

My inbox had more “failed delivery attempt” emails in it than actual mail.

I live on the third floor of an Amsterdam apartment. I bought it at the start of the year, and only after moving in did I find out the bell doesn’t work.

So a driver stands downstairs holding my package, presses a button that does nothing, presses it five more times, and leaves. Then I get the email.

Fixing it properly means getting quotes, paying someone, and giving up a full day while a man runs a new line and mounts a phone on my wall. Surprise, surprise, lazy cheap me isn’t in the mood for that.

You know what I am in the mood for? Trying to vibe code this in half a day.

The solution appeared to me like a vision from the doorbell gods while sitting at an Amsterdam cafe. There was a QR code on the table: scan it, get the menu, order from your phone. It notifies the waiter “hey, someone is here and they want a cappuccino.”

A doorbell isn’t far off that. Someone is at the door and needs a way to tell you they’re there.

2 hours and thirty cents later the final result is a piece of paper taped to my doorframe.

My requirements

• Two hours to set up

• Completely free

• Two-way chat

The last one matters more than it looks. A ping that only says “someone is at the door” leaves the driver standing there guessing. I want to be able to say “one minute, coming down”, or “not home, leave it with number 2”.

What the visitor does

• Driver stands at my door with a package

• Scans the QR code sticker

• Lands on my doorbell page and presses one button, and gets a sick ass motion graphic of a bell ringing

• I get a push notification on my phone

• We have a quick chat: “here to deliver package” / “1 min, coming down”

The software

I didn’t go looking for any of this. I opened Claude Code and talked through the requirements the way I’ve just described them to you, and it came back with the shortest list of parts that could do the job. Two of them, both free. It help me set up these tools, wrote the 563 lines of HTML code needed, made them live, all in one sitting.

ntfy.sh is the part that makes my phone buzz. Think of it as a walkie-talkie channel with a web address. You invent the address, anything sent to it shows up as a notification on every phone listening in, and there’s no account, no API key, and no server involved. I installed the ntfy iOS app and tuned it to my channel. Anyone who knows the address can listen too, so the entire security model of my front door is that nobody guesses the URL.

Vercel is the part that puts my doorbell page on the internet. A QR code is only a link, and the link has to live somewhere a stranger’s phone can reach. Vercel takes three HTML files, gives them a proper https:// address, and charges nothing for it. The https:// matters more than it sounds, because a phone camera refuses to open a scanned QR code without it.

Two hours and no money rules out anything with a server, a database, or a signup. ntfy removes the backend completely, because the visitor’s browser talks to it directly.

So the whole thing comes down to two moving parts. Vercel holds the web page, which is what the QR code points at. ntfy.sh carries the ring to my phone and allows me and the delivery man to have a two-way conversation. Claude wrote the page, generated the QR code, and laid it out on one A4 sheet at every size from 15mm to 60mm so I could pick the right one at the door.

Then came the tedious manual labour intensive installation.

I printed that sheet for thirty cents, the entire budget of this project, cut out the 40mm one, and taped it to the doorframe with sellotape, directly next to the dead doorbell so nobody has to hunt for it. Two of them side by side now: a button wired to nothing, and a square of paper. It’s a duct-tape solution and it looks like one. It also works.

Build your own

This is the prompt to open with in Claude Code:

Build me a QR code doorbell as static HTML, no backend and no database.

Three files:
1. index.html - what a visitor sees after scanning the QR at my door. One big
   ring button that POSTs to https://ntfy.sh/<my-secret-topic>. After ringing,
   the page becomes a chat: visitor messages POST to the same topic, and the
   page listens to https://ntfy.sh/<my-reply-topic>/sse with EventSource so my
   replies appear as bubbles.
2. reply.html - my console. Subscribe to both topics, merge the messages by
   timestamp, show the last 30 minutes, and give me an input that publishes to
   the reply topic.
3. print.html - an A4 sheet of the QR code at sizes from 15mm to 60mm.

Every message I send should carry an ntfy Click header pointing at reply.html
so tapping the notification opens the console.

Make the visitor page look good on a phone in daylight: big tap targets, high
contrast, and an animation on the bell when it rings.

Then generate the QR with npx qrcode and deploy it with the Vercel CLI.

The first ring

My phone buzzed. Someone was at the door, and for the first time in months I knew about it.

Seconds later, a second ping:

“Delivery!”

“Be right there!” a smiling me responded.

I went down. He was standing there holding my package and looking at the doorframe, and the first thing he said was: “why is no-one else using this?”

I don’t fucking know, man. I have the same question.

Since the sticker went up I haven’t missed a delivery. We all carry a phone every hour of the day, so there is no good reason the alert has to arrive through a box screwed to a wall in 1974.

The result is me being able to talk to the people who dare knock on my humble abode and tell them to fuck off from a distance.

Get the finished version

The prompt above gets you a working doorbell if you’re happy to go back and forth with Claude for an afternoon. I’ve also written the one-stop version: a single document you paste into Claude Code, followed by “build this.”

[Download the QR Doorbell build guide]

It carries the whole brief. What to build, what to deliberately leave out so Claude doesn’t over-engineer a doorbell into a startup, the two things it should ask you for before writing any code, and the limitations it should know about but not try to solve. Plus the human half: setting up ntfy on your phone, printing the QR, and testing it without locking yourself out.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

What are we optimising for?

Most people would say capture and retrieval. Get stuff in, find it later. And fair enough, you want that. Almost everything in your life is a project, and projects go smoother when you can actually find what you need.

That report for your boss. Tax returns. The plant pot you keep meaning to buy. Groceries. Travel to book. Learning guitar. Learning to juggle. Cleaning out the attic.

But capture and retrieval on its own is just a glorified filing cabinet. Stuff goes in, stuff comes out, nothing changes. I want more than that. I want the projects I finish to pay me. Every project you complete should drop XP and loot that makes the next one quicker, or lets you take on a harder one.

Like a video game. You don’t beat the second boss because you got braver. You beat it because you came back with better armour, a sharper sword, and a few levels under your belt.

That’s what a good system does: it takes the grind out of the mundane so you can spend yourself on the bosses actually worth fighting. Obsidian is the one I use to do it.

Here’s the structure I use and teach. Three levels:

• Areas

• Hubs

• Projects

One question sorts them: does it end?

Areas

Areas never end. They’re the big buckets of your life, and most people have seven to ten.

Mine:

• Health

• Wealth

• Career

• Personal Development

• Social

• Environment

• Leisure

Take Health. That’s where you should be able to find everything about your health. Yearly check-ups, dental, the sports you do, diet, injuries, skincare, sleep. If it touches your health and you can’t reach it from this one note, the system has failed you.

Areas are just the top shelf. The real stuff lives in the hubs and projects underneath.

Hubs

Hubs also never end, but they live inside an area. Inside Health I’ve got Gym, Jiu Jitsu, and Running.

A hub is the shelf for one subject. Everything on that topic piles up here: the YouTube video on running technique, a list of running clubs in your city, the shoes you’re eyeing for next season.

And it’s where your running projects live: the ones you’re planning, the one you’re in, and the ones you’ve already finished.

Projects

Projects end. That’s the whole test: an end date or an end condition.

“Run a marathon in 2026” is a project. So is filing your taxes or renovating the kitchen. There’s a finish line, and one day you cross it.

While it runs, the project collects your daily effort. When it closes, it leaves something behind: knowledge notes, the reusable stuff you figured out along the way. Those don’t get archived with the project. They stay up in the hub, waiting for the next project that needs them.

There’s plumbing underneath that makes all this automatic, your daily notes and a bit of tagging. But that’s a piece for another day.

The Marathon Example

Enough theory. Let’s run a marathon.

You sign up for the Amsterdam Kerst Marathon. That’s a project, so you make one: Project - Amsterdam Kerst Marathon.

From that day, everything about this race finds its way onto that one page. Not because you drag it there. Because you mention the project in your daily note and it gets pulled in on its own. Every run you go on. The physio visit for your runners knee injury. The article you read about carb loading. The plane tickets and the Airbnb booking.

You never open the project to “update” it. You live your life, you log it, and the page builds itself.

Then you run the thing. Four hours, twelve minutes. You blow up at kilometre thirty-two and spend the last ten wanting to die, but you finish. status: complete.

Here’s where every other system quietly fails you. The race is done, so the project gets archived. Filed away, never seen again.

And look at what just got buried. Three months of hard-won lessons. Which shoes fit your foot. How many gels to carry and when to take them. The warm-up that finally saved your calves. The training block that actually worked.

You’ll swear you’ll remember all of it. You won’t. Six months later you’re back on Reddit at midnight, asking a stranger the exact thing you already figured out yourself.

So you don’t leave it in your head. As you train, you write it down. And the trick is, those notes aren’t about the marathon. They’re about running.

Which means they don’t die with the project. They stay up in your Running hub, next to the finished race and everything else running you’ve ever collected. And the Running hub lives inside Health. So without filing a single thing, your Health area now knows you ran a marathon, knows it went 4:12, and knows everything you picked up getting there.

Then a year later you get cocky and sign up for Berlin.

This time you don’t start from a blank page. You open the new project, reach into the Running hub, and drag the whole lot across. Amsterdam took three months of flailing to figure out. Berlin takes an afternoon.

It’s the video game again. First marathon, you were a newborn marathon baby, no clue what you were doing, can barely grow a stubble. But then you crossed the line, slayed the level 1 boss (yasss queen), levelled up: beard grown, armour shining, a sword with a few magic spells on it.

You can almost see the next boss, Berlin, shaking in its lil boots.

Want a system like this?

Everything I just walked you through, the areas, the hubs, the projects that fill themselves, is already built and waiting.

If you’d rather not set it up from scratch, I’ve packaged the whole thing as a demo vault. The structure, the daily notes, the logging, the bases, all wired up. You open it, drop your own life in, and start living inside it.
→ the Evergreen Vault

And if your life doesn’t fit neatly into a template, which the interesting ones never do, that’s what I do one on one. We look at your actual projects, how your brain works, what you’re chasing, and build the system around you instead of the other way around.
→ work with me

And if you are still not certain, just book a free 30-min call where I can help you with whatever!
→ free 30-min discovery call

Whichever fits, you stop starting from zero.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

Us humans are short-term creatures.

Traditional task management apps know it, and they prey on it. They prey on the fear that someone out there is checking off tasks faster than you, and that if you slow down for a second you’ll fall behind.

Get things done as quickly as possible. Time is money.

Now put your head down and check off at least 12 of them cute checkboxes before you leave the coziness of your cubicle today.

Todoist, Things 3, OmniFocus, Notion, Google Tasks, TickTick... the list is endless. Some are just a long check-list, some are calendar-first, others pile on useful functionality like AI and notifications.

A lot of cool features. But every single one of them shares the same flaw: the tasks never tie back to a knowledge base. Or if one’s attached, it’s insultingly rudimentary.

A task has no real value the moment it’s checked off.

It sits there, waiting to be ticked. You get your phone reminder, you complete it, you feel good. But the next time a similar task lands in your inbox, one almost identical to the thing you finished two months ago, you start over from scratch.

You start from zero.

You spend half the time re-collecting the resources, links, and context. The other half is the actual work.

And here’s the part nobody in the “productivity” communities talks about: these apps treat the “how” you completed a task as worthless. They’re single-minded and short-term, and they do no favours to future you.

They treat you like a good little robot that loves chewing on the endless supply of tasks life throws at it.

So let me offer something better.

Obsidian, where every task is wired back to your knowledge. There are two ways to do it: the minimalist approach and the maximalist approach.

Why Obsidian beats every task management app

Sure, Obsidian has drawbacks.

There’s no slick mobile app for managing tasks. Setup takes effort. You need a knowledge management structure where every task relates to a project note. Connecting your calendar is an extra step...

But for all of that, it has the one advantage no other system can touch:

Obsidian makes completing your future tasks easier.

Every task is really a mini-project. When you create one, link it to a project, and leave behind whatever you pick up along the way, future you inherits all of it. The next time a similar task comes around, the relevant material is already sitting there waiting.

Take your 2025 taxes.

You create a project note. You start a work session, create the task, and as you grind through the steps you leave notes behind: links to the tax websites, your accountant’s contact details, a page on which expenses are deductible, a step-by-step of exactly how you filed this year.

A year passes. 2026 tax season rolls around. You open “Project - 2025 taxes” and there it all is, waiting for you.

Tasks stop being just tasks. They inject your second brain with reusable knowledge, sitting patiently for the next related task to land on your virtual desk.

The minimalist Obsidian task system

The simplest version is one-line checkboxes scattered across your notes and daily logs. You mark something as a task and tie it to a project:

- [ ] #task Do the thing [[Project]] 📅 2026-05-30

Dataview queries automatically route these tasks wherever you want to see them: an “All Tasks” page, the related project page, or anywhere else.

Back to taxes. You log the task from anywhere in your vault:

- [ ] #task Finish 2025 taxes [[Project - 2025 taxes]] 📅 2026-01-01

You create the [[Project - 2025 taxes]] note, and when you sit down to actually work on it, you spin up a work session to brain-dump everything you’re doing: [[2026-01-01 Work session on 2025 taxes]]. That work session is where the juicy reusable notes get born, the ones waiting for you the next time tax season rears its ugly head.

Because a task rarely gets finished in one sitting, every time you come back to it you create another work session, dump what’s on your mind, and do the work. The “how” accumulates across them.

For a lot of people, this minimalist approach is enough. The Tasks community plugin adds a bit more functionality on top, but you don’t need it to get the core benefit.

The maximalist Obsidian task system

The maximalist approach turns those one-line tasks into notes of their own.

This is where that idea pays off: the task literally becomes a note, a mini-project of its own. It doubles as the work session where the knowledge gets made.

A task-as-a-note gives you frontmatter. You can add any dimension you want: status, priority, due date, scheduled date, created date, energy required, location, and anything else your lil heart desires.

It can also be referenced in multiple places, which the minimalist one-liner couldn’t easily do. When you work on it across several days, you point to it from your daily note with [[link to the task note]], open the note, write out what you’re working on, and create the reusable knowledge right there.

You can check it off from anywhere you referenced it: a random note, a table view, the project note, or the task note itself.

But the multiple references aren’t even the biggest win. The real benefit is how you get to view and interact with your tasks.

In the minimalist approach, you go to the project page or your “All Tasks” page and read a long Dataview list.

With task notes, you pull every task into Base tables. Now you can see them as lists, tables, kanban boards, even a calendar. I lean on TaskNotes for this, because it comes with useful Base queries for all of those views out of the box.

Closing thoughts

After four years of PKM consulting, this is the thing I see people forget about Personal Knowledge Management. It isn’t just a tool to make today easier. Most of the value is in the knowledge you build up over years of working inside it.

There’s still a place for the throwaway tasks: buy groceries, do meditation. Not everything needs a project note. But for anything where finishing it could create knowledge that helps future you, take the road that’s a touch more cumbersome up front.

Yes, it takes time to set up. Yes, it takes more time to get your head around generating reusable information. But skipping it means staying stuck on the same boring problems life keeps throwing at you.

If you do something once, record the “how” of it. The next time the task comes around, it feels a little lighter.

Your job is to make future you’s life easier.

Suffer for them.

Getting started with TaskNotes

I’m putting together a video on this, coming soon to my YouTube: https://www.youtube.com/@construct_by_dee. It walks through a full GTD-style project and task management system.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

There’s a specific kind of tired that doesn’t go away when you close your laptop.

It follows you to dinner. It’s still there when you’re trying to sleep. It’s why you reached for your phone in the middle of a conversation that deserved your full attention. It’s the reason your evenings don’t fully belong to you.

The low hum of open loops: commitments half-made, threads you meant to pick up, the thing from last week’s meeting you haven’t let yourself forget.

At some point, something slipped.

Maybe it was small. A reply you forgot to send, a follow-up that came a few days too late. Embarrassing for a moment, but recoverable.

Or maybe it was bigger. The kind of fuck-up you still think about.

What would have happened if you hadn’t let it slip?
What opportunities did someone else pick up instead?

Meanwhile, someone with the same workload was juggling fifteen of those threads without dropping one.

You’ve seen this person.

The one people lean on because they know they won’t drop the ball.
Walks into every meeting with what feels like an abnormal memory.
Knows the thread from three conversations ago.
Never scrambles.

The person who makes you wonder if the Red Sea just parts for them. While the rest of us are fighting desperately to stay above water.

The one whose work stops when they leave work.
Who is actually present at dinner.
Whose evenings belong to them.

That’s who got those opportunities you missed out on.

Not more talented than you. Not working harder.

Just a better system.

And the frustrating part is you’ve tried to build one.

Everyone has some version of a system.

It works well enough when things are manageable. But as soon as life piles on, the same system that was holding you starts holding you back. A burden you carry out of habit more than hope.

Every tool that was supposed to fix that made the same promise: if you just captured it properly, nothing would fall through.

It still fell through.

Not because the tools were bad. Because every one of them asked you to adapt to it. None of them could keep up when the workload hit.

Here’s why: our brains were never built to hold open loops. They were built to think, to connect, to create. Not to act as a filing cabinet for every commitment made in every meeting ever attended. Every unresolved thread you’re carrying is taking up space that was supposed to go somewhere else.

The problem was never your discipline. It was that you were asking your brain to do something it was never designed for.

For the first time, the system adapts to you. Not the other way around.

What makes that possible is thinking in projects.

Every project gets a home: one page where everything connects back to it. Meeting notes, work sessions, tasks, decisions. Not scattered across three apps and a notebook. One place.

I had a call last week with a client I hadn’t spoken to in eight weeks. I typed two lines to Claude Code: “Create a meeting with Theo at 2pm tomorrow. Give me context.” Thirty seconds later the meeting was on tomorrow’s daily note, linked to the project, linked to Theo. Inside it: the summary from our last session, the full email chain, and every open item still outstanding. I walked in knowing exactly where we’d left off and what to prep.

That’s Obsidian as the structure, Claude Code as the assistant that handles the administration. The logging, the linking, the filing. All of it handled without you touching it.

This results in you having more time to think about the important things instead of spending an hour or two sifting through notes like a depressed European in the winter looking through their medicine cabinet for vitamin D pills.

The evenings come back. The dinners belong to you again. You become the person who walks into every meeting knowing exactly where things stand. Not because you have a better memory, but because you stopped asking your brain to do a job it was never built for.

That’s the system. The one that makes you the person who never scrambles.

Want to build it out?

Lifestack-OS

Website

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

Think about every area of your life that has some level of admin attached to it.

Your work. Your journaling. Your travel planning. Your finances. Your family responsibilities. Your social life: the birthdays you want to remember, the restaurants you want to try, the people you keep meaning to catch up with. Your hobbies. The ideas you have, business or creative. The things you read and watch and want to come back to.

Most people carry all of this in their flesh computer - aka their brain. Or spread across ten different apps, saved links, and notebooks they haven’t opened since the week they wrote in them.

Think about the last time you were at dinner and not fully there. Something churning in the back of your mind. Work, life admin, something unresolved. That is not a focus problem. That is a system problem.

Your brain was never built for storage. It was built for thinking. Every open loop you’re holding in it is taking up space that was supposed to go somewhere else.

After four years of consulting and helping people build PKM systems, I have found that almost all of it falls into three buckets. Three areas you can build out directly in Obsidian, that between them cover most of the overhead in your life.

Your days. Your work. Your curiosity.

Your Days

Not everyone needs a journaling system. Everyone can benefit from one.

We have all heard about the payoffs of a journaling practice. It gives you space to reflect on your life and where it is heading. It helps you understand yourself well enough that you get to the end and can honestly say: this was a life well lived.

The only way to know if your life is still on track is to have regular moments of reflection, and you can only reflect meaningfully if you have something to look back at.

The minimum effective dose is a short summary of your day and a rating. That is enough. If you want to go further, add three photos and a running log of what you did. Suddenly you have a window into every day you have ever lived.

Once your days are connected to everything else in your system, you can set long, medium, and short-term goals and actually see whether your daily life is moving towards them. Not at the end of the year when it is too late to course-correct. Every week. Every month. Your days tell the story. The system just makes it readable.

Most people attempt this kind of reflection at their yearly review. By then the substance of the year is already gone.

Your Work

Everyone has career, work, study, or entrepreneurial ambitions that take up a large portion of their life. And because the portion is large, the overhead is large. Meetings, projects, stakeholders, follow-ups, work sessions: all of it generating threads that need to be held somewhere.

Having a single place where everything related to a project lives (meetings, work sessions, tasks, people) does two things. It makes the current project easier to navigate. And it makes every future project easier too, because the context never disappears.

You walk into every meeting knowing the full thread behind it. Nothing slips. Your work stops following you home.

Your Curiosity

Your ideas are disappearing. Not because you are not curious. You are more curious than you think. But because there has never been a single place for them to land.

The book recommendation buried in a DM. The business idea from a walk. The interest you filed under “later” that is still there, three years on, waiting.

All of these have a fingerprint in them: the particular combination of things only you would find interesting. A picture of who you actually are, the potential that you have never been able to see because the pieces were never in one place.

Give them somewhere to land and the picture starts to form. Your hobbies stop being things you mean to get back to. They become things you actually do. Your interests compound instead of dissolve. You look at what you have collected and think: oh. So this is what I am like.

Why now

Up until the end of 2025 a PKM still required a lot of work.

People who built and maintained these systems genuinely had an edge. They thought better, worked better, dropped fewer balls.

But there was a catch: maintaining the system was in itself admin.

You still had to write the summary. Still had to manually link and tag the note. Still had to process the meeting before the next one started. The system offloaded admin from your brain, but handed some back in the form of upkeep.

For most people, when life gets busy, that upkeep was just enough friction to stop them entirely.

That wall is now gone.

Tools like Obsidian - with Claude Code stacked on top of it - offer, for the first time, what a perfect PKM should actually look like. One where the maintenance happens automatically. Where the input is as effortless as speaking a sentence. Where the system adapts to you, not the other way around.

The payoffs were always real. Now the cost of entry has almost entirely disappeared.

That is why now is the perfect time to start.

The tools: Obsidian, Claude Code, and voice-to-text

Obsidian

One year from now you will have a vault that knows more about your life than any app you have ever used.

Every project you worked on. Every person you’ve met. Every idea you had, connected to every related idea that came before or after it. All of it searchable, all of it yours, sitting on your own device with no subscription and no company that owns it.

That is what Obsidian builds over time. Calling it a note-taking app is technically accurate and completely misses the point. What makes it different is not what it stores. It is what it connects.

What makes Obsidian different from other note-taking apps, is linking.

Notes connect to each other. A project connects to its meetings, its tasks, its people. A person connects back to every conversation you have had with them.

The longer you use it the more the connections multiply, and the more the vault starts to feel less like a filing system and more like an extension of how you actually think.

Claude Code

Every system you have tried before had the same weakness: it needed you to do the admin. Write the summary. Link the note. Process the meeting. File the idea. And at the end of a long day, that was exactly the thing you skipped.

Claude Code removes that layer entirely. It is a personal assistant that sits directly on top of your Obsidian folder, reading it, writing to it, acting on it, without you copying and pasting anything into a chat window. You teach it how you work once. After that it handles the parts you used to skip. You instruct it in plain English, as you would a personal assistant sitting right next to you. You don’t need to be technical. You just need to know what you want.

But it doesn’t just stop at Obsidian. You can connect it to your email, your calendar, excel workbooks, presentations, and a plethora of other services.

Because it has the full context of your vault, it does not just fetch information: it connects it.

Ask it to summarise the emails that came in overnight and it ties each one back to the project it belongs to. It already knows the thread, the history, the people involved.

The more you use it, the more it understands you. The admin that used to fall on you gradually falls on it instead.

Voice-to-text

The last piece of friction in any knowledge system is typing. We now have highly accurate voice-to-text services - Wispr Flow is one - that work system-wide, in any app, in any language. You speak, it types. The barrier to capturing something drops to near zero.

Speak a thought on a walk. Dictate a meeting note in the car. Add a log entry while you make coffee. It lands in your Obsidian, claude processes it.

A PKM still needs structure

The tools are only half the equation. Without a tested structure underneath them, an empty vault is just another app you opened once.

The foundation lives in your Obsidian vault and is built around three things.

Daily notes first. Your daily note is your entry point to every single day. Everything starts there: what you are working on, who you are meeting, ideas you want to capture, how the day went. It is not just a journal. It is a living index that connects outward to everything else in your system: your weekly, monthly, quarterly, and yearly notes; the projects you are pushing forward; the topic pages that organise your thinking. One note a day. Everything else flows from it.

A project structure. Every project gets its own note. Your work sessions connect to it. Your meetings connect to it. Your tasks, your stakeholders, your decisions, all anchored in one place. The context for any project is never more than a click away, whether it started yesterday or two years ago.

A logging system. This is the connective tissue between the daily notes and the project structure. It is a methodology that allows information captured anywhere in your vault to flow automatically to the places it belongs. Nothing needs to be manually filed. The system routes it.

Claude Code sits on top of all three. In the beginning it handles the admin: capturing, linking, processing, retrieving. The more it learns how you work, the more it becomes your thinking partner that already has the context before you ask the question.

This is the combination we have not had until now. A personal assistant with the full context of your knowledge base, connected to your email and calendar, that you can speak to in any language without touching a keyboard.

If this is the structure for you

Without a tested structure, most people lose momentum within the first two weeks. Not because the tools are hard (they are not) but because an empty vault with no methodology is just another blank page. The question that kills it: is this really worth it?

The honest starting point: go to YouTube, find an Obsidian and Claude Code channel you resonate with, and start with their structure. Commit to one system. The specific one matters less than sticking with it.

If the approach in this article resonates: daily notes as your entry point, a project structure that connects your work, a logging system that ties everything together. You do not have to build it yourself.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

Last month, a friend texted me the night before she landed in Amsterdam. She’d just ended a long relationship and booked the trip on impulse.

“I don’t know anyone there, I don’t know what I’m doing, I just need someone to point me somewhere.”

So I run the same routine, as I have countless times before:

• dig through my notes

• pull out the restaurant recommendations

• find the museum that’s worth the queue

• what running clubs to join

• remember which neighbourhood to go to for a drink at sunset

I spent forty minutes digging through three years of Amsterdam notes. Typed it out. Sent it.

She wrote back: “This is exactly what I needed. It felt like having you there.”

That’s when it hit me: she was essentially talking to my notes. Not to me. I was just the translator.

Which got me thinking: what if she could just ask the notes directly, without me in the way?

Here’s what’s now possible that wasn’t two years ago.

You take a specific slice of your Obsidian vault, just the Amsterdam notes, with all your opinions on each place, why you like it, when to go, the spots worth the queue. Point a chatbot at it. Share a link.

She messages:
“Best cafes for sundowners in Amsterdam?”
“Any running clubs I can join?”
“Links to buy museum tickets?”

She gets your notes. Your opinions. Your experience. Not a stranger’s review on Google. A conversation with a version of you that knows Amsterdam and knows her.

You could do this for any knowledge you’ve spent years accumulating:

• Travel - every city you’ve lived in becomes a queryable guide

• Work handovers - a new team member inherits a knowledge base they can ask questions to, instead of documents that go stale

• Jiu jitsu - your training notes, recovery protocols, gym recommendations. Your training partners get to ask your experience questions while you’re asleep

• Parenting - your frameworks on money, health, relationships, passed to your kids not as a letter but as something they can have a conversation with

The problem

Most of us don’t capture our lives.

Not because we can’t. Because we don’t think our story is worth telling. We assume our experiences are ordinary, our opinions unremarkable, our daily life too mundane to record.

But you are the most interesting story you will ever have access to. Nobody else will live your specific combination of cities, relationships, failures, and hard-won lessons. Nobody else will have your take on things.

The laziness excuse used to hold up. Writing takes effort. But AI has quietly removed that barrier - you can speak, ramble, half-finish a thought, and something useful comes out of it. That excuse is gone.

What’s left is the belief problem. You have to actually think your story is worth capturing.

And if you do? The sharing part takes care of itself. Ask an AI to extract the relevant subset. Strip out anything sensitive. Hand it over. It’s like a surgeon carefully slicing out a piece of your brain, making a copy, while making sure the important bits stay attached and safe.

So far we’ve been talking about sharing with people who are here. Friends visiting your city. Colleagues inheriting a project. Your kids asking questions you’ve already answered in your notes somewhere.

But there’s a version of this nobody thinks to ask until it’s too late.

What happens to your story when you die?

Right now: nothing. It disappears with you.

My grandfather died before I could ask him anything useful. And it’s not just the big questions I missed out on. It is the mundane. The boring questions that only a grandson would find interesting. What he ate on a random Tuesday. His favourite book. What his parents were like. His favourite childhood holiday. The best place he ever lived. What he actually thought about my grandmother the first time he met her.

What having my father as a son was like, watching someone become the person I’d one day call Dad.

Nobody thought to write it down. He certainly didn’t. And now the knowledge is just gone.

But plain text files outlive every app ever built on them. An Obsidian vault from today is still readable in 500 years. The AI layer that makes it conversational: it exists right now, and it will only get better.

The conversation I couldn’t have with my grandfather is the one your grandchildren could have with you. Just by you writing things down, in your own words.

A vault full of clipped URLs and half-finished notes doesn’t become a conversation partner. A vault full of your actual thinking does: your opinions, your reasoning, your specific experience. Not what you know but how you know it.

That’s what makes your journal more valuable every year you keep it. Not because you’ll reread it. But because someone else might need it, in ways you can’t predict, at a time when you won’t be around to answer the question yourself.

But you choose to be selfish...

Questions I’m still sitting with:

• Does knowing someone might query your vault one day change how you write in it?

• How do you structure a vault to be navigable by someone who didn’t build it?

• What’s the line between public and private? Amsterdam restaurant notes: yes. Personal reflections: obviously not. But where exactly is that line?

• Could this become social? Friends seeing who else has used your recommendations, connecting through them.

I’m building the Amsterdam version first and I’ll document the whole setup.

If there’s a slice of your vault you’d want to make shareable, what would it be?

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

-2 is the worst day you could imagine. +2 is the most blissful. 0 is dead average - not good, not bad, just another Tuesday.

I thought it was a slightly obsessive thing to do. Turns out it’s the most useful thing I’ve ever done.

Here’s what I found. My best days almost always have at least one of three things in common:

• Novel - learning to snowboard, traveling somewhere new, trying a drawing class, long-boarding through Vondelpark

• Difficult - producing a YouTube video, giving a speech, competing in something, cooking a complicated dish

• Social - dinner with friends, a coffee catch-up, a phone call with family, celebrating someone’s birthday, a boat-ride with fellow humans

Now here’s what’s interesting. Theoretically I could do all three of those things in one day and still have a terrible day.

But that’s very unlikely.

And that’s the point. You have more influence over whether today is a good day than you think. The deck is stacked in your favour - you just have to know which cards you’re holding.

But you can only know which cards you’re holding if you’ve actually been paying attention.

Most people only attempt this at their yearly review. And by then, the substance of the year is gone. You try to reconstruct what made a day great or awful. It’s like asking someone to list every thought they had last week Tuesday.

That’s pretty much it. Start tracking your days, and whenever you sit down to reflect, whether that’s weekly, monthly, or at the end of the year, you’ll actually have something to work with. The enlightened Obsidian brethren/sistren reading this are already nodding.

So here’s your task, starting today.

Open whatever app you want. I will say that if you open a Google Doc I’m going to jump out of the screen and gut punch you, but honestly, any tool works. The software is the easiest problem to solve later. What you can’t do is go back and capture the days you didn’t track.

Give yourself a rating. Write two sentences about why. That’s it.

Just start capturing. Every day you wait is another day without data. And if your life is already great, good. Now find out exactly why, so you can make it even better.

If you want to see what this looks like in practice, here’s a screenshot of my own yearly review in Obsidian, filtered down to my best days. This is just one way to do it, the principle works in any tool.

See you at the end of the year.

And remember: all roads lead back to the daily note.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

My dad is severely addicted to his phone. He’d never admit it. But I’ve watched him pick it up and put it down more times than he’s looked his children in the eye (except of course for the grandchildren my sisters have bestowed upon my parents). Facebook’s algorithm has gotten very good at this.

If an alien species came to earth and only had my dad to make inferences about humanity, they’d conclude that a superior species in the form of a glowing brick had enslaved us.

Even as I’m writing this in a cafe, I can see people getting dopamine-punched every 90 seconds when the telepathic siren-call of their phone goes off. Looking up from their laptops, checking their phone, hot-potatoing through every app, putting it down, picking it up again - oeh, what is this? Another hot potato.

Cal Newport, the author of Deep Work, believes one of the biggest competitive advantages you can have in the coming years is the ability to sit down and do deep, focused work.

Easier said than done with fresh “hits” just a button-click away. Now more than ever we are being robbed of our attention.

One way to cure this is to lock your phone in a kitchen safe for two hours. Another is to build the willpower to sit down and do the work.

But what I want to talk about is activation energy. The bottleneck isn’t willpower. It’s activation energy. How hard it is to start.

Most of us try to optimise with productivity apps, timers, cold showers, journaling routines. My approach is embarrassingly simple.

Restrict yourself to the confines of a meeting.

I stumbled onto this while trying to stop losing track of what I was working on. I started with what I call “logs” - a timestamp on my daily note with a bit of information about what I was working on.

Sometimes the log got long enough that I turned it into its own meeting - a work-session. Every work-session ties to a project, which makes everything easy to find. But something else started happening after a few weeks of working this way.

Four benefits I didn’t expect.

You never start from zero

I have a briefing command built into my setup that looks at the project, the previous work-sessions, outstanding tasks, and what my next steps are - and gives me a summary of my progress so far.

I don’t need to start from a blank screen anymore. I don’t have to go back through previous sessions to extract what I’ve done or what I need to do next. My AI assistant does it for me.

Never. Starting. From. Zero.

Four walls go up

Think of a cinema. You’ve created conditions to pay attention to one thing - the story playing out in front of you. You’re in a dark room, talking is frowned upon, your phone sits untouched in your pocket. The result is full immersion.

When I’m in my work-session, that is the only page that is open. It feels like being thrust into a cockpit where all other distractions are blocked out - at least for the 30 minutes I dedicate to it. It allows me to get into a flow state easier.

Your work-session is the stranger glaring at you when you try to open Instagram.

(Here’s the rabbit hole I’m saving for a future article - the more immersed we are in whatever we’re doing, the more value we derive from that time. Similar to being completely absorbed by learning a new skill that we are interested in, practicing a song on the guitar, or listening to your favourite musician live. Attention divided is attention lost.)

You see the progress

On this very article, I can see the Hansel and Gretel trail I’ve left for myself. If someone looked at this project and dove into each work-session, they’d see my thought process, the paragraphs I decided to remove, the iterations I worked through, the research I drew on to formulate my opinion - from inception to completion.

That trail doesn’t just feel good to look at. It’s receipts.

You create material for reflection

What gets measured gets managed.

Just “completing” the project doesn’t make the next one any easier. Having notes of how you completed it lets you see the shortcuts, the inefficiencies, and - above all - build something that helps future you.

This isn’t just about getting into flow. It’s about getting better at the work over time.

A chef won’t relearn a recipe. They write it down. When they need to whip up that Italian dish again, the note is ready. They don’t go back to an old lady in rural Italy to learn everything from scratch.

Yet here we are. Back in rural Italy. Every. Single. Time.

If you’re tired of being a slave to the pitiful (yet powerful) distractions of life, try the work-session approach. But be warned, it might make your life better - not everybody is ready for it.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

Last year, the tech geeks were quietly talking about AI tools in the dark corners of the conference. Still niche, still early adopter energy.

This year, that was everyone.

Old, young, no coding background whatsoever. Building read-it-later apps, book catalogues, study assistants, home automation systems, mindfulness applications. Someone even built a personalised music teacher for their kid. All of it inside Obsidian, all with Claude Code.

You can now build something like that in 10 minutes. And the best part? It’s yours. No subscription. No company that owns your data. Yours.

The age of traditional apps is dead. The age of personalised, self-built apps is here.

Clive Thomson is a tech journalist who has spent decades covering the intersection of people and technology. He attended the summit and said it gave him a glimpse of what the future of PKM will look like. And he’s right. This is where we’re heading.

But Clive also said this:
“Every time new technology arrives, we’re excited and overwhelmed in equal measure. 20 years ago, what tools did we have? A word processor. Today we have too many to choose from. How do we manage this?”

My take-away question became: how do we stay essentialists now that almost anything is possible?

Most people never figure it out. Most hop from one app to the next promising one that promises to relieve all the tension that has been building up in their disorganised minds. But sooner or later, they all collapse under their own maintenance.

What finally fixes it isn’t discipline or a better app. It’s removing the friction entirely.

PKM systems don’t fail because people aren’t disciplined. They fail because the friction of filling things in is just enough to stop you from reaping the benefits. In the past it almost felt like an unteachable lesson - that you need to start and stop 10 different systems and methodologies, that you need to grind your teeth and build up the mental resilience to input the data in the right way, to use the correct template, to diligently sit and fill in your journal, to give that meeting a summary. End of a long day, you tell yourself you’ll process it later. You don’t. Three weeks later you’re piecing together fragments of memories you should know by heart.

I can finally say: it’s no longer a discipline problem. It’s a design problem. And it’s fixable.

And the scale of what’s now possible is bigger than most people realise. You can have multiple Claude instances running at the same time. One is processing your meeting. Another is doing your week review. You’re making coffee. This isn’t one assistant. It’s a department.

But with that power came a question nobody had a clean answer to. Another speaker at the conference, Thijs Pepping, said it best: “The more you measure, the more you behave like a robot.”

With data about ourselves so easily acquirable, how can we remain in the driver’s seat of our own lives? Will we quietly hand over the thinking to an algorithm?

We’re building systems to track sleep scores, day ratings, step counts, productivity metrics. At some point the measurement starts to erode the thing you’re measuring. Worth sitting with, especially for those of us building quantified-self systems in Obsidian.

What I’m trying to hold onto is this: AI wraps around your methodology, it doesn’t replace it. The structure of your life - your life areas, your projects, your daily rhythm - stays yours. Designed by you, for you. AI handles the admin layer around it. The logging, the summarising, the linking, the processing. The thinking stays human.

I don’t know if that’s the answer. But it’s the line I’m trying to protect.

I’ve never been more excited about where this is going. And I’ve never thought harder about what to protect from it.

We’re launching a 4-week live cohort in April to help you build this yourself.

Interested in building your own system?
Sign up and you’ll get a free email series + a free live webinar before the cohort opens - and first access when sign-up goes live.

https://lifestack-os.com/

Even one workflow that removes one piece of friction is worth it. That’s how the avalanche starts.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

Duncan is the knight. Without his sword and armour, he can still throw a punch, but can’t do the things a knight needs to do.

The sword isn’t just a thing Duncan holds. It’s an extension of him. With it, he reaches further, strikes harder, moves differently. Take it away and he’s a different fighter.

If you are here you already know this feeling. Your vault isn’t just where you store notes. It extends what you’re capable of thinking. Your links surface connections your brain forgot it made. Your structure holds context across weeks and months that your memory never could.

If my Obsidian vault got deleted tomorrow, I wouldn’t just lose files. I’d lose a part of how I think. A small piece of my brain, chipped out. That’s how deeply these tools become part of you. In cognitive science they call this extended cognition: your mind doesn’t stop at your skull.

But there’s a lot of admin that comes with these tools. The sword needs to be sharp. The armour needs to be polished and maintained. All of it makes Duncan a better knight, but none of it is fighting.

That’s where Egg comes in. The squire. He sharpens the sword, polishes the armour, handles the logistics. Egg removes the friction so Duncan never has to stop being a knight to do maintenance.

That’s what AI is for me. That’s what Claude Code is. It handles the formatting, the tagging, the linking, the filing. All the stuff that keeps the system running but isn’t the actual thinking.

But here’s the thing about Egg.

Egg can sharpen the sword. Egg can’t (or rather, shouldn’t) swing it.

Skills are built through doing. Duncan becomes a better fighter by swinging his sword, getting bruised, taking that feedback to better approach challenges in the future. In cognitive science they call this enacted cognition: understanding created through action.

When you decide what to capture, how to structure a note, what connects to what, that’s you swinging the sword. The thinking happens in the doing. Not before it, not after it, in it.

This is the part most people get wrong with AI. You should know your structure. You should be the one enforcing it.

If you don’t, if you let AI run wild in your vault and do the structuring in a black box without you understanding it, it starts modifying the sword in ways that don’t reflect you. The tool stops being an extension of your mind. It becomes an extension of something else.

Once you don’t understand your tools anymore, the connection between you and the sword is broken. And a sword you don’t understand isn’t useful anymore.

The optimal way is to use your squire to talk through how to make the tools better. You understand what it’s doing. You can see how the sword is changing to your benefit. You’re in the loop.

Anything more than that, where the squire starts making decisions without you, that’s when things break down. The sword changes in ways you didn’t choose. The structure no longer reflects how you think. And one day you pick up the sword and realise you don’t recognise it anymore.

The outcome?

A guy with a shitty sword and a bold squire.
A guy who’s squire will outgrow him.
Congratulations.

If this article made something click, this is the fastest way to make it stick:

1-on-1 Consulting
Obsidian x Claude Code Workshop

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

I looked at my daily note and thought, “this is ugly.” About as pleasant to look at as my family WhatsApp group’s AI slop videos that my dad keeps spamming. At least those have colour. My vault was a wall of grey plain text. Every tag, every link, every log entry, identical. So I decided to see if my personal assistant (AKA Claude Code) can prettify it for me.

Obsidian on the left. Claude Code on the right.

I started simple. I wanted all the tags starting with #log to be green. So I asked, “can you turn all the log tags green?” BAM. Hulk has entered the chat. Green baby!

While admiring my genius I noticed the green only showed up in Reading View but not in Live Preview. I told Claude it was a dumb piece of shit. It politely asked me to inspect the element so it could figure out what went wrong. A few back-and-forths later, everything was coloured:

Patient as a saint. Gently held my hand through the whole process while I insulted it.

I didn’t write a single line of code. I don’t know what a CSS selector is. Still don’t. I just looked at my vault, said what I wanted to see, and watched it update in front of me.

Now my tags are green and blue. My project links are red. Work sessions glow green. Log entries have clean little orange dots instead of checkboxes. I open my daily note and I can feel what kind of day it was before I’ve read a word. Like a trader scanning green vs red.

These colours don’t just live on the daily note. The dataview queries that pull logs into project pages keep the formatting too.

And the whole time, the only thing I did was complain. ”Too bright.” Claude toned it down. ”The links all look the same.” Claude coloured them by type. ”Those checkboxes are annoying me.” Claude turned them into dots. Every time I looked at my screen and didn’t like something, I said so. And a split second later, it’s done.

And when I get bored of it in three months, I’ll just tell Claude to change it again.

If you can complain about it, you can build it.

I’ve been using Claude Code not just for colours. It processes my meeting notes, pulls in YouTube sources, runs my weekly reviews, helps me with taxes.

You used to scroll through hundreds of community plugins hoping someone built the thing you wanted. Now you just describe it.

How To Get This For Your Vault?

Open Obsidian on one side. Claude Code on the other. Look at your vault and find something that annoys you. Tell Claude to slave away until your Obsidian looks like an extension of you.

If you want these CSS snippets, I made them available for download, but they’re based on my structure (projects are red, meetings are green, log tags are green, area tags are blue).

My honest advice? Learn how to do this yourself.

I run Claude Code x Obsidian workshops every couple of weeks. We set up Claude Code from scratch, walk through use cases like this one, and by the end you have a toolkit you can take home and use immediately. No dependency on me. Just you and your vault.

If you want more of this, I write a weekly newsletter about building systems in Obsidian, Claude Code, and the weird things I discover along the way.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

At this point I feel like a Pokémon trainer watching Charmander become Charizard.

Every evolution unlocks a new ability.

Obsidian gave me structure.
Claude Code gave me read, write, think.
Wispr Flow gave me hands-free capture.

This evolution? It learned a new move: control the screen.

The problem Up Until Now

Claude Code lives in my terminal.
Obsidian lives in its own app.

I’ve been running them side by side for months, and Claude has learned exactly how my vault works. It reads my notes, creates new ones, logs my day, and processes meetings.

Claude could do everything behind the scenes, but it had no way to control what I was actually seeing on screen. It couldn’t open a note for me. It couldn’t navigate.

Then I Found The Obsidian CLI (Command Line Interface).

“Anything you can do in Obsidian can be done from the command line.”

It’s built directly into Obsidian (not a plugin), bridging your terminal and Obsidian and letting anything running in your terminal, like Claude Code, control what happens on your screen.

You know that scene in Iron Man where Tony talks to Jarvis?

Tony walks in. No keyboard. No mouse.
“Wake up. Daddy’s home.”

“Pull up the schematics.”
“Enhance.”
“Cross-reference last Tuesday.”

Holographic screens flying around. Data surfacing exactly when he asks for it. He never clicks. He never searches. He just speaks, and the system responds.

That’s the vision.
That’s where this is going.

Before the CLI, I had Jarvis locked in the basement.
He could do the research, crunch the numbers, write the report.

He just couldn’t put anything on my screen.

Now Jarvis has access to the holographic screens.

“Jarvis, process this meeting.”
Summary created. Action items extracted. Note opens.

“Jarvis, open the lasagne recipe.”
There it is.

“Jarvis, log gym, 45 min biceps.”
Logged. Note opens.

“Jarvis, find the last two Charmander feet pics.”
Both open. Right in front of me.

No searching. No clicking. No sweating through your folder structure like a drug addict tearing apart the couch cushions looking for the next hit.

I speak. Immediate relief.

More upgrades. More power.

One problem:

You Can’t Multiply By Zero.

If your vault is empty, Jarvis has nothing to pull up on those holographic screens. No notes to open. No meetings to reference. No recipes to find.

All that functionality, and your Jarvis is just standing in the corner staring at his feet. Granted, they are handsome feet.

But if you’ve been building, even casually, your system just learned to talk back.

And it has a lot to say.

How Do You Access It?

Two things to do manually in Obsidian:

1. Update to Obsidian 1.12+ and ensure you have access to the CLI feature (currently Early Access, planned to be free for all users)

2. Go to Settings → Core plugins, toggle on CLI, and click “Register CLI path”

Keep Obsidian running. The CLI talks to it as a server.

Then launch Claude Code in your vault:

claude -p /path/to/your/obsidian/vault

And prompt it:

“Verify the Obsidian CLI is working by running obsidian help. Then add instructions to CLAUDE.md for how to use the Obsidian CLI for all vault interactions: searching, reading, creating, appending to notes, managing properties, daily notes, backlinks, tags, and using --format=json for parseable output. Also check what vault this is connected to with obsidian vault.”

That’s it. Future Claude Code sessions in your vault will automatically use the CLI instead of raw file reads.

A Quick Note on AI and Your Vault

I know I’ve been talking a lot about AI in knowledge management lately. That’s because I genuinely believe this is the next evolution, not a trend.

I hear the privacy concerns. I’m actively working toward running these tools locally (mac mini gang) so your data stays yours.

Even if you’re not ready to evolve your Pikachu yet, at least be aware of what’s possible for when you are.

Your Obsidian’s Next Evolution

If you want to hand your system its next evolution, I run Claude and Obsidian workshops regularly.

In the workshop, we run through different use cases to get the gears turning on how you can use this inside your life and your Obsidian:

• Meeting processing

• Source processing

• Daily logging

• Work sessions

• Content creation

You pick the ones that fit your workflow.

The course that comes with the workshop teaches you exactly how to upload each skill.

Think Neo in The Matrix:

“I know kung fu.”

Except it’s:

“I know meeting processing.”
“I know project management.”
“I know daily note logging”
Upload. Done.

Each skill is a new ability your vault didn’t have five minutes ago. You simply choose which ones you want to learn.

If you’d rather skip the learning curve and have it all working by the end of the day, book a one-on-one consulting session and we’ll set everything up in one go.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

It’s Valentine’s Day.

Most people’s effort ended with Googling “romantic restaurants near me.” Panic booking something with a suspiciously enthusiastic review from someone named Gary.

Gary once left a five-star review for a petrol station.
“Really enjoyed the atmosphere.”

You outsourced intimacy to Gary.

Allow me to shame you.

Shame.

Romance isn’t hard, kids.

Three months ago she mentioned her grandmother’s recipe while on the phone with a friend as you walked past. She called it “the best thing I’ve ever tasted.” One sentence. In passing. Immediately into the void. Forgotten by everyone else thirty seconds later.

Except you.

Because you, my dearest, you have the void by the nuts.

You have Obsidian. You don’t rely on memory. You rely on capture.

So when Valentine’s Day arrives, it resurfaces. You call her parents. You get the recipe. You make it.

She stares at you like a baby trying ice cream for the first time. Confused. Emotional. Slightly destabilised.

This is not romance.
It is attention.
And attention is attractive.

Most people confuse effort with attention. They are not the same thing.

Naming your firstborn after me is optional, but encouraged.

That pattern, capture once, forget completely, resurface on command, is where a system stops being storage and starts becoming a thinking partner.

Until January 2026, retrieval was the weak link. We built structure. Daily notes. Project systems. Tag taxonomies that made us feel organised and morally superior. It helped. But you still had to remember what tag you used. What you called it. Where past you decided to bury it.

It worked. Technically.

But not. easy. enough.
Friction kills systems.

So what killed my friction?

Claude Code.

You now have an AI chatbot pointed at your Obsidian folder. Something you can instruct in plain English. Not a search bar. Not a filter. A butler who has read every note you have ever written and still shows up for work.

Instead of remembering tags or digging through your own clever structure, you tell Claude to sift through everything and bring it to you. Handkerchief on your lap. Silver tray. Exactly what you ordered.

And it does not get tired. Or annoyed. Or forget.

Instead of trying to remember where you stored something, you type:

“It’s Anna the chemist’s birthday soon. Give me gift ideas based on all the logs and information I have on her.”

And it combs through everything. Passing mentions. Old ideas. Patterns. Context. It brings you high confidence hits. Experience gifts. Creative callbacks you forgot existed.

The image above is not creativity.

It is compounded attention.

She is getting a horse piñata filled with her favourite treats. Not because it is random. Because I want to see the self-proclaimed horse girl’s moral hesitation before the first swing.

Character is revealed under pressure.

I did not sit for three hours brainstorming like a Pinterest mother. I captured things over time and built a system that squeezes the vault for me.

Here is the uncomfortable part.

You actually have to use the system. If you did not capture anything about her, this does not magically save you. You cannot retrieve what is not there.

This system compounds effort. It does not create it.

But when you do capture, casually and consistently, your life becomes queryable.

And this is not just useful for saving your relationship.

Your meeting prep is reactive because you rely on memory.
Your study prep is chaotic because you rely on willpower.
Your travel planning is 17 tabs peppered with hope.
Your writing process mostly consists of staring at a blinking cursor.

You keep trying to think harder in the moment instead of remembering better over time.

Obsidian holds the raw material of your life.
Claude lets you interrogate it in plain English.
Together, they give you leverage.

Systems create asymmetry. Memory does not.

You can build a system that makes you look unfairly prepared in life.

Or you can keep outsourcing your life to Gary.

What steps do I take next?

If you want to set this up properly instead of half-installing it and hoping for the best, there are three ways to work with me:

1. Hands-on Workshop
Small groups. We install and configure everything on your device and work through real use cases immediately.

2. 1-on-1 Consulting
For people who want a custom setup tailored to their vault, workflow, and business.

3. Obsidian Demo Vaults
If you prefer to reverse-engineer systems at your own pace, my demo vaults show the exact structure I use.

If this article made something click, pick your level and start there.

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

I’ve been running hands-on Claude + Obsidian workshops. Seeing what actually works, and where people get stuck, shaped the last 30 days.

A Warm Coffin

Most knowledge systems don’t fail because they’re bad.
They fail because they’re comfortable.

They work just well enough that you never feel an urgent need to change them, even as they slowly stop helping you learn, think, or adapt. There’s no dramatic breaking point. Just gradual stagnation.

That’s what I mean by a warm coffin.
Something comfortable enough to stay in while it quietly restricts you.

That’s also why people defend bad systems so aggressively. Not because they’re good, but because they’re familiar.

Knowledge management evolved roughly like this:

oral stories → writing → computers → folders → search → databases → networks

Each step was a genuine leap forward. Each one also reframed the previous approach as a polite historical artefact.

Computers gave us scale.
Folders gave us containment.
Search gave us hope.
Databases gave us structure.
Networks give us relationships.

Obsidian lives in that last shift.

People still stuck at the folder, search, or database stage are essentially cave painting, just with better fonts and recurring billing. Obsidian already escapes that trap by treating knowledge as a network instead of a filing system.

But that’s a low bar.

Compared to how people will be using Obsidian once AI is properly integrated, you might already be falling behind. You might feel technically savvy compared to your parents. Compared to people actively using AI to learn better, you’re still operating inside familiar limits.

What’s different now isn’t the tools themselves. It’s the rate at which they evolve.

We’ve seen transformative tools before. But they spread slowly enough that falling behind was temporary.
Today, the pace of change is accelerating.
When change compounds, falling behind gets harder to recover from.

The silver lining is that these are also the most powerful tools we’ve ever had for learning faster.

The irony is that the faster you learn these tools, the faster you can learn everything else. The advantage compounds.

That’s why the most important skill over the next five years isn’t a specific technology. It’s learning how to learn, and these tools finally make that practical instead of aspirational.

And uncomfortably, the only real defence against AI isn’t avoiding it.
It’s wielding it properly.

I wrote about my first 48 hours with Claude Code. Then my first 14 days. Now it’s been a month.

I wish I could say the novelty is gone and the honeymoon phase is over. But I’m still the guy cornering strangers at dinner parties to explain how fucking sick my Claude + Obsidian system is.

What stuck

Daily logging without the admin hell

Everyone loves talking about “building a second brain.” Nobody loves manually creating notes, linking people, tagging projects, formatting timestamps, and choosing the correct emoji.

That’s not thinking.
That’s data entry.
And I refuse to do data entry in 2026.

Claude does it now

I say, “logged meeting with Dan at 9:30,” and Claude handles the timestamp, tags, emojis, and links. Even when the actual filename is “Dan Blonde stocky pierced nipples marathon gym sauna guy.” It matches by first name and carries on with the level of professionalism and lack of eyebrow-raising I would expect from a loyal personal assistant.

Unlike some of you.

What used to take minutes now takes seconds. Capture stopped being optional. My vault went from patchy to embarrassingly detailed. My future biographer is going to need a lot of hugs.

The most valuable thing Claude does isn’t clever.
It’s clerical.

What evolved

From capturing to producing

This surprised me.

At first, Claude was purely for input. Logging. Formatting. Organising. Then that became normal. With speech-to-text dictation (thanks, Wispr Flow) feeding straight into my Claude personal assistant, capturing felt effortless.

Then it clicked.

When I sit down to work now, I don’t start with a blank page. I start a work session and ask Claude to surface everything I’ve already thought about the topic. Relevant notes. Previous work sessions. Old arguments. Half-formed ideas I forgot I had. All of it appears directly in the note I’m working on.

Instead of asking Claude to generate ideas, I ask it to resurface mine. This article came together that way. Nothing here was invented. The raw material already existed.

Writing feels easier because the hardest part was never writing. It was locating your own thinking. Now I start from my past work, decide what matters, and assemble the argument.

I’m not starting from other people’s information.
I’m starting from mine.

The same applies to sources

Instead of letting articles, videos, and threads pile up and radiate low-grade guilt, I process them as they come in. Key points are extracted, tagged, and stored.

“I’ll read this later” is a lie.
“This is already processed” actually ships work.

The thinking is mine. The organisation is automated. When I sit down to produce something, I start from my material, not a blank page and not something generated for me.

I’m building on my own knowledge, not outsourcing my brain.

Speed via a vault graph

I asked Claude to generate a simple JSON graph of my vault to speed up note lookups. It’s essentially an index. Instead of searching through files every time, Claude checks the graph first.

This came from a comment on the previous article, and it turned out to be a clear improvement over the default “go search the vault” approach. Faster lookups. Fewer broken trains of thought.

Security concerns didn’t magically disappear

One thing that hasn’t faded with time is my discomfort around security.

Using AI like this means sending personal context to a large corporation. Not abstract data. Your notes. Your work. Your thinking. That’s a real concern, and it should be treated as one.

Most setups today involve some level of external processing. That’s the tradeoff. There are people running fully local or hybrid systems where sensitive data never leaves their machine. I haven’t settled on my own version of that yet, but I’m actively looking into it.

For now, it’s a conscious tradeoff. You’re trading some privacy risk for very real upside in leverage, speed, and learning. That doesn’t make it reckless. It does mean you should understand what you’re opting into.

The point isn’t to ignore the risk.
It’s to be aware of it and decide whether the reward-to-risk ratio makes sense for you.

For me, so far, it has.

What I dropped

One massive CLAUDE.md file

Long instruction files fail for the same reason terms and conditions exist: they’re technically correct and functionally ignored.

Claude behaves exactly like a human here, just with a slightly better memory. The longer the instruction file, the more slips through the cracks. Past a certain length, you’re not giving guidance. You’re just yelling into the void.

The fix was to compartmentalise.

Specific Claude Skills for specific actions. Small, focused instruction files loaded only when needed. Claude follows these. The monolith never stood a chance.

The four I use constantly:

Daily Logging
Handles timestamps, tags, emojis, and links. Used constantly. Losing this would hurt emotionally.

Process Meeting
Turns raw transcripts into summaries, action items, and follow-up emails. An 80-minute call becomes something usable in minutes.

Process Source
Extracts summaries and key points from articles, videos, and threads. My source folder went from graveyard to tool.

Claude feels like a very fast intern who never complains and never questions why I saved a hobby horsing video tagged #bachelor-party-idea.

Week Review
Reads all seven daily notes, summarises the week across life areas, and selects three photos prioritising people, experiences, and milestones. What I used to skip now takes 90 seconds and actually happens.

What 30 days taught me

The compound effect is real.

When logging takes no effort, you log everything. When reviews are automatic, you actually review. When sources are pre-digested, you use them. When connections take seconds, your vault stops being a graveyard and starts being useful.

The biggest shift wasn’t any single feature. It was the collapse of the gap between thinking something and having it captured, connected, and retrievable. My vault went from storage to a thinking aid. That’s what everyone promises. Almost nobody delivers.

I’ve used Obsidian for three years. Capturing was never the problem. Producing was. Having your own processed material changes that completely.

If you want to learn this setup properly

I’m running small, hands-on workshops for people who haven’t touched Claude Code yet, or who want to move past “I installed it” into actually using it.

We set everything up on your device, and immediately work through real use cases:

• Daily log automation

• Project and meeting creation

• Pulling web content straight into your Obsidian templates

If this article made something click, this is the fastest way to make it stick.

Sign up here

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter

The Setup (Minutes, Not Hours)

I wasted hours searching for “Claude Code + Obsidian” tutorials. Found nothing. Everyone’s using it for programming, not note-taking.

So I stopped searching and just installed it. Pointed it at my vault. Started talking to it.

It worked immediately. Setup was the only hard part, and even that took minutes.
(A short setup video will be live on YouTube sometime next week).

The Magic Glue: CLAUDE.md

Claude is brilliant, but forgetful. Every conversation starts fresh.

The fix is a file called CLAUDE.md. It’s not documentation. It’s an instruction manual for your personal assistant.

When Claude makes a mistake, you teach it. Add a rule. It doesn’t make that mistake again. Claude can even update the file itself.

My CLAUDE.md started with three lines. It’s now 370+ lines covering:

• Where my daily notes live

• My exact logging format with emojis and tags

• How to create projects, work sessions, meetings, etc

• How projects link to work sessions

• Patterns for gym visits, social interactions, content creation

Every correction became a new rule. Every repeated explanation became documentation.

(You might hear about “Skills”, separate instruction files for specific tasks. For Obsidian workflows, CLAUDE.md alone works fine at the beginning. Start simple.)

Teaching Your 80/20

Most of what I ask Claude to do doesn’t change much from day to day.

Log something. Capture a meeting. Create a project. Update today’s note.

Instead of teaching Claude everything, I taught it those few actions first.

The teaching loop is simple. I ask for what I want. If it’s wrong, I correct it once. The correction becomes a rule in CLAUDE.md. The mistake doesn’t happen again.

After a few days, the interaction stopped feeling like “using AI” and started feeling like talking to my system.

Now I’ll say something like:

“Had a catch-up with Stefan about the Q1 roadmap”

Claude knows where meetings live. It creates the note, links Stefan, adds today’s date, fills the metadata, and connects it to the right project.

I don’t think about structure. I don’t think about format. I just describe what happened.

That’s the 80/20. Teach the handful of actions you repeat every day, and let the system handle the rest.

Some other findings…

Once the basics were set up, I discovered what Claude is also useful for.

Metadata at scale.
You can update a lot of notes at once. I no longer write summaries. Claude reads each note, summarizes it, and fills the field. I batch this at week’s end: “Check all meetings and fill missing summaries and tags.” Housekeeping in seconds.

Tip: having a base file where you can see all your meetings at a glance makes this oddly satisfying. You can watch Claude fill it in.

Search replacement.
I ask Claude instead of Googling. It knows my vault. It’s a personalized search engine.

I can ask it to find a super simple salmon recipe online, create the note using my recipe template, and add a cooking log to today’s note.

It beats the back-and-forth of copying, pasting, refining, choosing templates, and linking everything manually.

Embedded in Obsidian.
I run Claude inside Obsidian using the Terminals plugin. It lives in the sidebar. No extra window. Oddly cozy.

The Real Shift

The real gain isn’t speed. It’s friction removal.

When logging takes three seconds instead of two minutes, you actually do it. More notes. More connections. More ideas captured.

Parkinson’s Law says work expands to fill the time available. The reverse is true too — remove the busywork and time appears.

It feels good to hand off a task, let Claude run for a couple of minutes, and go grab a coffee.

Fourteen days in, I can’t imagine going back.

The Elephant in the Room: Privacy

Before you point an AI at your personal notes, know what you’re signing up for.

Your data leaves your machine. Claude Code feels local, but files go to Anthropic’s servers. Your notes travel over the internet.

Claude explores. It reads files you didn’t explicitly mention to understand context. It might stumble into your journal, your client notes, your half-written thoughts.

Things to consider before you do this:

• Passwords, API keys, credentials

• Client or patient confidentiality

• Personal information

• Compliance (GDPR, HIPAA)

What I do:

• Keep sensitive files outside my vault

• Use .claudeignore to exclude folders

• Review changes before confirming

• Keep regular backups

Is it worth it? For me, for now, yes.

The worst that could happen is Anthropic gifting me a few free therapy sessions after reading my journal.

But here’s the irony. The reason I love Obsidian is that your data is yours. Plain markdown on your machine. No cloud unless you want it. No company reading your notes.

AI tools strip that away. The data lives on corporate servers. It might leak. It might train models. It might be used against you.

So be cautious. Know what you’re sharing. Decide if the benefits outweigh the risks for your situation.

I’m digging into offline Claude Code options. When I find these sandboxed solutions, you’ll be the first to know how to set it up.

Want to Learn?

I’m running small, hands-on workshops for people who haven’t touched Claude Code yet.

We’ll set it up on your device and immediately run:

• Daily log automation

• Project and meeting creation

• Web scraping into your templates

Just setup and real use cases, right away.

Sign up here

Much love,
Dee
Construct by Dee

Website | YouTube | X / Twitter | Newsletter