The ritual outlives the reason

The original teacher made a practical decision. The cat was loud. Tying it up solved the problem. There was no deeper meaning.

But without that context, the action became doctrine. The ritual outlived its reason. People followed the practice without understanding what made it useful in the first place.

I keep thinking about this story when I watch how people talk about AI.

Two camps, same mistake

There are two camps right now:

• “AI is the future, it will change everything, you need to adopt it or get left behind”

• “AI just hallucinates, it’s overhyped, it’s useless”

Both stated with absolute certainty. Both camps are tying up the cat.

The believers copy prompts and workflows without understanding the context that made them work. Someone got a useful result once, shared their process, and now people follow it like the gospel. They treat AI outputs as authoritative without the verification that made the original insight useful.

I see this constantly. Someone shares a prompt that worked for their specific situation. Within days, it’s being passed around as “the right way to use AI.” The context gets stripped away. The situational becomes universal.

The skeptics do the same thing in reverse. “AI hallucinated once” becomes “AI is always useless.” A situational observation becomes universal dismissal. They tried something, it didn’t work, and now they’ve closed the door entirely.

Both responses share the same flaw: turning a specific experience into a general rule.

What actually works

The original teacher tied up the cat for a practical reason: it was making noise. The value was never in the ritual. It was in understanding why something worked in the first place.

Same with AI.

When something works, ask why. What was the context? What made this situation different? What would need to be true for this to work again?

When something fails, ask the same questions. Was this the right tool for this task? Was the input good enough? What would I need to change?

The value is in understanding your context, your use case, and what makes something useful for you. Not in copying what worked for someone else. Not in dismissing what failed somewhere else.

The middle ground nobody wants

The uncomfortable truth is that AI is situationally useful. Sometimes remarkably so. Sometimes not at all. It depends on the task, the context, the quality of what you’re feeding it, and what you’re trying to get out.

This is unsatisfying. People want clear answers. “Always use AI” or “never trust AI” are easy positions to hold. “It depends” requires ongoing judgment.

But that’s the reality. The teacher didn’t tie up every cat in the world. He tied up one specific cat that was causing one specific problem. When the problem changed, the solution should have changed too.

Stop tying up the cat

The next time you see someone sharing the “right” way to use AI, ask what problem they were solving. The next time you hear someone dismissing AI entirely, ask what they actually tried.

The ritual is never the point. Understanding why something works is the point.

Stop tying up the cat.

It was sitting in my drafts, something I’d written about using AI to consume my personal knowledge base rather than generate new content. I wasn’t sure it was ready. Then somehow it went live, got picked up by Hacker News, and suddenly thousands of people were reading something I hadn’t meant to share yet.

The conversation that followed was fascinating, and not because people agreed with me. What struck me was seeing my own ideas reflected back through different lenses, watching people take the same core observation and arrive at completely different emotional responses to it.

The capability that cuts both ways

My post was about empowerment. I’d connected my Obsidian vault to AI: three years of meeting notes, reflections, observations about building software. Instead of asking AI to write things for me, I started asking it to find patterns in what I’d already written. Hidden connections. Forgotten insights. The evolution of my own thinking over time. It felt like gaining access to a version of myself with perfect memory.

But the Hacker News crowd went in wildly different directions with it. One camp immediately jumped to surveillance: yes, AI is excellent at consuming information, they said, and that’s exactly what makes it terrifying at scale. The panopticon problem. Mass behavioral analysis. Authoritarian potential. The other camp went the opposite way: LLMs hallucinate constantly, they said, so any “patterns” AI surfaces from your notes are probably fabricated anyway. The whole thing is useless.

I found myself amused by the polarization. Here I was, sharing a specific experience of connecting my notes to an AI and finding it genuinely useful, and the reactions split into “this enables sophisticated mass surveillance” versus “this is all hallucinations and garbage.” Both extremes, almost no middle ground. The dramatic contrast said more about how people are processing AI right now than it did about my actual use case.

The work that came before

Another thread in the conversation stopped me. Someone pointed out that at least half of the “AI superpower” I was describing came from something that had nothing to do with AI: the fact that I’d spent years putting everything into Obsidian in the first place. The years of capturing. The discipline of documenting. The habit of reflecting after meetings, writing down observations, processing what I was learning. AI just made that investment liquid.

This reframed something for me. I’d been so focused on the retrieval breakthrough that I was undervaluing the capture that made it possible. AI can only surface patterns from what you’ve given it. The real work happened long before any AI was involved: the decision to document, the consistency of showing up to write reflections even when I was tired, the thousands of small moments of capture that accumulated into something queryable.

I think about this watching others work effectively with AI, too. The ones who benefit most from AI tools are often the ones who were already organized, already thinking systematically, already documenting their decisions. The tool amplifies what was already there. If you haven’t built the foundation, there’s nothing for AI to work with. This isn’t an argument against AI. It’s a reminder that the investment in capture and reflection pays dividends in ways I couldn’t have predicted when I started.

Verification as the emerging skill

The skeptics in the thread raised something worth sitting with: how do you know the patterns AI surfaces are real? LLMs can hallucinate. They can find connections that don’t exist. They can present fabricated insights with complete confidence. Several commenters questioned whether AI actually summarizes or just abbreviates, potentially missing critical details along the way.

This is a legitimate concern. I’ve caught my AI assistant making up quotes from my own notes, inventing patterns that weren’t there, confidently asserting things that a quick check proved wrong. But I think the pitfall where detractors fall here is about this comparison: the question shouldn’t be whether AI is perfect, but whether AI with verification beats pure human recall at the same task.

Could I personally read three years of notes and find the patterns myself? Not really. My memory is partial and biased. I remember recent things more than old things, dramatic things more than quiet things, things that confirm what I already believe more than things that challenge it. AI with verification is better than what I could do alone, because treating AI outputs as hypotheses to investigate rather than conclusions to accept gives me a starting point I wouldn’t have otherwise. The skill is in the verification loop: when AI surfaces a pattern, I ask it to cite the specific notes, I read those notes myself, and I check whether the pattern actually holds. Sometimes it does. Sometimes it doesn’t. But the verification process is faster than trying to find the patterns manually in the first place.

The accidental lesson

I didn’t mean to publish that post. I wasn’t sure it was ready, wasn’t sure the ideas were fully formed, and wasn’t sure it matched with other content I put out. But sometimes the things we hesitate to share are the ones that resonate most. The post reached more people by accident than most things I carefully plan and polish.

There’s something in that. The draft that felt unfinished connected with people precisely because it was honest about being a work in progress, about figuring things out in real time, about the gap between having an insight and fully understanding its implications. The Hacker News conversation filled in perspectives I’d been too close to see: the surveillance angle I’d been compartmentalizing, the importance of the capture work that came before, the necessity of verification as a practice rather than an afterthought. These weren’t criticisms of my thesis. They were extensions of it.

Maybe that’s what publishing is for. Not to present finished ideas, but to put half-formed ones into the world and see what comes back.

This week

If you’re using AI as a thinking partner, a few questions worth asking yourself. What work did you do before AI that makes it useful now? The capture systems, the documentation habits, the reflection practices. AI amplifies those investments. And how do you verify? When AI surfaces an insight, what’s your process for checking whether it’s real? Building that verification habit is probably more important than optimizing your prompts.

Here’s to more accidental publishes!

You capture the idea. File it somewhere. Promise yourself you’ll review it later. You won’t.

I’ve been experimenting with something different. A system that maintains itself. That remembers what I’m working on without me having to tell it. That synthesizes my past thinking when I need it.

This week, something has changed significantly.

The problem I was solving

My days are fragmented nowadays as director of engineering at Buffer. Strategy conversations. Hiring interviews. Team syncs. Context switches constantly.

The mental overhead isn’t the meetings themselves. It’s holding everything in memory between them.

• What did we decide last week about the API redesign?

• What feedback did I give this engineer two months ago?

• What was that insight from the 1:1 I had on Tuesday?

I used to rely on notes. But notes require you to remember what you wrote, where you put it, and when it’s relevant.

What I wanted was a system that could surface the right context at the right moment without me asking.

The next layer

Back in August, I gave Claude Code access to my Obsidian vault: 1,400+ notes of meeting transcripts, reflections, strategic thinking. That was transformative. But it was still just a powerful tool I had to direct.

This week, I pushed further. Three additions:

• Skills: Dedicated modes for specific tasks: planning my day, thinking through problems, drafting team communications. Each skill loads with the right context and approach already baked in.

• Working memory: A persistent layer that knows my current projects, preferences, and patterns. Every conversation starts with this baseline context loaded automatically. No more re-explaining what I’m working on.

• Learning extraction: At the end of each session, the system tries to identify what’s worth remembering: decisions made, insights surfaced, tasks completed. It updates its own memory without me having to curate anything.

The shift: from a tool I use to a system that knows me.

What actually changed

Two moments stood out this week.

First: I asked it to help plan my day. Instead of starting blank, it already knew my current priorities. Then it surfaced tasks from my Obsidian notes that were still marked unfinished. Things I’d forgotten about. It was holding context I’d dropped.

Second: I finished a conversation where I’d brainstormed about a tricky strategy concept. Later, I checked the working memory. The decision and philosophy was already there. Captured without me doing anything.

These are small moments right now in these early stages, but they add up to something significant: the system is starting to hold cognitive weight I used to carry myself.

The principle underneath

This isn’t about AI features. It’s about cognitive architecture.

Most people use AI as a slot machine. Put in a question, pull the lever, hope for a good answer.

The breakthrough comes when you treat AI as an extension of your own memory. When you give it enough context that it can reason about your work the way you would, if you had infinite recall.

• Memory without maintenance

• Context without cognitive load

• Your own insights, accessible at the speed of thought

What I’m building toward

I’ve been thinking about how to share this more concretely.

Not a course. Not endless content. Something small and opinionated.

Something like a 60-minute setup that shows knowledge workers how to structure their notes, configure AI as a reasoning partner, and create simple rituals that keep the system useful.

The goal would be to help people reduce mental overhead while operating at a higher level. Building calm through better cognitive architecture.

If this resonates with how you work, I’d love to hear from you. Reply to this email or comment on the Substack post, and tell me what weighs on your mind.

An urgent message. Bad news. A reminder of that thing you’ve been avoiding. The thing you know is important, the one that’s been sitting there waiting for you.

Your body reacts before your mind catches up. A tightening in the chest. A heaviness. You know this feeling. It’s dread.

And then the thoughts begin.

The branching

Your mind takes that single event and starts projecting. One thought leads to another. Each branch spawns three more. Suddenly you’re not dealing with the thing itself. You’re dealing with every possible consequence of the thing, and every consequence of those consequences.

The urgent message becomes a difficult conversation, which becomes a damaged relationship, which becomes a failed project, which becomes questions about your competence, which becomes...

You’re no longer in the present. You’re living in a future that doesn’t exist.

This is overthinking. And it happens so fast that by the time you notice, you’re already several layers deep.

The paralysis

Here’s what makes this insidious: overthinking disguises itself as preparation.

It feels like you’re being responsible. Thinking through the possibilities. Considering the outcomes. Being thorough.

But you’re not preparing. You’re paralyzing.

The project doesn’t get started because you fear the feedback. The conversation doesn’t happen because you fear the confrontation. The decision doesn’t get made because you fear being wrong.

Procrastination and overthinking feed each other. The more you think about what could go wrong, the less you want to start. The longer you wait, the more time your mind has to invent new catastrophes.

Mark Twain captured this perfectly: “I’ve had a lot of worries in my life, most of which never happened.”

Most of the futures you’re dreading will never arrive.

The bias

Your brain has a bias toward threat detection. This made sense when threats were physical. A rustle in the bushes could mean a predator. Better to assume danger and be wrong than assume safety and be dead.

But now the threats are emails and deadlines and difficult conversations. Your brain applies the same threat detection, the same worst-case pattern matching, to situations that don’t warrant it.

So when you start thinking about possibilities, your brain doesn’t give you a balanced sample. It gives you the scary ones. The embarrassing ones. The ones where everything falls apart.

This is important to understand: your overthinking is biased. It’s selecting worst-case scenarios from the full spectrum of outcomes and presenting them as likely.

The reframe

Here’s an experiment: take whatever you’re overthinking about and imagine it going perfectly.

The difficult conversation leads to understanding. The project gets praised. The decision turns out to be exactly right. People are impressed by your work. You handled it well.

Does this feel harder to believe than the catastrophe version?

That’s interesting, because it’s equally imaginary. Neither version is happening right now. Neither version is real.

Your mind chose to simulate the disaster. It could just as easily have simulated the success. Both are fiction. Both are projections. Both exist only in your head.

The difference is how they make you feel. And how they affect what you do next.

The present

The antidote to overthinking is simple but difficult: come back to now.

Right now, the worst-case scenario isn’t happening. Right now, you have a thing to do. A next step to take. A decision to make.

The future will be handled by your future self. And here’s something worth remembering: your future self will be better equipped than you are now. They’ll have more information. More context. They’ll know things you don’t know yet.

You don’t need to solve every downstream problem before taking the first step. You just need to take the first step.

The email needs a response. Write it.

The project needs starting. Start it.

The conversation needs having. Have it.

Your future self will handle what comes next. Trust them. They’ve got this.

The practice

Noticing overthinking is itself a skill. Most of the time, we’re deep in the spiral before we realize we’re spiraling.

When you catch yourself projecting into catastrophic futures, pause. Ask: is this happening right now? Usually the answer is no. What’s actually happening is you’re sitting somewhere, feeling anxious about a story you’re telling yourself.

The story might come true. It probably won’t. Either way, it’s not happening now.

Focus on now. Do the next thing. Trust that you’ll handle whatever comes.

This is the only strategy that actually works against overthinking. Not arguing with the thoughts. Not trying to out-think them. Just recognizing them for what they are: projections, not reality.

And returning to the present, where your life actually happens.

This is the routine I’ve been rebuilding since coming back from Bali. The closet cleanup I wrote about was the visible first step, the quick win that reduced morning friction. The phone and desktop reset was about creating space in my digital environment. But underneath both of those was something I’d been craving for a while: getting back to the consistency I’d practiced for years but had let drift.

I’m a few weeks into it now, and I feel amazing.

There’s a general prevailing wisdom about routines that I used to believe, which is that they’re restrictive. Waking up at the same time every day sounds rigid. Training for a race sounds demanding. Tracking protein and cutting empty calories sounds like you’re constraining yourself, limiting your options, removing spontaneity from your life. From the outside, people doing these things consistently look like they’re operating under strict discipline, trading freedom for optimization.

What I’ve rediscovered is the opposite. Consistency creates freedom.

The 7am alarm eliminated a negotiation I didn’t realize was exhausting me. Throughout 2025, every night used to involve some version of: what time should I wake up tomorrow? Do I have early meetings? Can I sleep in? Should I set the alarm for 7 or 7:30 or 8? That micro-decision, repeated daily, added up. Now the alarm is set to 7am, and when it goes off I get up. No negotiation. The decision is already made, which means I’m not spending willpower on it.

The half marathon gave the morning movement a purpose beyond just “exercise is good for you.” I have a training plan. Tuesday is an easy run, Thursday is intervals to build tempo, Sunday is the long relaxed run. I don’t wake up wondering whether today is a gym day or a rest day or whether I should just skip it because I’m tired. The plan decides. I just execute.

The food changes followed naturally. You can’t train for a half marathon while eating garbage. Your body tells you what it needs when you’re asking it to run 15 kilometers on a Saturday morning. So the focus becomes hitting protein targets, cutting the snacks that made me sluggish, paying attention to what actually fueled good training sessions versus what left me dragging. The constraint of the race created clarity about everything else.

What this makes me reflect about is how the structure created space rather than taking it away. Waking at 7am means my day doesn’t start with work. I have several hours before my first meeting where I’ve already accomplished something that matters to me. The consistent training means I have more energy in the afternoons, not less. The food discipline means I’m not managing the downstream effects of whatever I grabbed when I was hungry.

I’ve known this for years. But knowing something and living it consistently are different, and it’s easy to let routines drift when life gets busy or travel disrupts your rhythm. Bali was the reset I needed. Coming back with intention, rebuilding these foundations deliberately, has reminded me why I built them in the first place.

The pattern across everything I’ve been writing about since returning is the same: what feels like adding structure actually removes friction. The closet with fewer clothes means faster mornings. The phone with fewer apps means less distraction. The consistent wake time and training schedule mean I’m not constantly deciding what to do with my body. The structure handles the small decisions so I have capacity for the ones that actually matter.

As the days get longer after winter, I’ll probably push the alarm back to 6:30, then 6. Earlier sunrises make earlier mornings easier. The consistency adapts to the season, but the foundation stays.

I’ve been thinking about this for years, actually, and I realized I wrote about the digital version of this exact concept back in 2020, before this Bali trip, before the last closet cleanup, when I was figuring out how to create space in the digital environments where I spend most of my working hours (especially in the middle of COVID lockdown). Re-reading that piece now, what strikes me is how consistent the principle has remained even as the specific tools and contexts have changed: whether it’s physical objects in your closet or digital objects on your phone screen, the goal is the same, which is to reduce the cognitive load of navigating options that don’t serve you so you can reserve your mental energy for things that actually matter.

The piece I wrote back then was about space, the deliberate creation of emptiness in digital environments that are designed to fill every pixel with something demanding your attention, and reading it again after the wardrobe experience made me realize these aren’t separate optimization tactics but expressions of the same underlying system, which is that sustainable focus, calm, and clarity come from intentionally reducing the small sources of friction before they accumulate into the kind of tension that lives in your body for months.

So here’s what I wrote in 2020 about creating space in your digital life, and I’m sharing it now because if the closet cleanup resonated with you, this is the next place to look.

“One can furnish a room very luxuriously by taking out furniture rather than putting it in.”

– Francis Jourdain

I’m pretty much obsessed with the idea of space. It is something I always strive for: on my wardrobe, few but selected, high-quality clothes, on my desk, a few essential items that are a joy to use, on my backpack, objects that I use all day or that bring me joy.

I believe that is something we can apply to our digital lives as well.

On the phone

This was my iPhone’s only screen back in 2020:

No notifications. No red balloons with a number. No apps reachable in one tap to get distracted. I put all the apps together for one reason: they are very hard to find manually.

“How do you find anything here?”, I hear you ask. I type the name of the app in the search, sliding down from this view. That forces me to be a bit more mindful of what I’m reaching out to, and sometimes allows me to stop the urge to check something unimportant.

I still have three apps available with one tap of a finger: phone (because this is, at the end of the day, still a phone), email, and my to-do app of choice.

Also, having all this space allows for a fantastic side-effect: I can enjoy my phone background for real. I switch around between some of them that serve as reminders for when I reach out for the phone.

2025 version: what’s changed

The setup now has a calendar preview widget visible, and the dock holds email, calendar, and Todoist (my task manager). Keeping it simple. The principle is the same but the practice hasn’t been constant. I’ve fallen off this setup multiple times over four years, usually during periods of high stress when I convinced myself I needed more apps readily available, more notifications enabled, more ways to stay on top of everything that felt urgent. The phone screen would gradually fill up with apps again, the notification badges would return, and I’d find myself checking my phone first thing in the morning out of anxiety rather than intention.

Every single time, going back and purging it all out made things exponentially better. Cleaner. Calmer. The forcing function of typing to find apps saved me from countless context switches, those moments where I reach for my phone with intention to do one thing and the presence of other apps would have derailed me into checking something else entirely. But I had to relearn that lesson multiple times before it stuck, each time recognizing that the temporary feeling of control I got from having everything visible was actually the opposite of control, it was letting my environment decide what deserved my attention instead of making that decision myself.

The other practice that makes this setup actually work is physical distance: keeping my phone out of reach during the workday, or in my backpack while commuting, creates just enough friction that I have a moment to check with myself “am I reaching for it because I need it, or is it out of anxiety or boredom?” Most of the time when I actually have to get up and walk across the room to grab my phone, I realize I didn’t need it at all, and the minimal screen setup means that even when I do grab it, there’s nothing pulling me into a fifteen-minute distraction spiral. The combination of physical distance plus digital minimalism is what actually works, not just one or the other.

On my computer

This was my desktop in 2020:

I do my best to hide everything: dock and menu bar, gone, notifications are hidden while I am working, no files on the desktop (I know screenshots can pile up there, don’t let them), and a soothing background that gives me a sense of calm and, you guessed it, space.

Then I operate in two different modes: I either have a handful of apps open, as I have now, having some space between them and not filling every single pixel of my screen, or I have one app working in full screen, allowing me to focus (what I might do more when coding, or doing any other task that requires deep work).

How this has evolved

The desktop has followed the same pattern as the phone: periods where I maintain it perfectly, followed by periods where stress or overwhelm convinces me that I need to see everything all at once. The dock creeps back. Files start accumulating on the desktop because I’m moving too fast to properly file them away. Notifications start breaking through because I’ve told myself that this one thing is important enough to warrant the interruption.

And every time I notice the desktop has gotten messy, every time I realize I’m getting pulled into distractions instead of staying focused, the solution is always the same: strip it all back down. Hide the dock again. Suppress the notifications. Clear the desktop. What I’ve learned over four years isn’t that I’m disciplined enough to maintain this perfectly, but that I know how to recognize when things have drifted and I know exactly what to do to reset back to calm.

The hardware has changed too: I now work from a MacBook Pro 14” instead of a larger display, and the menu bar stays visible for cosmetic reasons since the screen has a notch anyway. Working from a smaller screen forces me to be even more deliberate about what I have open, since there can be fewer things competing for my attention at once. To be honest, I’d love to go back to having a bigger screen as soon as I manage to create a fixed desk space to work from again, but the constraint has taught me that focus isn’t about screen real estate, it’s about intentionally limiting what’s visible regardless of how much space you have available.

The desktop setup creates what I think of as a neutral starting point for focus: when I sit down at my computer, there’s nothing screaming for my attention, no visual reminder of the seventeen things I’m not currently working on, just the blank canvas that lets me decide what matters right now rather than having my environment decide for me.

The pattern underneath

Making space is a continuous effort. Like in the physical world, you have to handle things and keep them under control. Like in the physical world, you need to let go of what doesn’t work for you anymore.

What I’ve learned over the years since writing that original piece is that this isn’t about minimalism for its own sake or some aesthetic preference for empty screens. It’s about recognizing that every icon, every notification badge, every open tab represents a small claim on your attention, and those claims add up in ways that aren’t immediately obvious until you experience what it feels like when they’re gone.

The closet cleanup last week removed friction from my mornings. The phone setup removes friction from the dozens of times I reach for my device throughout the day. The desktop setup removes friction from the moment I sit down to do deep work. Each one seems trivial on its own, but collectively they’re the difference between fighting your environment and having your environment support the work you’re trying to do.

This week, make some space either on your phone or your computer. I’m sure you know where to begin.

He told me I had months of accumulated stress living in my abdomen. Physical tension from carrying too many things mentally for too long.

I came back to Barcelona Wednesday night with four days before work started again on Monday. Four days to intentionally build calm instead of letting the chaos decide for me.

The first thing I did Thursday morning: cleaned out my closet.

Why the closet came first

I’ve been thinking about simplifying my wardrobe for months. The idea kept surfacing: capsule wardrobe, uniform-style dressing, keeping only what makes me feel good.

But I never acted on it because it felt like something I’d get to eventually. A nice-to-have optimization, not urgent.

Coming back from Bali changed that prioritization.

I had four days to set up systems that would support calm instead of fighting against it. The closet was first because it’s a daily decision point: every morning, I stand there choosing what to wear from options that don’t all make me feel confident.

More options makes decisions harder. Harder decisions become friction. Friction erodes calm, which I’m trying to protect.

What actually happened

I piled everything on the bed. Everything went into one massive pile to force real decisions instead of category-by-category rationalization.

First pass: keep only what fits now and makes me feel good wearing it.

That eliminated half the pile immediately. Clothes that don’t fit best. Things I liked the idea of but never wore that often. Items that fit but didn’t make me feel confident.

What remains is:

• Black jeans that fit perfectly

• Two pairs of Uniqlo Airism pants (black, comfortable, versatile)

• Light gray sweatpants for casual days

• A few polos and linen shirts I actually wear

• Black and gray tank tops and t-shirts in different fits

• One sleeveless silk shirt that’s slightly more dressed up

• A dark blue wool shirt from Arket I love

• A couple of structured shirts that work for everything

Not extreme minimalism. I kept a linen suit and some items that have emotional value. But the core principle held up: if it doesn’t fit now or doesn’t make me feel good, it goes.

What this actually solved

The next morning I opened my closet and every option was something I’d intentionally kept. No decisions about whether something still fits. No trying things on and changing three times. No low-level anxiety about whether this choice makes me feel confident.

Just pick something, it works, move on.

That’s maybe three minutes saved. But it’s not about the time. It’s about the cognitive load.

Decision fatigue is real. Every small decision throughout the day uses the same mental resource as big decisions. Your brain doesn’t distinguish between “what should I wear” and “how should I approach this project.”

Reducing small friction points protects capacity for things that actually matter.

The other benefit: clarity about what’s actually missing. When everything in your closet is noise, you can’t see the real gaps. Now I can see exactly what I need: a few crisp white shirts, a couple more t-shirts that actually fit well, maybe one more versatile jacket. The signal emerged once the noise was gone.

The pattern underneath

This connects to everything I’ve been learning about sustainable high performance. You can’t build calm by trying to be calmer during chaotic moments. You build calm by eliminating the small sources of friction before they compound into stress.

The closet is one system. But the same principle applies everywhere:

• Too many browser tabs open creates low-level anxiety about what you’re forgetting.

• Too many apps fighting for notification space fragments your attention constantly.

• Too many items on your task list that you’re never going to do creates guilt every time you see them.

Each one seems trivial. Collectively they’re the tension that accumulates in your body over months of carrying too much mentally.

What I’m building toward

I’m not trying to eliminate stress by working less or caring less or lowering my standards.

I still want to achieve at a high level. Strong performance matters to me. Building things that matter drives me.

But I don’t want to brute-force it anymore. I don’t want to accumulate months of tension in my body because I’m carrying too many small frictions that could be eliminated.

The closet was the first move. Small, tactical, immediate.

Walking into Monday morning, I have one less source of daily friction. One more system supporting flow instead of fighting it.

That’s how you build sustainable high performance: not by trying harder during chaotic moments, but by intentionally reducing the small things that make chaos harder to handle.

What held me back before

I’ve thought about learning to scuba dive for years because the idea always fascinated me, rooted in my relationship with water that started when swimming was my sport growing up. I competed on a swim team, spent hours in the pool every day, loved being in water (and still do), but there was always this fundamental limitation: your breath, the way that no matter how good you get at swimming, eventually you have to surface because you’re constrained by your lung capacity. Scuba diving removes that constraint entirely, letting you stay down, explore, and move through the water without the clock of your breath counting down, and that possibility always appealed to me, even though I never committed to actually learning it.

Then I started dating Julia, who’s an Advanced Open Water diver who has dived around the world, and hearing her experiences when we got together made the fascination grow because these weren’t abstract possibilities anymore, but real places she’d been, real creatures she’d seen, real experiences I could have too. I initially thought about getting certified in Barcelona, where it would be close to home and familiar, but talking with Julia made me reconsider, since the sea life would be more beautiful somewhere tropical, and Indonesia is known for strong currents, so learning somewhere with changing conditions from the start might make me a better diver rather than optimizing for comfort.

But there was still resistance, the kind that’s harder to name, because learning something new means being a beginner again and being a beginner means performing in front of others: an instructor watching you struggle with skills that are probably second nature to them, other divers who already know what they’re doing, that discomfort of being incompetent while people watch that you have to embrace if you want to learn anything worth learning.

The decision to start

This trip to Bali was planned as a diving trip from the start, where Julia was going to dive and I came with the intention to finally learn, though what made it real wasn’t just the decision I’d made before we left but booking the course itself, committing to those specific days and putting myself in the position where backing out would mean wasting the opportunity I’d created for myself.

The first day was pool training, what they call confined water, where you learn the basics in a controlled environment before going into the ocean, and I thought being comfortable in water from years of swimming would make this easy, but it didn’t because breathing underwater through a regulator is completely different from anything swimming teaches you.

Every swimming instinct says hold your breath when your face goes under, and empty your lungs and refill them again with every stroke, but the regulator keeps feeding you air and your nervous system has to unlearn what it knows and trust something new, where breathing needs to be slow, controlled, and not explosive. This is what took most time to learn properly over the course of the week.

But what surprised me was how systematic the training was, with each skill broken down into small steps followed by clear demonstrations and practice until competence, before moving to the next skill, reminding me of how good software engineering training works too, where you don’t try to learn everything at once but master one piece and then build on it.

The moments that shifted everything

The second and third days, we went into the ocean for real diving, descending to 12 meters and then 18 meters over the course of four dives, and this is where the breakthrough moments happened. The first ocean dive was disorienting because while the pool had clear water, a flat bottom, and controlled conditions, the ocean has currents and limited visibility and fish everywhere and a reef below you that’s constantly changing depth, so my brain was processing so much new information that I could barely remember the skills we’d practiced since I was too busy trying to take it all in.

But by the second dive that day something clicked as my buoyancy control got better and I stopped fighting the water and started moving with it, where years of swimming gave me something useful in the form of comfort with being weightless and letting the water hold you, though diving added a new dimension since in swimming you’re always moving toward the surface or parallel to it while in diving you’re suspended in three dimensions where you can hover in place and move up, down, forward, backward. That freedom took getting used to, but once it started to feel natural, the mental space that had been consumed by “am I doing this right?” started opening up, and I could actually look around, notice the coral, watch the fish, be present in the experience instead of just surviving it.

The third dive was when it really came together, where we descended to 18 meters and I felt calm and comfortable with steady breathing and controlled movements, no longer thinking about every micro-adjustment because the skills had started becoming automatic, and in that space I could actually enjoy being underwater with its weightlessness and silence and this completely alien environment that somehow felt peaceful. That progression from overwhelm to competence to enjoyment happened in about six hours of diving, which is what struck me: how quickly you can go from “I have no idea what I’m doing” to “I can actually do this” when the learning is structured well and you’re fully immersed in it.

What learning to learn actually looks like

Tim Ferriss talks a lot about meta-learning, the skill of learning how to learn by breaking down complex skills into components and finding the minimum effective dose of practice and identifying what actually matters versus what’s just noise, and scuba diving gave me a direct experience of those principles in action though what made it work wasn’t just the formal instruction but the conversations between dives.

Julia and I would talk through what I’d just experienced, questions like how do you build intuition around buoyancy and how do you know when it feels right after you’ve adjusted your BCD at depth, where she’d explain and I’d ask more questions about how you breathe, whether it’s through the diaphragm or more shallow lung breaths, and I even put my hand on her chest to feel the rhythm of it. Those conversations filled in what the instructor couldn’t convey in a demonstration, because while the course taught me the essential skills like buoyancy control, equalization, and emergency procedures and communication underwater, with each skill practiced deliberately until it became reliable, understanding how to make those skills feel effortless came from debriefing with someone who’d already internalized them to the point of not needing to think about them.

What I learned was that efficiency underwater comes from doing less: stay neutral in the water, don’t kick your fins too hard, don’t breathe so hard, and if it’s a drift dive, just relax and let the current take you forward. Each dive got longer because my oxygen consumption improved, not from trying harder but from being mindful of all the small things that waste energy, and the feedback loop was immediate where you try something, feel how it works, talk through what happened, adjust, and try again. The water doesn’t lie because you can feel when your buoyancy is off, and you can see when you’re using too much air because you’re working too hard, so the environment gives you constant information while the conversations help you interpret that information and know what to change.

That’s the meta-skill: recognizing that learning happens in layers, where formal instruction gives you the framework and practice gives you experience, but conversation with someone who’s already done it helps you build intuition faster by compressing the learning curve through understanding not just what to do but how it should feel when you’re doing it right.

What learning from zero unlocks

I’m sitting here with my PADI certification, but what I’m noticing isn’t about self-improvement being a new concept to me. I’ve always been focused on getting better at things. What’s new is remembering what it feels like to learn something from absolute zero, to have no existing skill to build on, and to start as a complete beginner.

I’ve spent years sharpening existing skills like getting better at engineering leadership and improving how I communicate and refining my productivity systems, and all of that is important, but it’s iterative improvement on things I already know how to do. Learning to dive was different because I had no foundation here beyond swimming, helping with water comfort, while diving itself was completely new, and experiencing that arc from zero to competent in four days cracked something open.

If I can learn something from scratch like this then I can also keep sharpening the skills I already have, where the confidence works in both directions because learning from zero proves you can still build new capabilities and that proof makes you trust that your existing capabilities can keep growing too. This connects to something I wrote about a couple weeks ago about the gap between taste and skill, where your ability to recognize good work develops faster than your ability to create it and that gap is brutal when you’re learning because you can hear exactly what’s wrong before you know how to fix it.

Diving replicated that experience because I could feel when my buoyancy was off before I knew how to correct it and I could see I was using too much air before I understood how to breathe more efficiently, where the gap between what I could perceive and what I could execute was right there, immediate, undeniable. But the gap closed fast because the feedback was so direct, and watching it close reminded me that this is how all learning works, that the gap is supposed to be there and it’s not a sign you’re failing but the training ground itself. What cracked open wasn’t just “I can learn new things” but “I remember how this feels”.

The discomfort, the incompetence, the gradual closing of the gap, and remembering that feeling is what makes everything else feel more possible again.

Why starting from zero matters

When you spend years working on iterative improvement you can forget what building from scratch feels like, where you optimize what you already know and get incrementally better at existing skills, which is valuable but also limiting because starting from zero reminds you that competence is buildable and you’re not just refining what you can already do but proving you can acquire completely new capabilities, and that proof changes what feels possible.

I’m leaving Nusa Lembongan with a PADI certification, but what matters more is remembering that learning from zero is still available to me, that the skills I have aren’t the only skills I can have and the things I’m good at aren’t the only things I can get good at. And that reminder ripples everywhere because if I can learn to dive then I can also keep sharpening the skills I already have, where the confidence isn’t just about acquiring something new but about trusting the process of improvement itself, whether you’re starting from zero or building on years of experience.

That’s what adventure unlocks: not transformation or reinvention but just the reminder that growth is renewable, that curiosity can take you places you haven’t been, that being a beginner at something even for a few days cracks open possibilities you’d forgotten were there.

Our second villa? Booked yesterday, Saturday.

We’re half a week into a vacation we have not planned to the last detail. And it’s working perfectly.

The key is to know exactly what you can’t compromise on, and stay flexible about everything else.

What we decided early

Before we left, we were clear on a few things: this was going to be a diving trip (you can’t change the whole vibe overnight), we wanted access to yoga spots, spas, coastal areas, and great restaurants.

The specifics could be decided later.

Switching from a diving trip in Southeast Asia to a hiking trip in the States or city tourism in Europe the night before wouldn’t be an iterative change. It would be drastic. The diving trip was the whole idea that shapes everything else.

But Bali or the Maldives? Where exactly do we stay? Which specific dive sites? Those are iterative decisions that get better with more information and the closer you get to the actual dates.

What waiting actually gave us

Booking villas a couple of days before arrival meant we got places with massive discounts. We’re talking 60% off in some cases, but more than the money, we got better information. The weather changes. Our energy levels change. What we wanted from each location shifted as we experience the trip.

It seems like waiting to decide would close up options, but it actually creates clarity.

The framework under the surface

This is straight from Lean and Agile practices; it’s called the “last responsible moment” principle.

You don’t decide early just because you can. You decide when you have the most information and the decision actually needs to be made.

For Bali, that meant:

• Deciding this would be a diving trip months in advance (foundational decision that shapes everything)

• Deciding on Bali vs. Maldives a few weeks out (both support diving, other vibes and pricing became the differentiator)

• Deciding on regions a few days out (weather patterns become clearer)

• Deciding on specific villas 24-48 hours before (prices, availability, and our actual needs aligned)

• Deciding on specific dive sites once we got there (local conditions and our energy levels mattered)

Each decision happened at the last moment it could have been made responsibly. Not sooner. Not later.

Where this shows up everywhere

This same pattern runs through engineering decisions constantly.

I’ve seen teams defer choosing database technology until they have working prototypes. Build the feature skeleton first. See how users actually interact with it. Let that usage pattern reveal what data model and database design you actually need.

Deciding the database too early locks you into constraints before you know how the feature will be used. You optimize for access patterns you haven’t validated yet.

Waiting until you have real user interaction gives you the information that matters. Then you can choose the right tool with actual context.

The same applies across engineering decisions:

• API contracts need to be decided early (other teams depend on them)

• Performance requirements need to be decided early (they shape architecture)

• Specific implementation details can wait (you’ll learn better approaches as you build)

• UI polish can wait (user feedback tells you what actually matters)

Deciding the flexible parts too early locks you into choices that made sense three weeks ago but don’t fit what you know now.

Waiting until the last responsible moment means you’re making decisions with maximum context.

What makes this different from procrastination

Procrastination often occurs when individuals avoid making decisions due to uncertainty, anxiety, or fear.

The last responsible moment delays decisions intentionally because you know more information is forthcoming.

The difference is clarity about what’s essential versus what’s iterative.

This framework works when you’re honest about which is which. It’s the “responsible” part of it.

What flexibility actually requires

Staying flexible sounds easy. In practice, it means tolerating uncertainty longer than feels comfortable. You have to trust that waiting will give you better options, even when booking now would feel more secure. You have to know your constraints well enough to recognize what can’t be delayed and you have to move quickly when the moment arrives.

Last responsible moment means compressing decision time once you have clarity.

What this trip reminded me

The best planning means knowing what you need to control and leaving everything else open.

We didn’t plan every detail of this trip. We planned what was essential. Everything else waited for better information. We took action when that clarity appeared, or we were about to cross the line from “last responsible” moment into the danger zone.

The same principle applies to shipping code, building teams, and designing systems. Decide what’s foundational early. Let the iterative decisions breathe until you have real context.

Waiting isn’t a weakness, sometimes it’s the most strategic thing you can do.

Then my hands try to play it, and it sounds terrible.

This gap between what you can hear and what you can execute has been the defining experience of learning guitar. Some passages took me months to get right. Others took years. Many are still works in progress.

Your ears develop faster than your hands, and that is really frustrating. You know it sounds wrong long before you know how to fix it.

Ira Glass talks about this for creative work. Your taste develops before your skill. You can identify great work before you can create it.

Most people quit in this gap. They hear the distance between their vision and their execution and assume they’ll never close it.

Twenty years of guitar taught me something different about that gap.

What the gap actually teaches you

The space between vision and execution does specific work:

• You learn to hear what’s wrong before you know what’s right

• You develop tolerance for iterating on something that sounds bad

• You practice making small improvements without seeing the full result yet

• You build trust that the gap will close if you keep working

That last one is critical. The gap closes slowly. Week by week. Month by month. You can’t force it faster than your hands can learn.

The same gap shows up everywhere

Every junior engineer faces this. They can read beautiful code, they might recognize an elegant architecture, they can spot technical debt.

Their own code looks nothing like what they admire. The gap is massive.

Every new engineering manager faces this. They know what good leadership looks like. They’ve experienced it, they can describe it, they value it deeply.

Their first one-on-ones feel clumsy. Their first team decisions miss important context. The gap between vision and execution is brutal.

Every product builder faces this. You can see the product in your mind. Clean. Intuitive. Exactly what users need.

The first version is rough. The interactions feel off. The gap between what you imagined and what you built is painful to look at.

What the gap actually means

The gap between your taste and your skill is where all growth happens.

Your taste shows you the destination. Your current skill shows you the distance. The gap between them is the training ground.

You close the gap the same way in every domain:

• You keep showing up to the work

• You tolerate sounding bad while you improve

• You make small adjustments and practice them until they’re automatic

• You trust the gap will close even when progress feels invisible

The gap isn’t a sign you’re failing. It’s proof that your taste is developed enough to guide your improvement.

In guitar, I learned this gap never fully disappears. I’m always hearing things I can’t quite play yet. That gap is what pulls me forward.

The same is true for engineering, management, product building, everything that matters.

Your taste develops first. Your skill follows. The gap between them is where you do the work.

That gap is the whole point.

But then, on Wednesday morning, I woke up with the flu and couldn’t work.

No systems. No optimizations. Just forced rest. And turns out I’m terrible at the kind of recovery I can’t control.

What forced rest actually looked like

On Wednesday and Thursday, I stayed home sick. Here’s what happened:

• Checked Slack multiple times even though I wasn’t working

• Felt guilty about canceling meetings

• Kept thinking “I could just do this one thing”

• Never truly disconnected from work mode

The irony hit me hard. I’ve been writing about recovery as a performance multiplier. But when my body actually forced me to stop, I couldn’t let go.

The difference between chosen and forced recovery

When I choose recovery, I control everything: I pick the activities, I schedule the timing… I feel productive even while resting.

When recovery is forced on me, I have no control. No productive activities to point to. Just sitting, being unavailable and incapable of doing what I set myself to do.

That difference revealed something I wasn’t expecting.

The question I kept asking myself: what should I do to feel better?

The better question I realized I should have asked earlier on: what actually makes me feel nurtured right now?

What I learned about myself

I’m much better at active recovery than passive recovery.

When I can do something productive while recovering, I’m great at it. A long run builds capacity, and a museum visit generates inspiration. Both feel like progress.

When I just have to rest and wait for my body to heal, I struggle. The lack of visible progress makes me antsy.

The system I’ve built for intentional recovery works because I’ve learned what activities build my capacity. Long runs. Walking. Reading. Strength training. Meditating. Unstructured time away from screens.

But forced rest requires a different awareness. What feels nurturing when you can’t be productive? What actually helps when you can’t control the timeline?

Building better awareness

When I have energy and capacity: active recovery works. Movement, exploration, and inspiration-seeking.

When my body says stop, maybe the answer is actually resting, not checking Slack, not doing “just one thing, " actually disconnecting.

The system for intentional recovery is important, but so is awareness of when you need a different kind of rest.

Sometimes that’s a long run. Sometimes that’s just lying on the couch while sick.

Both are valid. Both are necessary.

I did it a few days this week. When it worked, it felt really good. Mornings that followed that pattern felt better. The work itself flowed better. By noon, the rest of the day felt like bonus time.

But I’m not consistent yet. Some days, I nailed it. Other days, I didn’t. But now I can see the things that make a good day feel like one.

And this is not about waking up early. It’s about having systems that work and help you create flow and do your best kind of work, and rest in the moments that are best set up for that, too. So, even if you wake up later in the day, I think that applying this concept of working first thing can do wonders: the science of flow says so. So remember: the sequence matters more than the clock.

The three factors that mattered

When the early work sessions actually worked, these three things were present, in this particular order of importance:

• Goal clarity: Knowing the exact task the moment I open my eyes. No deciding what to work on. No wandering. The discipline of waking up consistently and jumping on doing what you said you would depends on having something specific pulling you forward

• Caffeine timing: A v60 with decaf coffee at the beginning to ease into the ritual. Then 2-3 coffees throughout the day, strategically spaced to maximize benefits without the crash

• Sleep consistency: Staying within a 1-hour window every day helps. The same window helps more than just “go to bed early.” Your body knows when to wake up without forcing it

Especially the first one: when I don’t have goal clarity, I can feel the whole thing falling apart. If you start the day without a clear goal you’ll wander into email or Slack, and we all know that’s not where your most productive work happens.

Tactics vs. end goals

So many people think the win is working early. But the real win is the system that makes early work sustainable.

Most people try the tactic and they wake up early. They skip the supporting structure. Then they burn out after a few days and assume early mornings don’t work for them.

The tactic sounds easy. Sustaining it to actually reach the goal requires a different approach.

Still figuring it out

I’m not writing this as someone who has things dialed in. I’m writing it as someone who’s excited to keep going because things feel better to me, and I am excited to share it in case it has a positive impact on somebody else.

The days I get everything aligned, the quality of work is noticeably different. Flow states that used to take an hour to reach happen in the first few minutes. Hard problems feel easier to think through.

The days I don’t, I still get the work done. But there’s more friction. More force is required.

That’s the thing I’m paying attention to: when the system is right for me, everything feels easier.

I’m going to keep going with this and see what else I learn. I’m curious what happens with more consistency.

Three things that inspired me this week

• The Power of Full Engagement: Jim Loehr and Tony Schwartz’s work on energy management keeps showing up in these experiments. The concept that relates with this week’s content: capacity building happens in the recovery periods, not during the work itself. The early morning run isn’t taking time away from work. It’s building the capacity to make the work count.

• Creative Personal Branding: Jürgen Salenbacher’s book on building authentic personal brands. Just started this one today. It resonates with me the concept that great personal brands are not manufactured, but documented. That’s what I hope to do: share my reflections in here and on social media. Not sharing hacks or positioning myself as an expert, just sharing what works for me, and excited to hear what resonates with everyone.

• How to Use Claude Code for Everyday Tasks: I’m defaulting to using AI for more and more things lately, testing the limits of what works and what doesn’t. Some results are underwhelming, but some others change how I work entirely!

The question isn’t whether you can wake up early. It’s whether you’ve built the system that makes a life that is more aligned with your goals feel natural.

Still building mine.

• Ran an easy (conversational pace) long run.

• Wandered through the city (walked 30,000 steps).

• Went to a museum exhibit.

I used to think that these activities were almost as relaxing as lying on the couch watching Netflix. But over the years, I have come to understand that there is a huge difference: Netflix feels like rest, while a Sunday like today feels like capacity building.

The weeks I do it, Monday mornings feel clearer. The work flows better. There’s less forcing it, less friction.

That’s what high-quality recovery does. And I can trace my best periods of productivity directly to protecting it.

The pattern I keep seeing

I’ve run this experiment enough times now to trust my experience.

The weeks where I protect Sunday as a full recovery day:

• Monday starts with flow states that used to take hours to reach

• Hard problems feel easier to think through

• Output quality is noticeably higher

The weeks I skip it because something feels urgent:

• Diminishing returns by mid-week

• More hours produce less output

• I end up running on empty

The difference is measurable. I can see it in what gets shipped, how fast decisions happen, and how clearly I think through work.

What actually works for me

Taking a step back, I think that what works for me on Sundays are: a long physical activity that’s not as intense as what I might do during the week. Seeking inspiration (museums, books, etc.,). Breaking patterns and agendas. Spending more time away from screens. All of it it’s very different from regular cognitive work.

When I run for an hour or two at easy pace, my heart rate variability might dip that night from the physical load. But something else happens: my cognitive system gets a full reset. Different systems working means other systems recovering.

Walking through a museum gives my mind space to wander without directing it. Unstructured time to think. No screens. No inputs demanding a response.

The biggest mistake I used to make

I thought any downtime was recovery.

Watching Netflix after a hard day feels like rest. Scrolling through feeds feels like decompressing.

But I’d start the next day in the same state I ended the last. Sometimes worse because I’d spend so much time in a gray zone: neither fully working nor fully recovering. Thinking about work while watching TV. Never fully off.

That’s not recovery.

What changed my thinking

The Power of Full Engagement is a book that talks about active recovery in such a different way than how I used to understand it.

The concept that stuck with me: knowledge workers have to perform all day, every day. We work all day and treat recovery as optional downtime. Professional athletes do the opposite. Most of their time is training. Capacity expansion. Only a small percentage is actual competition.

What I’m still figuring out

I feel I protect weekly recovery more consistently nowadays. But I’m not as disciplined about daily recovery practices. I would love to explore all different time scales: daily, weekly, monthly, quarterly.

The thing that I’m excited to try now is: waking up early more consistently, jumping in to work in flow and silence for one or two hours before the daily demands get to me. Then taking a break for active recovery before any meetings. Going to the gym, running, journaling, meditating, etc.

Whenever I’ve done this, the results feel good. Flow states come easier. The first few hours of the day feel better. By noon, I feel like the rest of the day is just bonus.

But it’s still not a consistent practice.

What I learned this week

The Rian Doris podcast on recovery gave me better language for patterns I’d been observing. The concept of allostatic load: chronic stress that degrades your ability to access flow states. That’s exactly what I’ve experienced when I skip recovery.

When I don’t clear the accumulated stress, the threshold for flow gets higher. Same work feels harder. More friction. More force required.

Recovery lowers that threshold. Makes flow accessible again.

That’s why my best productivity periods correlate directly with high-quality recovery. It’s not downtime. It’s support.

The question I’m asking myself, and you should, too

Am I treating recovery with the same commitment as the work itself?

Because the weeks I do, everything else gets easier.

Three things that inspired me this week

• How To Actually Recover For Maximum Productivity: Rian Doris breaks down the science of active recovery and why entrepreneurs have the athlete model inverted. The allostatic load concept finally gave me language for what I’ve been observing in my own patterns. When you don’t clear chronic stress, you literally can’t access flow states.

• The Making of a Corporate Athlete: Jim Loehr and Tony Schwartz’s (authors of The Power of Full Engagement) piece on how professionals should train like athletes. Professional athletes spend most of their time training (expanding capacity) and little time competing. We do the opposite. And wonder why we plateau and burn out.

• Tips from the Top: Picked this up at the museum today. It’s career advice from 60+ architects, organized into themes like designing with purpose and learning from other disciplines. Cross-pollination works. Reading how other people think about their craft gives me new angles on my own. Different domain, same fundamentals about building things that matter.

The question isn’t how many hours you are working. It’s whether you’re building the capacity to make those hours count.

Mike

They’d been working alone for three hours trying to recover a system after a critical error. Every method they knew had failed: database backups, VM snapshots, everything in their toolkit.

By the time I logged in, later that day, as it was a bank holiday for me, and I got pinged due to the criticality of the situation, the conclusion felt certain: nothing could be done.

But here’s what actually happened: repeated failure had transformed “I don’t know what else to try” into “nothing else can be done”.

That shift happens so gradually you don’t notice it happening.

The certainty trap

You start with uncertainty. You try everything you know. Nothing works.

• Database backup fails. That’s one data point.

• VM snapshot fails. That’s two data points.

• Every method in your toolkit fails. Now you have evidence.

At some point, without realizing it, the question changes.

From “what else might work?” to “see? Nothing works”, each failed attempt builds a case. A logical, reasonable case in which you’ve tried everything and nothing is possible.

The most dangerous part is that it feels like evidence-based thinking, like you’ve done your due diligence, like giving up is the only rational conclusion.

Fresh eyes, different question

I am not particularly smart, and I definitely wasn’t a subject matter expert about the infrastructure or databases surrounding this crisis.

I just had something the person dealing with the crisis didn’t: no history of failure on this problem.

Fifteen minutes exploring with AI revealed data we didn’t know existed. Within an hour, the “impossible” recovery was done.

The solution wasn’t hidden. It was sitting right there in the documentation. But when you’ve built a case that nothing works, you stop looking for what might.

How many “impossibles” are just failed attempts?

Think about the last time you gave up on something. Was it actually impossible? Or had you just accumulated enough failures to make it feel that way?

The bug you couldn’t fix after trying five approaches, but solved when someone suggested a sixth. The conversation you thought was hopeless after three failed attempts, but there was a breakthrough on the fourth with a different framing. The problem that felt insurmountable until someone who hadn’t tried and failed yet asked one question you hadn’t considered.

We confuse our history with reality. When we’ve failed repeatedly, “I couldn’t” becomes “it can’t be done.” Our experience becomes a universal truth.

The lonely crisis effect

Working alone amplifies this. There is no one to challenge your conclusions, no one to ask “what else?”, and no one around who is free of your history of failure on this specific problem.

Distributed teams make isolation the default during a crisis. You don’t want to bother people. They’re in different time zones. You think you should be able to handle this.

So you work alone. And each failed attempt builds more evidence that nothing else exists. Your track record becomes the only data that matters.

What actually helps

(spoiler alert: it’s not trying harder with the same methods)

• Stop working alone on hard problems: Get someone else involved before your failed attempts become false certainty. Not because you’re not capable, but because a fresh perspective hasn’t built the case for impossibility yet.

• Recognize when certainty feels earned: When you start feeling certain that something is impossible because you’ve tried everything you know, that’s precisely when you need someone who hasn’t tried yet.

• Use tools that explore beyond your attempts: AI can search spaces you haven’t looked in because it has no history of failure on your specific problem.

The goal isn’t to never give up. Some things actually are impossible. The goal is to recognize when you’re deciding something is impossible because you’ve failed repeatedly, not because it actually is.

What I learned

I got lucky this week. I happened to log in without any failed attempts weighing on me. But luck isn’t a system.

We need ways to catch ourselves when our track record starts feeling like the truth. When “I’ve tried five things” becomes “there are only five things to try”.

The actual learning is not how to solve that specific problem in the future; it’s learning to recognize how often we all accept defeat simply because we’ve accumulated enough failures to make giving up feel logical.

And that matters far beyond any single crisis.

Three links that inspired me this week

• Whateverest: A mockumentary about a failed musician in a small Norwegian town caring for his ailing father while making dance videos for YouTube. His philosophy of “No Bad Days” deeply resonates with me: acceptance is finding a way to dance even after facing defeat.

• How to use Claude Code as a Second Brain: Noah Brier talks about using AI to search through personal notes and act as a research partner that asks clarifying questions. The insight that resonates with me, and I have shared before: AI’s superpower is reading and processing everything you’ve thought, not just writing new things.

• Merca Bae DJ Set: Caught Merca Bae live at La Mercè festival in Barcelona last night and still vibing to the beats. Such a good way to replenish energy through good vibes and movement.

The question isn’t whether you should give up. It’s whether you’re giving up because something is actually impossible, or because you’ve failed enough times to make it feel that way.

Mike

I had a recent conversation with someone that was feeling overwhelmed that way. "It's the small tasks accumulating mental burden over time," I shared with them, "it's the stuff that's on the back of your mind that only you know it's there, and that is weighing you down".

I've been here before. We all have. The quick Slack question that takes 5 minutes. The "can you just check this?" that takes 10. The random production issue that needs eyes for 15. The ideas you have that realize are game changers, but never have time to follow through thoughtfully.

None of them make it to Linear. None get discussed in retros. By Friday, you've lost a full day to things you can't even remember.

The invisible workload

Teams track sprints. Estimate story points. Measure velocity.

But they don't track:

• Context switching costs

• Mental residue from unfinished micro-tasks

• The cognitive load of keeping 20 small things in memory

• The energy drain of being perpetually reactive

It's very easy to get pulled into random requests constantly and become reactive by default. Not because you want to be, but because saying no to a 5-minute ask feels petty. That is, until those 5-minute asks compound into burnout.

Why small tasks hurt more than big ones

Big projects get respect. They have kickoffs, documentation, clear ownership.

Small tasks get none of that. They just... appear. In Slack. In passing comments. In "quick favors."

The major migration was stressful but bounded. It had a start, middle, end.

The accumulation of small tasks is unbounded. It never ends. It just grows.

The boundary problem

In those situations, I ask people to surface concerns about unplanned work. But here's the thing that everyone replies at first: each individual task seems too small to complain about.

That's why the most expensive phrase in engineering is "it's just 5 minutes". Because it's never just 5 minutes. It's:

• 5 minutes to understand the context

• 5 minutes to do the thing

• 5 minutes to communicate back

• 15 minutes to regain focus on what you were doing

That's 30 minutes. For a "5-minute" task.

The solution isn't what you think

Most teams try to solve this with process and friction. Route through managers. Add approval gates. That slows down the whole team, and doesn't really change the system that created this in the first place. It also has a nasty side-effect: makes other teams slower, too, and slowly switches the expectations for everyone to move slower and accept process as-is.

The actual solution is to make the tasks visible, not just harder to request. Capture everything, add it to your Linear backlog. Then keep working on it as you were. No approval needed. No complex tracking. Just visibility.

After a couple of weeks, this will show you how many hours of "quick tasks" you have accumulated. Possibly, a big percentage of your capacity that was invisible until now.

The conversation that changes everything

Once you have that data, you can change things:

"We're spending 30% of our time on unplanned micro-work. What should we stop doing to continue this? Or should we stop this to focus on our roadmap?"

That's not being difficult. That's being strategic.

Three things that inspired me this week

• Code From Anywhere with Claude Code (GitHub Action): The future of async engineering. Create a GitHub issue from your phone, tag @claude, and have code waiting when you return. I have been using Codex for this purpose for a while while going on walks, and it feels so magical to come back to my laptop and have several Pull Requests ready for me to review already.

• The Right Time To Read: We expect 5-minute tasks to be trivial, so we never see them compound into 30% of our capacity. It's only when we step back (as Pico Iyer suggests) that we see the whole canvas.

• Introducing Alterego: the world’s first near-telepathic wearable that enables silent communication at the speed of thought. The gap between thought and action is collapsing entirely. We live in the future.

The small tasks won't go away. But they can become a conscious choice rather than death by a thousand paper cuts.

Your energy (and your team's energy) is finite. Protect it accordingly.

This week, I kept thinking about all the APIs I've built on top of over the years and the ones I wanted but couldn’t. Some became career milestones and product breakthroughs, while others made me desperate enough to build alternatives and workarounds.

The difference always came down to one philosophy.

APIs as ecosystem accelerators

Here's what fifteen years of consuming APIs taught me:

Closed platforms think they're protecting value

They limit access, restrict functionality, or charge for basic features. Every interaction feels like a negotiation.

Open platforms create value multipliers

They give developers superpowers. Problems are solved in ways the platform never imagined. The ecosystem innovates 10x faster than any internal team could.

I've watched this play out repeatedly:

• Salesforce opened its API and created a $5 billion ecosystem

• Stripe became the default payment processor through the developer experience

• Twilio turned programmatic phone systems from an enterprise nightmare into a few lines of code

Each platform that trusted developers multiplied their own success.

A rising tide lifts all boats

When developers can build on your platform:

• They solve problems you haven't even identified

• They reach audiences you couldn't imagine

• They create value that makes everyone more successful

I've been on both sides. I've built on dozens of APIs. I know what it feels like when a platform trusts you with real access. And I know the frustration when they don't.

Building for builders

The best API decisions I've seen share these principles:

Generous rate limits

Developers need room to experiment and scale. Artificial scarcity breeds resentment.

Transparent pricing and/or tiering

Clear, predictable steps that scale with success. Published tiers that work for startups and enterprises alike. If, in order to get started, I need to "contact sales", I will do my best to work around needing you, rather than with you.

Real functionality

The same capabilities the platform uses internally. Full access to create, update, and integrate.

Developer experience first

Documentation that actually helps. Error messages that explain solutions. Webhooks that work reliably.

The ecosystem advantage

Companies that embrace open APIs gain something powerful: thousands of minds solving problems in parallel.

No internal team, regardless of size, can match the creativity of a motivated developer ecosystem. Every hackathon, every side project, every integration is free R&D.

The platforms that understand this don't just survive. They become the essential infrastructure in their market.

Three things worth checking this week

• Buffer’s new developer API: We are walking the talk. After years of discontinuing our old public API, we are building one that embodies everything we have learned about developer experience.

• Platform Revolution: How Networked Markets Are Transforming the Economy―and How to Make Them Work for You: This book was my personal bible when I was building Otter, a platform to build custom furniture by connecting consumers with carpenters. It crystallized my thinking about network effects and my understanding of ecosystem dynamics.

• A powerful reminder

  

  ---

May you get the weather you like!

Until next Friday,

Mike

“The job is what you do when you are told what to do. The job is showing up at the factory, following instructions, meeting spec, and being managed.

Someone can always do your job a little better or faster or cheaper than you can.

The job might be difficult, it might require skill, but it's a job.

Your art is what you do when no one can tell you exactly how to do it. Your art is the act of taking personal responsibility, challenging the status quo, and changing people.

I call the process of doing your art 'the work.' It's possible to have a job and do the work, too. In fact, that's how you become a linchpin.

The job is not the work.”

– Seth Godin, Linchpin.

These words have an even more impactful meaning in 2025. AI handles more of our tasks daily. Remote work shows us we're more than our office presence. Job titles mean less when entire industries transform overnight.

Your job is what pays the bills. Your work is what lights you up.

A teacher's job is to deliver the curriculum. Their work might be to inspire curiosity. An accountant's job is to balance the books. Their work might be helping businesses thrive. A designer's job is to create interfaces. Their work might be making information accessible to everyone.

I lead engineering teams. That's my job. My work is creating environments where people do their best thinking and achieve their full potential, both through growing as individuals, learning best practices, and creating systems around them to reduce friction. Sometimes that means writing code. Often, it means clearing roadblocks, designing a strategy, protecting focus time, or celebrating wins.

The magic happens when you stop confusing your job description with your identity. Your value comes from the problems you choose to solve, not the tools you use to solve them. Your meaning comes from the impact you create, not the title on your LinkedIn profile.

To figure out who you are, you need to stop identifying with what you are not.

And then half of the work is already done.

Until next Friday,

Mike

Many folks start with what they have today. They look at current resources, consider existing constraints, and then plan incremental steps forward. This makes perfect sense, and it works.

But I've always found myself starting somewhere else entirely.

The backwards approach

Instead of moving forward from today, I imagine the best achievable outcome—not fantasy, not unlimited resources, not breaking physics, but the best version that could actually exist in the real world.

Then I reverse engineer from there.

This quote from The Imaginary Foundation's mysterious founder has always been on the back of my mind:

"Imagination is the factory that makes legends. It is the beginning of all achievement. To imagine is to perceive many potential futures, select the most delightful possibility, and then pull the present forward to meet it. Imagination has transported us from shivering in dark caves to triumphantly floating above our precious blue earth. It reminds us that reality is malleable and we are the architects of our own fate."

That phrase: "pull the present forward to meet it" captures exactly what this backwards approach feels like.

How working backwards actually works

Here's a high-level look at what I do:

1. Define the best achievable outcome

What would amazing look like if we got everything right? Not perfect, but the best version within reality's constraints.

2. Trace backwards from that outcome

What needs to be true for that to exist? What capabilities, systems, or changes are required? If you walk backward, what does reality look like one step away from the goal, then two, then three?

3. Map the gap from today

Now look at where you are. The path becomes surprisingly straightforward when you know exactly where you're headed.

4. Execute forward with clarity

Each step has a purpose because you know how it connects to the destination.

Why does this create less waste?

When you work backward from a clear outcome, you naturally filter out activities that don't contribute meaningfully to it. You're not experimenting to find your goal; you're building toward something specific and clear.

At Buffer, I practiced this when writing our Engineering Direction. That's why there's not a vague goal like "improve developer experience" but rather "30 people, 300 person impact," all with specific outcomes, such as removing CI/CD blockers and improving our shipping cadence without negatively impacting quality.

Working backwards from that vision revealed exactly which constraints to tackle and which were distractions.

Reality is more malleable than we think

The Imaginary Foundation quote reminds us that we've literally gone from caves to space stations. Not through incremental planning, but through imagination, pulling the present forward.

It applies everywhere:

• Team structures that seem fixed but aren't

• Technical debt that feels permanent but isn't

• Market positions that appear locked but aren't

The distance between where you are and the best achievable outcome is usually smaller than it appears. Most constraints are assumptions, and most barriers are design decisions.

The trick is starting with clarity about where you want to end up.

Three things that inspired me this week

1. "Working Backwards: Insights, Stories, and Secrets from Inside Amazon"

Amazon's entire product development process starts with writing the press release first. Then they build backwards from that vision.

https://www.amazon.com/Working-Backwards-Secrets-Amazons-Insights/dp/1250267595

2. I've vibecoded an app to stop using my phone (that much)

Iñaki Tajes, from calisteniapp, has this amazing video in which he shows how much of a difference vibe coding is having in his workflows as the single technical member in his team, and how it inspired him to create a new app. This is what excites me about AI.

3. The AI-Driven Leader: Harnessing AI to Make Faster, Smarter Decisions

This audiobook I’ve been listening to recently has been the key to empowering me to rely on AI further and further in my day-to-day. Figuring out how to leverage AI to escape operational overwhelm to free your time to focus on strategy, shrink the time it takes from data to decisions, and improve your decision-making are things that any leader and entrepreneur should dedicate time to improving.

https://www.amazon.com/AI-Driven-Leader-Harnessing-Smarter-Decisions/dp/B0DB8QL3ZK

Who would have thought that "doing it backwards" would be so positive?

Until next Friday,

Mike

I keep noticing a pattern: many folks, engineers or otherwise, try AI tools, get mediocre results, and go back to doing things the old way because it's faster and better. They're right, in a way. If you approach AI as a one-shot replacement for your existing workflow, you'll probably be disappointed and will feel slower.

But what if we see beyond that and flip our understanding and intention?

The AI-first mindset

For the past couple of months, I've been experimenting with a different approach. Before starting any task, I ask: how can AI help me do this faster and better?

Not to do it for me. Enable me.

This shift led me to share my recent experience with Claude Code, where I used it to rapidly iterate on Buffer’s engineering direction. The breakthrough came from enhancing my entire personal knowledge management system. Being able to access, navigate, and synthesize years of thinking at a completely different speed and depth.

When you approach tasks with AI as your assistant rather than your replacement, and lead with curiosity, something shifts. The friction disappears. Even when the output isn't immediately perfect, you learn something. And that learning compounds.

I believe so strongly in the impact of embracing AI to assist with everything that I made it one of our strategic themes in Buffer's engineering direction. Not as a nice-to-have, but as a fundamental shift in how we work. We must become AI-native in order to thrive.

The intention behind these Friday notes

I'm launching this newsletter as a weekly practice in making explicit my reflections and learnings. Each Friday, I step back from the immediacy of my work to ask: What is different this week? What are the learnings? What is the highlight?

Some context on the format: this is an experiment in sustainable reflection. No rigid structure, no content calendar, no engagement metrics to chase. Just a commitment to pause, synthesize, and share.

The goal is to document the intersection of:

• Building products and leading teams

• Experimenting with AI as a productivity multiplier

• Finding sustainable rhythms in ambitious work

• Turning chaos into clarity through systems thinking

• Keep growing and improving in all facets of my life

Think of it as that Friday afternoon moment when the week's chaos morphs into something coherent. That's what I'm aiming to capture and share.

Three things that inspired me this week

• Carl Barenbrug's workspace setup: after working from a 14" laptop for over a year and a half, Carl's minimalist workspace tour hit differently this week. His attention to detail and intentional constraints remind me that environment shapes output. Time to invest in a proper setup again.

• "The days are long but the decades are short": Sam Altman's 36 life reflections resonate deeply with my own 36 years on this planet. I return to this piece regularly. Even the smallest things compound when you zoom out into decades.

• "The Fastest Way To Fix Your Broken Attention Span": This video on Attention Deficit Trait (ADT) versus ADHD hit home. Protecting our ability to find flow in an increasingly fragmented world is a challenge for all of us. Worth the 11 minutes for anyone feeling scattered.

Here's to more experiments, more documentation, and less perfection.

Until next Friday,

Mike