A change in one area of life cascades into other areas – popularized by Charles Duhigg in The Power of Habit. If we think of this strategically, then we’d want to identify which areas, or habits, have the greatest blast radius and highest number of potential cascades.

Our mindset, self-image, and habits are upstream of almost everything else in life, so doing something that improves these will also improve everything downstream. Focusing here will provide the highest leverage.

So what are some good candidates? Reading, meditating, therapy, and journaling are some logical ones. They can all influence your mindset, worldview, and self-image, which itself can lead to better habits. But they are not the most reliably powerful drivers of change.

The issue is that they are exclusively cerebral exercises with low physical effort. You don’t need to push yourself to read 10 pages or to sit in silence for 10 minutes. You don’t break a sweat when writing on a piece of paper or talking to a friend or therapist. This lack of effort in the physical realm leaves a lot of downstream potential on the table.

Furthermore, confidence and self-image is also influenced by how we look, so we need to pick something that spans the cerebral and physical domains. More domains, more blast radius.

And that something is to get fit. It’s the highest leverage thing you can do for yourself and for others. Not only does pushing yourself physically cascade further and in ways that pushing yourself mentally doesn’t, but it also lowers the activation energy of everything else in its radius.

Consider that there are many out-of-shape bookworms but not many jacked people who don’t read. The same goes for meditating and journaling. As for therapy, things often don’t even get there because hard exercise dissolves many of the common problems like stress and anxiety. This tells you everything you need to know about which activity is upstream from the rest.

Getting in shape is as much a mental and spiritual exercise as it is a physical one. It’s the gateway drug to self-improvement across every area of life.

Getting fit makes you multidimensionally jacked.

Physically jacked is the baseline – this, and “being good for you” is what’s expected of getting fit, and most people think that’s it. They don’t know what they don’t know, so they underprice its value. Let’s see how else it changes you.

Emotionally jacked

You’ll be faced with countless physical and mental challenges in the gym. You’ll tweak a muscle, get recurring pains, or hit plateaus – each is a setback you overcome and become more resilient and perseverant. Your threshold for what counts as a problem in day-to-day life goes way up, and so things that used to bother you no longer register.

You also become more confident – and no I’m not talking about feeling better because you look better. I’m talking about that foundational confidence that comes with beating all those obstacles and knowing you can handle whatever is thrown your way. You build an undeniable body of evidence that forces your brain to rewrite your self-belief.

Intellectually jacked

To reach your fitness goal you’re forced to learn new things and try new things. You might start with nutrition and energy balance – what causes weight change? From there you might make some changes to your diet, and once it’s habitual explore recovery. Maybe you’ll start to go to bed at a more regular time, and then move on to learn about hypertrophy, then supplements, different techniques and adaptations, and the list goes on.

You learn to become proactive and high agency. You identify your current knowledge gaps and seek to fill them. You become more curious because you’re seeing how much it’s improving your life. You experiment and become more adaptable, adjusting your plan based on the data that you are seeing – either you’re getting closer to your goal or you’re not. You learn how to learn, and this skill is universally transferable.

Spiritually jacked

Along the way you’ll naturally be more inclined to explore meditation, breath work, yoga, ice baths, reading, or journaling. Each of these helps inspire a deeper connection with yourself, with life, and with a higher purpose.

I’ve never met someone who started working out because they wanted to “get closer with themself”. But inevitably, somewhere along the journey, it happens anyway. This is the cascade at work.

Financially jacked

No, going to the gym isn’t magically going to make you a millionaire – but many disciplines in building your physique apply to building your wealth. Specifically, that of consistency, compounding, and risk management.

Consistency – this one is obvious, you need to show up and get your reps in. You can’t reach your body goals by going to the gym or hitting your calorie target when you feel like it, the same way you can’t reach your financial goals by investing or following a budget when you feel like it. Fair-weather is better than nothing – at least you are lifting weights and investing something – but nobody has built anything they are proud of by doing it only when they felt like it.

Compounding – when you first decide to get fit, you might start with the basics of lifting and eating at a calorie target. Over time as these start to become habitual, you’ll reach the next level – perhaps you’ll fix your sleep schedule or start to hit specific fat/carb/protein targets, in addition to your calorie target. Each small improvement eventually compounds into a whole new lifestyle, and it’s this new lifestyle you are creating which brings you to your goals. If you were already the person capable of reaching your goals, you would have already achieved them. This lived experience with compounding makes you understand its importance in your finances as well: money makes more money, and little by little, a little becomes a lot.

Risk management – you want to stay in the game by not kicking yourself out of the game. Ego deadlifting and blowing your back, using steroids or drugs with long term consequences, or simply pushing too hard too fast and getting injured. All of these are poor risk management. It’s better to go 80% hard year-round than 100% hard for a month and then be taken out by tendonitis for 3 weeks.

Investing is the same: rule #1 of investing is to never lose money, and rule #2 is to never forget rule #1. So many people blow their portfolios taking catastrophic risks and set themselves back by years. I almost did this twice and never again did I put 40% of my portfolio into weekly and 0DTE options (if you don’t know what that is, just think of it like the stock market equivalent of a lottery ticket). You can’t play a game you’re no longer in.

Jacked in your career and personal life

The intellectual, emotional, and spiritual gains spill into the rest of life as well. When you’re more confident, you’ll speak out and share your ideas more. When you’re more resilient, you’ll handle difficult conversations and setbacks better. When you’re connected to a sense of purpose, you’ll be more at peace with yourself and the world.

“What if my life is already good?”

It will become even better. I’m not saying fitness should take priority over everything else in life, just that the multidimensional gains from fitness will make you strictly better off in every scenario, all else equal. Happy to debate this.

Nothing starts with a grand plan

It’s totally common (and perfectly acceptable) to start with shallow goals. In highschool, I just wanted to look good and get girls. That’s literally it. I didn’t care about the health benefits, I didn’t care about the character development – in fact, I didn’t even know what character development was. I lacked the awareness that working out did anything other than make you healthy and your muscles bigger.

Then the cascade happened.

Within a few years I began meditating, then tracking my macros, then sleeping at similar times each night. Then I started listening to podcasts, reading, journaling, and tracking my finances. I became more confident socially and in private (like believing in myself more and taking more chances).

I expected none of this on day 1. I didn’t know what I didn’t know.

The physique is nice, but it’s nothing compared to everything else that came with it. Who knew that such a shallow goal can change your life.

Subscribe now

Well that was me last summer. I was sitting at my desk and my mind felt like a bloated todo list. It was taking more effort to remember, organize, and sort through all the stuff rather than just doing the stuff. I also felt like I lost my sense of direction – what am I working towards? What do I want? What were my goals again?

It felt like my days would end and I wouldn’t really know what I did, or if what I did was even important and taking me somewhere. And if it was taking me somewhere, did I even want to be going there in the first place?

So I grabbed a pen and wrote down every task, todo, and project I could think of. Reach out to so-and-so people, complete this Excel course, figure out when I’m going to visit my parents, finish my M&A class assignment, research my shortlist of stocks to buy, build my next workout plan, do interview prep, fix the bug in my internship project.

One brain dump and 30 minutes later, it felt like I had the mental equivalent of finally going to the bathroom after holding it in for a while.

The math problem

Take the most complicated math problem you’ve solved. You probably didn’t do it in your head. Now consider that your life is infinitely more complicated than that math problem. You didn’t solve that one in your head, so don’t try to solve this one there either.

Your working memory only holds a handful of things at a time. Add another, and something falls out. So get it out of your head and onto paper where the solving happens. You can’t manage a pile you can’t see.

The exercise

We’ll call this a spill and sort – you’ll see why in a sec.

Spilling

Sit somewhere quiet with a notebook or a sheet of paper (I used a 3x5” index card). It helps to first close your eyes for a couple minutes and just exist – notice the sounds around you, the sensation of the seat beneath you, and your belly or chest expanding and relaxing with each breath. This helps slow your mind and put it into a place of calm so it can work better.

Then think of any tasks, decisions, goals, ideas, appointments, projects, events, people you need to get back to, things you want to try, and yes even the stuff you’ve been avoiding for a month, and write it all down. Don’t filter, go in any order, or organize anything yet, just write whatever comes to mind, whenever it comes to mind.

Sorting

Once you’re done, it’s time to start grouping. First look at everything you wrote and absorb it for a few moments. Don’t think about it yet, just notice what’s there.

Then draw a circle beside items that feel similar, a square for ones that feel like a different category, a triangle for ones that feel different still, and so on. Don’t decide categories ahead of time. We are vibe-grouping right now – the categories will emerge naturally.

Now you can start labeling things based on vibes – health, work, family, trading, building, school, whatever. These are your categories. Some large categories can even merit sub categories, like diet and exercise within health.

Once you do this, you’ll see everything that was weighing on your mind and can take this wherever feels right for you: prioritizing tasks based on urgency? Go for it. Organizing based on importance? Sounds good. Removing stuff you don’t actually care about? Amazing.

What feels right will come up naturally because you made a move forward and the map has changed compared to when you started. It feels like playing a strategy game with the map blacked out, then the fog of war lifting all at once – you can finally see the whole game and know exactly what to do and where to move next (not to mention your brain also feels 10 lb lighter).

You might even notice a disconnect between what you think is important to you, vs what your time and energy are saying is actually important to you. If you do, sit with it – it’s usually pointing at something deeper (mine pointed at a career I was chasing but didn’t actually want. We’ll get to that).

Subscribe now

The next day

The morning after a spill and sort is the best: you wake up and you just know what to do. You don’t waste time or mental bandwidth on deciding because you already have the clarity from the exercise. Your job now is to just execute.

Where this exercise becomes really powerful though is in the downstream effects.

When you just “know what to do”, it makes it easier to start because it removes the friction of figuring it out first. Once you get started, you now have momentum, and the momentum carries forward. It also clarifies what to do next, because the next step only reveals itself after you take the first one.

Next thing you know, you just cranked out 2h of deep work and did more in that time than what feels like all of last week. And to top it off, you know that everything you are working on is the right thing to be working on because you already clarified what matters to you from the exercise.

Now you can go to bed at night knowing what you did, and that what you did is moving you closer towards what you actually want to move towards. It’s a great way to end the day.

What it looked like for me

Reading about the exercise doesn’t seem revolutionary on the surface. It’s because it’s not, it’s quite boring actually: you’re literally just writing your todos on paper, grouping them, and maybe getting rid of some stuff. But boring works – allow me to share what happened to me.

The last “big” time I did this was summer 2025, midway through my MBA. I felt like I didn’t know what I wanted – I started the program wanting investment banking, and everything I did was in that direction: electives, case competitions, networking, even my internships.

So naturally when I did the spill and sort, a lot of my todos were investment banking stuff: financial modeling practice, networking with bankers, an Excel course, stuff like that.

But the things that came out which I felt more excited about were all tech and AI related: experimenting with different LLMs, building workflows, and overall exploring what I could do with AI.

I didn’t make any big decisions that day, but I felt the cracks forming in what I thought I wanted vs what I actually wanted. I decided to deprioritize some banking stuff in favour of some AI stuff. I was following my energy.

A few months and a mid-MBA crisis later, I had discovered that I do not in fact want a career in investment banking because I love tech too much (and trading, which banking would restrict me from). I still think IB is really cool, and maybe in another lifetime that’s what I’ll be doing, but not in this one.

That’s why this exercise can be so powerful. At worst, it helps you organize and prioritize your mental todo list to execute more effectively. But at best, it can make you slow down, do some soul searching, and make you see with your own eyes where your energy is pulling you.

The daily system

This exercise isn’t meant to be done every day. It’s more useful when something shifts, or when you feel like you’ve lost clarity on where you’re going, or at the start or end of a new chapter.

This exercise is strategic and helps clarify the direction, but underneath it you need a tactical layer. There are many ways to do this, but for me I like the Power List by entrepreneur Andy Frisella.

This is a daily five-item list of critical tasks that you need to finish for the day. Get all five done, and you win the day. Win more days than you lose in a week and you win the week. Win more weeks than you lose in a month and you win the month, and so on. Every night I decide what my critical tasks are for the next day, which are generally based on what I learned from the spill and sort.

This is a system that requires maintenance! If you stop, you’ll slowly lose clarity and you’re back to a head full of static. Except now it’s static plus a touch of guilt from losing the clarity you worked hard to gain.

Clarity on paper

Lots of advice tells you to think more about what matters, and that’s very important. But it’s hard to think your way out of a head that’s too full to think. So you empty it first, and then you think.

If you try this, I’d love to hear what you discover.

Thanks for reading.

– Melad

Subscribe now

He noticed investors would sell their winners and hold their losers, locking in a gain about 50% more frequently than locking in a loss. Interestingly, the stocks they sold kept going up and outperformed the ones they held on to, meaning the pattern wasn’t just emotional, but sub-optimal.

This behaviour is called the disposition effect, and the cause is loss aversion: selling a losing stock means turning an unrealized loss into a realized one, and the emotional pain of doing that outweighs the logical math of cutting it. Research since then has shown that even experienced investors who can explain the mechanism in detail still hold losers too long.

Knowing the right move and executing the right move are two very different skills, and nobody is immune from that gap.

The visibility trap

Loss aversion gets triggered by the metric you can see. The P&L, the scale, the effort of dieting, the cost of hiring a mentor, or the reflection in the mirror. These are the metrics your brain weighs heaviest because they’re immediate, concrete, and right there.

But the metrics that matter are usually invisible. Muscle built under a layer of fat, skills acquired through losing trades, competence gained in the years before you’re actually good at something new.

Your brain defends the metrics you can see at the expense of the ones that matter.

Subscribe now

My first bulk

Seven months ago I started my first real bulking phase after years of “lean gaining”. At 169 lb and 12% body fat, I decided to aim for 200 lb at 8%. After a few years of lifting (and if you’re natural), you can’t build muscle and lose fat at the same time, so you cycle: bulk, cut, and repeat.

I’m now 182 lb at 15%. Of the 13 lb I gained, 6 are muscle and my lifts are up across the board. On paper, this is working how it’s supposed to.

But mentally, the paper left me unprepared. I felt fat and gross. I no longer saw the abs, vascularity, or muscle definition I got used to over the years. By most visible standards I looked worse, even though the invisible metric (muscle) was moving in the right direction.

“It’s so cliche, but it really is one of those trust the process things”

I was chatting with a friend recently who happens to be one of the biggest guys at my gym. I asked about his experience with bulking and cutting and he talked about the middle phase, when you’re a month or two into a cut and you look smaller, feel weaker, and not that lean yet either.

Then when you bulk, you’re already lean from your cut and you look bigger almost overnight from the new calorie surplus, making you feel like a king. But then growth slows, you stop being lean, and it starts to look like you’re going backwards.

I realized I had developed an identity around being lean over the last 10 years and losing that visible marker during my bulk didn’t just feel like a regression, but also like losing a part of who I was.

People quit, switch plans, or convince themselves there’s a secret when they don’t see the results they want quickly enough, and that usually happens in the middle. “It’s so cliche, but it really is one of those trust the process things” he said.

This intellectual vs experiential gap doesn’t get captured anywhere in a plan.

The trade

The mental game shows up in trading too, but faster.

I’ve had countless trades where I was almost stopped out, the position recovered to break-even, and I closed it for a small win rather than letting the plan do its thing. Many of those trades would have worked, but I didn’t want to feel the pain of almost losing again so I took the small win only to watch the stock hit my target without me.

A novice trader can be given the exact plan of a professional and still lose money, not because the plan is wrong, but because the years of mental skills developed to execute the plan are not included in the handoff.

The ability to take a loss, and another, and another, and still put on the next trade without deviation is what the professional can do that the novice cannot.

The actual skill

The skill isn’t in finding the perfect plan, but in detachment from the visible metric. You learn to watch it move wrong and tolerate that pain without intervening because you trust what’s happening underneath.

That’s what you really build as a trader, the ability to let wins and losses register as just lines and numbers on a screen. Once you get there, the plan just runs itself. And the same is true everywhere else.

You only build detachment through the reps of watching the visible numbers get worse before they get better. No amount of reading or preparation will compensate for the practice of lived experience.

Next time you catch yourself in doubt at something you intellectually know should work, don’t ask whether you’re following the plan right – because you probably are. Ask instead whether you’re falling for the visibility trap: too focused on the metric you can see, and not focused enough on the one that actually matters.

Share

But yolo, if I was gunna learn it one day then I’d have to start somewhere, so I downloaded it and told myself I’ll figure it out.

It was all in my head

First thing I realized was that it’s actually not that scary. Sure, it runs in your terminal, but interacting with it is very much like chatting in the normal Claude web app. It was only scary because I made it scary in my head without even trying it (how often do we do that in our life?).

I told it that I’m new to Claude Code and have basic programming knowledge (functions, objects, separation of concerns, etc), with the goal of learning more about architecture and best practices. Then, list out some simple apps I can build that will teach me about foundational skills, while guiding me through why it’s doing things one way vs another.

It gave me the usual suspects: a notes app, a calculator, and some to-do list, which are technically useful for learning state management and other things, but useless for actually building something I’d ever open again.

So I ignored all that and told it I want to build a neighborhood finder for Airbnb.

The problem: Airbnb’s search sucks for this

I go to San Francisco semi-regularly. The first few times I had to do the same annoying research: which areas are close to the gym, grocery stores, and whatever event I’m attending? And which areas should I avoid because they’re sketchy?

Airbnb just gives you the pins on a map and calls it a day. So now you have to open up Google Maps and see where the amenities are, and then hop on to Reddit to see what the good and bad parts of town are. Takes maybe 10 extra mins.

So why not build the tool that does this for me?

The fix: Airbnb matchmaker

I’ve never built a proper frontend + backend app before, so I told Claude Code to “walk me through building this. I want a separate frontend and backend, so lets build the backend and tell me how to connect it to my frontend in Lovable. Explain what you’re doing and why.”

Here’s what I used:

• Lovable for the frontend (user inputs city + preferences)

• Claude Code for the backend (processes the request)

• Firecrawl API to scrape Reddit for neighborhood insights

• Google Maps API to find gyms, grocery stores, etc.

• Vercel to deploy it (which also forced me to learn a bit of GitHub)

The workflow was simple: I’d tell Claude what I wanted, and then it would build it and explain to me what it did. It was teaching me architecture, API integration, and deployment as we went. The whole thing took maybe 2-3 hours spread across a few days.

A side-quest gamechanger

Before getting into what I built, let me share a quick tip I learned that made this whole process way faster and easier.

If you’ve been using AI for a while, then you know they are only as capable as how well you prompt them. Garbage in = garbage out applies well here.

But as humans, we subconsciously default to what’s easiest. It’s not the end of the world, but typing can create enough friction that if we have for example 200 words in our brain that we want to tell the AI, then we might actually end up typing only 150 out of convenience.

So I started using dictation on my Mac, and prompting became so much easier. I could just talk and give way more detail in probably a quarter of the time it would have taken me to type.

This is a low-hanging fruit that I haven’t seen many people do. If you’re still typing long prompts, you’re doing it the hard way.

Subscribe now

The result: it works (mostly)

The tool works like this:

1. You enter a city

2. You describe what you want (e.g., “close to a gym, grocery store, not sketchy”)

3. The backend scrapes Reddit and queries Google Maps

4. It returns a list of neighborhoods that match your criteria

5. You then find an Airbnb in that area

On the main page you add city and preferences.

Here it shows the four Crunch locations in SF, and avoids an area called the Tenderloin which is mega sketchy at night.

It’s not perfect, for example sometimes instead of specific areas it will literally just say “Toronto” or “Waterloo” and draw the circle over the middle of the city.

But even though it only kinda works, it accomplished its goal of teaching me a bit more about something I didn’t know, which was building a backend, deploying it, and hooking it up via an API to a frontend – all in a few hours!

What I learned

1. Build something you’ll actually use (or solve a problem you personally experienced), not what tutorials tell you to build

The notes app would’ve taught me some concepts, sure, but I probably would’ve quit halfway through because I didn’t care about it. The Airbnb tool was technically harder, but I stayed motivated because I wanted the result.

2. AI is more than just a code generator, but only if you ask

You can build amazing things with AI, but it’s not going to explain to you why it did what it did – it will just tell you what it did. So just say “teach me while we build this” and it’ll make the whole process way more valuable.

3. Speech-to-text is a cheat code

Do you want to automatically make your prompts better AND save time writing them? Then this one’s for you. Use dictation mode on your Mac (or try Wispr Flow, it’s even better and you get a month for free with this link 🙌) and experience the difference.

4. Ship broken stuff

If you’re just starting off learning, you need reps. Who cares if what you made isn’t polished – if it’s 70% of the way there then it’s good enough and it has taught you >90% of what you need to know. The last 30% will take 10x longer and only teach you marginally more, making the learning-to-time ratio not worth it.

The takeaway

You don’t need to know how to code or even be skilled at using AI to build something useful. You just need to:

• Pick a real problem you actually have (or something you’re curious about)

• Tell the AI what you want to build and ask it to teach you

• Use speech-to-text so you’re not wasting time typing

• Ship it even if it’s imperfect

If you’ve been sitting on an idea for whatever reason, here’s your homework: open up Lovable, Bolt, Replit, or Manus (it literally doesn’t matter, just pick one), turn on your speech-to-text, and just start talking about an idea. If you don’t know what to build, say “I don’t know what to build, so ask me a couple questions to help brainstorm”. Once you have an idea, ask it to teach you along the way.

You’ll be surprised at what you’ve built in an hour.

Ranch-style homes exploded in California after World War 2. They’re not inherently better, they just emerged when specific conditions aligned: cheap land, affordable car ownership, and a climate perfect for indoor-outdoor living.

Today, most Californians couldn’t tell you why ranch homes dominate. They’d just say “that’s what homes look like here”. The environmental conditions that created the default have been forgotten – what was once a choice shaped by temporary circumstances now feels like natural order.

You’re the same way.

The role of environmental luck

In Outliers, Malcolm Gladwell tells a story about Bill Gates most people don’t know. We think of him as a genius who built Microsoft through brilliance and hard work, but Gladwell shows how environmental luck stacked in his favor.

His parents were upper-middle class, father a prominent lawyer, and mother from a banking family. They sent him to Lakeside, an elite private school in Seattle. In 1968 when Gates was in eighth grade, the Mothers’ Club spent $3,000 from a rummage sale to get the school access to a computer. Most colleges didn’t have computer access in 1968, let alone middle schools.

And the timing mattered too – Gates was born at the right time, in 1955. Too old, and he’d be employed at IBM making mainframes when the PC revolution hit in the mid-1970s. Too young, and he’d still be in high school when the opportunity opened. Gates was college-age right when personal computers became viable, and he’d already logged 10,000+ hours of programming by then.

Strip away the well-connected parents, the private school, the one-in-a-million computer access, and the perfect birth year, and there’d be no Microsoft.

Your strengths work the same way. Maybe you’re “naturally good with people” because you moved schools three times and had to keep making new friends. Maybe you’re “analytical” because you had a teacher who constantly asked you “why?”

These feel like inherent traits, but they might be environmental defaults you’re mistaking for identity.

And your weaknesses too:

• “I’m not good at public speaking” might mean “I’ve never been in an environment that required practice”

• “I’m not that creative” might mean “I was never in an environment that encouraged experimentation”

• “I’m not good with money” might mean “I grew up in a household that never talked about finances”

Both your strengths and weaknesses might be accidents you’ve mistaken for who you are.

Subscribe now

When defaults become visible

I’ve been working out for over 10 years and meditating daily for almost as long. Through these, I’ve built discipline, perseverance, and a cascade of other habits. I could say that these were conscious decisions I made, which would be true, but I also got lucky.

In highschool, some of my friends happened to be into fitness. They worked out, so it rubbed off on me. Without making those friends in grade 9, would I still have the same habits and resulting traits today? Traits that I could credit to my character might just be environmental defaults I was lucky to inherit from the right people at the right time.

Redesigning Your Defaults

Try this: List just 3 things you think describe you. These could be:

• “I’m indecisive”

• “I’m easily distracted”

• “I’m bad at networking”

• “I’m a confident speaker”

• “I’m disciplined about fitness”

Now for each one, think about what environment created it and whether that environment still exists.

• “I’m indecisive” → Parents who made all the decisions for you.

• “I’m easily distracted” → TV always on in the background growing up.

• “I’m bad at networking” → One awkward career fair interaction at 19.

• “I’m a confident speaker” → Having to speak up at large family dinners to be heard.

• “I’m disciplined about fitness” → Friends who worked out in high school.

Most of these environments are gone and you’re running on outdated defaults.

For the ones that don’t serve you anymore, design one specific environmental change that nudges you in the right direction.

• “I’m indecisive” → Remove choice paralysis by creating a default (meal prep same lunch M-F).

• “I’m easily distracted” → Put your phone out of reach (floor, behind laptop, another room) when trying to focus.

• “I’m bad at networking” → Put yourself around people – go to the same coworking space once a week.

Like ranch homes in California, most of what you think is “just who you are” is an inherited default from a past environment. Your weaknesses aren’t inherent flaws, but a lack of exposure. Your strengths aren’t purely inherent either, but environmental luck you built upon.

Once you see your identity as a collection of defaults you can override, you become the architect of your environment rather than its recipient. The more intentional you are, the more opportunities will find you.

As I was just sitting there letting my mind wander, my thoughts at some point landed on AI: I've heard of all these different tools – but how do you actually build with them?

Lovable, Bolt, Replit, Cursor, LangChain, n8n, Next.js, Agent Builder, Antigravity. I kinda knew what each one did in isolation, but had no idea how they fit together to actually make something.

Turns out, that mental question mark was keeping me from starting anything.

But a month later, I had a live product.

How I Got Unstuck

Right there on the bus, I opened Gemini and just told it I was lost:

I'm a beginner at building apps with AI. I've heard of tools like Lovable, Bolt, Replit, Cursor, LangChain, n8n, agent builder, Antigravity and Next.js, but I only have a high level idea of what they are, but in isolation. Idk how they fit together to actually build something.

Can you put on your expert AI developer and app creator hat and explain each tool in < 100 words? Then, given my goal of building a simple app, explain how they work together (e.g. tool X is the "brain" of the app, tool Y is the front end which takes/shows data from X, tool Z does this, etc).

Ask me clarifying questions if needed.

That’s it. I asked it for help and once it broke it down, things started clicking. I realized it wasn't as complicated as it seemed.

Then I asked for a few project ideas to get started with.

Why Drafted?

Gemini suggested a few options based on what it knew about me. One of them was a tool to turn blog posts into LinkedIn posts.

This one clicked. I like to write, and when I finish a post I usually make a shorter version for LinkedIn anyway.

I would fire up Claude, go to my project folder with custom instructions (including samples of my writing style), tell it to turn my blog into a LinkedIn post, and finally spend time tweaking it to sound like me.

I figured why not turn this process into a simple app? It would be easier to set up, it would be something I could actually use, and I would learn in the process.

So I started there.

Enter Lovable

For the platform, I went with Lovable. I’ve heard it’s a simple and pretty no-code tool, and they had a 50% educational discount. Could I have used Replit or Bolt instead? Totally, but I decided to just pick one and go instead of spending more time comparing options.

Gemini then gave me a starter prompt for Lovable. I pasted it in, and then I saw what Lovable made.

That’s when it felt like something in my brain unlocked. I looked at what it made and thought: This is sick, I can do this.

The next step only reveals itself after you take the first one.

If you sign up for Lovable using this link you will get 10 extra credits – and if you end up publishing a website, I will also get 10 credits 😁

What Surprised Me About Lovable

How easy it was.

The UI was clean and took minimal effort to start using. It handled database stuff, and I could add AI features inside my app without touching any backend API stuff which I thought was pretty cool – it just took care of everything on its own.

Were there frustrations? Absolutely. Sometimes it wouldn't follow instructions, or it would "fix" something that wasn't actually fixed and I'd burn credits asking it to try again. But overall, solid experience.

The whole thing took me about a month from start to finish, spending a few hours a week on it.

What Drafted Actually Does

Drafted takes a blog post, brain dump, or half-baked ideas and turns it into LinkedIn-ready posts.

Here’s a quick demo:

You put in your content, it generates three variations, and you can edit them. The edits you make can then train the AI for your next generations, so it automatically learns your voice over time.

Could you do this with Claude or ChatGPT? Totally, that’s what I used to do, but there's friction. You can either create a project or folder with custom instructions, or you can re-prompt it every time. And if you want it to learn your voice, you could add some past content you’ve created into the project files (or paste it into your prompt each time) and instruct it to write that way.

Drafted removes that friction. Similar outcome, but less work and technical know-how.

What I Cut (And Why)

I started with a longer feature list. Here's what I parked for v1:

• Generation history – Each generation creates 3 variations, so this would fill up fast. And what's the value of a history if you already have a favourites section? I couldn't answer that clearly, so I cut it.

• Analytics/gamification – Fun, but not core to the job-to-be-done.

• AI pre-training – This would let the AI learn a user's writing style before their first post generation, but it’s more of a nice-to-have considering this feature already exists when you edit the generated posts.

• Google sign-in – This was taking too long to implement without creating a sketchy-looking URL (apparently I have to upgrade to a paid version of Supabase, the database Lovable uses). Not worth the delay/cost. If people actually use this, I'll revisit.

The filter I used: Does this help someone turn a blog/brain dump into a LinkedIn post faster? If not, it's v2.

What v2 Could Look Like

If there's traction, here's what I'd consider improving/adding:

• Chain-of-thought prompting – Right now the post generation prompt is one-shot including the user settings and training data. To make this better, I can make the prompt execution be multi-step. For example: understand the content → choose the best LinkedIn post style → write the post.

• Comprehensive writing style analysis – Currently it uses the saved edits as examples to train the AI, but a more comprehensive approach would be to perform a full analysis with its own prompt (using chain-of-thought like above). It would look at word choice, sentence structure, flow and rhythm, formatting style, persuasive techniques, and emotional appeal, to create a "rulebook" of how the user writes.

• AI pre-training – Option to paste some of your past content on first use (maybe prompted via onboarding flow) so the AI learns from it and performs the style analysis from above.

• Ability to paste YouTube, X, Instagram, etc URLs – Right now it only works on blog post URLs and brain dumps. There’s no reason it can’t be made to work with these links too though.

• Analytics dashboard – Total generations, usage streaks, content sources used (e.g. blog posts, social media, YouTube, brain dumps, etc), maybe some gamification to increase stickiness.

• Google sign-in – Proper implementation without the sketchy URL, and a lower-friction (and probably more perceived trustworthy) option than creating a profile using email.

• Generation history – But only if I can figure out a clear use case beyond "nice to have".

The Real Challenge Wasn't Technical

The hardest part of building Drafted wasn't actually the code or the tools. It was me.

At one point I spent over an hour designing the logo. A logo. For a tool that wasn't even launched yet.

You could argue branding matters – and I believe it does – but not at this stage. Not before anyone has even seen the thing. The priority must be launching. Branding can come later (to be fair, I think the logo is pretty sick).

That's perfectionism. And perfectionism is procrastination in disguise.

I've seen this pattern in myself before: When I was working on a startup a while back, I'd constantly tweak and polish instead of shipping. I'd tell myself I was "making it better", but really I was just avoiding the fear of putting it out there.

At its core, procrastination is the avoidance of some kind of fear: Fear of rejection. Fear of judgment. Fear of not being good enough.

I haven't fully gotten over this, but I've gotten a lot better at catching myself doing fake work instead of real work.

You're not gunna get strong by watching videos about how to work out. You have to actually go to the gym, lift the weights, mess up a few times, look silly, feel uncomfortable, and only then will you start to get strong.

Shipping is the same – you only get better at shipping by shipping.

The Backwards Motivation Trap

Many people wait to feel motivated before they take action. That’s wrong, and it’s counterproductive.

Motivation comes from action, not before it.

It's like getting out of bed early: You're not always going to feel like it when the alarm goes off, but once you're up and moving, you're fine. The resistance was only at the start.

I wasn’t motivated to build anything when I was on the bus. I was literally doing nothing!

But then I had a question, and then I wanted answers. When I opened Gemini to ask, that tiny action alone was enough to create momentum, and then that momentum created motivation.

Action creates momentum, momentum creates motivation, and motivation creates more action.

Drafted Is Now Live

You can find it here: https://drafted.meladkhajepour.com/ (I didn’t get a separate domain for it, maybe that can also be for v2 if there’s traction 😛).

It’s free, it’s simple, and it works. Is it perfect? No, but it exists in the world, which is more than I could say about any of my ideas just a short month ago.

If You're Stuck in the Same Loop

Maybe you've heard of all these AI tools. Maybe you've been wanting to build something. Or maybe you're overwhelmed by how many options there are and you don't know where to start.

Here's my not-so-expert-and-just-based-on-my-own-experience advice:

You don't need to know everything before you start or how every tool fits together. You also don't need a big idea or a plan laying out the next 50 steps.

You just need to take the first step, and the next step will show itself.

Maybe your first step can be copy and pasting this into your favourite AI:

I’m a beginner and I want to build a simple app using AI. I've heard of Lovable and Bolt and other stuff, but where should I start? Short answers only, and give me a prompt I can use. Make it so simple that anyone could do it.

Make the first step so simple it’s impossible to not start. 

Check out Drafted at https://drafted.meladkhajepour.com/, or in the Side Quests tab on my website. If you try it, I'd love to hear what you think – what works, what doesn't, what's missing. And if this post helped you get unstuck, let me know.

Is this a value trap, or is it the deal of the year?

A value trap is when a stock looks cheap, but stays cheap for good reasons. It’s like finding a house listed way below market price – it seems like a bargain until you realize that it’s cheap because the foundation is cracked and rotten and needs to be completely redone.

But Lululemon? It’s far from a rotting house, at least in my view. Something doesn’t add up. Nike, one of the most well-known brands in the world, has had its revenue decline 7% since last year. Their gross, operating, and net margins are 42%, 7%, and 6% respectively, and their return on invested capital (ROIC, a measure of how effectively a company generates profits from each dollar of investment) is 13%. 

Lululemon on the other hand, has increased revenues by 9%, their margins are 59%, 23%, and 16% respectively, and their ROIC is 35% – much better than Nike in all measures. The disconnect? Nike is trading at a price/earnings (PE) of 37x, while Lululemon is trading at a PE of 11x. 

Either the market knows something that I don't, or we’re about to get rich 🤑 (jk, not financial advice).

The runway that’s being ignored

The Street is so focused on slowing US growth, increased competition, and product fatigue (not to mention tariffs), that they are missing the international opportunity.

Lululemon's international revenue last year was $2.9 billion, or about 27% of sales. The size of the international activewear market? Over $200 billion¹ – that's not even a 2% market share. Even in the US, Lululemon’s brand awareness is estimated at around 40%.

China growth is moderating to 20-25%, but Europe and the rest of the world are barely penetrated. Unlike Nike, Lululemon isn't a mature company that’s running out of room. They still have years of international growth ahead, but the market is pricing it like the story is over.

What successful turnarounds look like

The market is also treating Lululemon's challenges as permanent, when history says otherwise.

Abercrombie & Fitch had similar product fatigue in 2015 and returned over 1,300% by refreshing their product line. Of course, just because one company did it doesn’t mean Lululemon automatically will too. But strong brands with loyal customers, capable management, and financial flexibility have historically navigated these cycles successfully:

• A&F overcame product fatigue and logo staleness; returned 1,300% 

• Crocs overcame product bloat and brand dilution; returned 2,600% 

• Domino’s overcame being ranked last in taste tests; returned 15,000% (❗️)

• Chipotle overcame an E.coli crisis that broke customer trust; returned 700%

• Best Buy overcame the “Amazon will kill them” narrative; returned 600% 

Lululemon's management has beaten earnings in 31 of the last 32 quarters (Covid being the only miss), and is planning to increase product newness from 23% to 35% for spring 2026. On top of that, they are enhancing their design and go-to-market process to reduce lead times and better respond to consumer demand.

Essentially, the stock is priced like this is the end of Lululemon and not a fixable issue.

The numbers

My target in the next 12-18 months is $300, reflecting a 22x PE on 2026 earnings per share (EPS) of $13.35.

This is based on the following simple assumptions:

• Consensus EPS for 2026 is $12.99. Management has historically beaten EPS by 5-15% with a median of 7%, so I took a conservative 5% beat ($13 x 1.05 = $13.65)

• Lululemon’s PE has historically been about double their annual revenue growth (see below), which implies that 22x PE reflects a ~11% revenue growth for 2026

• This 11% revenue growth, I am further assuming, will be the result of a successful spring refresh and a reacceleration of North America sales, leading to a PE multiple expansion

Value trap or opportunity?

As much as the numbers look good, Lululemon could still keep dumping. After all, it’s been on a downtrend for almost two years. The refresh could fail, tariffs could hurt margins permanently, and competition could get worse.

But when comparing the valuation to the fundamentals – best-in-class margins, pristine balance sheet with no debt, proven leadership, and a massive international opportunity – the risk to reward feels highly skewed to the upside. 

Analysts have aggressively cut their estimates, and the bar is low for management to outperform. Pessimism and catastrophic thinking have all been priced in, but the potential for a successful turnaround is not.

Policymakers (including Premier Smith!), CEOs, and business leaders came together with the fitting theme of building bridges in a fragile world, to discuss topics from arctic defence, international relations, and our lagging productivity.

We've gone from 90% of US levels in 2010 down to 74% today, and it shows:

The median household now falls 12% short of covering basic costs (vs 7% in 2000), with mortgage debt swelling from 80% to 131% of income. Our wages haven't kept up with inflation (2.6% CAGR vs ~3%), and our income inequality is actually growing faster than in the US.

Some reasons for our productivity crisis that stuck with me:

• Our tax burden falls disproportionately on lower incomes: we lack enough high earners to balance the load, with the top 1% of earners leaving Canada at record rates

• Our tax laws favour startups: 10-13% tax rate for revenues under $50 M but doubles past that, with scale economies not enough to offset the additional tax burden (and hence drives businesses south)

• We invest in the wrong things: our investment rate is more than most countries (25.5% of GDP), but too much goes into domestically-oriented sectors (e.g. construction and housing, which are important but don’t compete globally) and not enough into IP, machinery, etc to make us a more externally-oriented economy

• We have too many regulations: trade barriers and bureaucracy prevent businesses from squeezing out as much from every dollar of input (we lag behind the US in every major sector of our economy)

It would take 3% real GDP growth/year (more than double our current rate) for 10 years to get back to where we were in 2010, assuming the US only grows at 1.5% per year (it will likely be more).

What closing the gap looks like in practice is $500 B in strategic investments by 2035:

• Inland ports (1/4 the cost of ocean terminals with similar productivity boost)

• Energy infrastructure to attract advanced manufacturing

• Expanding critical minerals industry

• Increasing oil production by 70% (assuming we can get it to market)

• Building ~5x more LNG facilities

These are things that will make every investment dollar work harder by creating ecosystems. But it’s not going to happen on its own – we need a step function in our country's advancement with bold and transformational thinking, the same way the lightbulb wasn’t created through optimizing the candle.

What can we as citizens do? Change doesn't wait for permission – by advocating for pro-growth policies, supporting businesses that scale, or just simply talking about this stuff, all help to move the needle bit by bit towards building the country we want.

(These insights came from Robert Palter’s McKinsey presentation, with additional context from other speakers throughout the two-day forum)

This summer I got to intern at a search fund, which is where a group of investors provide capital to an entrepreneur to buy a small to medium size business and grow it over the long term. Prior to this, I’ve only really had experience with financial due diligence – the spreadsheet. But there’s more to a business than just a good spreadsheet, the same way there was more to mortgage-backed securities than just a good credit rating.

Here, I got to experience the transaction process and what goes into it beyond just the financials. When analyzing a business, we looked at a confidential information memorandum (CIM) – a sometimes 70+ page marketing document about the business, with info about its customers, management, operations, and financial performance.

I thought I’d have a decent grasp at due diligence from my experiences in the stock market, but there was so much outside of the financials that I’ve never seen before. Here are five interesting things I learned about due diligence that don’t appear on the spreadsheet.

The Owner Dependency Signal

Why it matters: A business that can't function without its founder isn't really a business you can own – it's someone else’s job you're buying.

How it showed up: A service business seemed like a great target: high customer retention, strong margins, and growing year-over-year sales. So then why was it being offered at only a 4x multiple when companies of its kind typically sell for at least 6x?

In the stock market, when a good company trades at a multiple lower than what is typical, it’s signaling either a bargain or a red flag. It’s your job as an investor to figure out which one it is.

In this case, it turned out the low multiple was because the owner was still very involved with the day-to-day operations of the business. What suggested this was that there was no general manager or management team, and the owner indicated that not only is he willing to stay on for two years to help with the transition, but also would recommend a buyer with specialized knowledge and high domain experience. The numbers looked great, but the rest of the CIM a little less so.

Scalability is desirable, and desirable things come with a premium. This 4x multiple was not a bargain, it was the most the market was willing to pay for a business that could struggle the moment its founder stepped away. If something seems too good to be true, it probably is.

The Invisible Asset Problem

Why it matters: Sometimes the most valuable parts of a business don’t actually belong to the company – it belongs to individuals, and can't be transferred in a sale.

How it showed up: A Canadian company provided a range of services for water and fire damage. Again another solid-looking company with good margins, increasing revenue, and had management in place to scale the business.

The challenge this time was one of the selling shareholders played a key role specifically in building relationships and handling claims associated with the Indigenous communities. One of their largest customers was from the Indigenous community and made up almost 20% of their revenue, and another major contract the business won was also from the Indigenous community.

These relationships were built and fostered over years and represented cultural bridges that couldn't be transferred to a new owner. The revenue the business generated through these relationships ultimately made this acquisition too risky to pursue with the shareholder exiting and no suitable successor to fill their shoes.

The Broken Habit 

Why it matters: When a consistent behaviour suddenly stops, there's usually a reason. 

How it showed up: A professional services company had been acquiring one or two competitors every year for nearly a decade. This company was executing on a roll-up strategy, which is to buy smaller competitors and combine them into one larger company for cost savings and economies of scale. The numbers showed that the strategy was working, so then why did they stop?

That was the question we had going into our call with the broker – maybe there was a good reason for the shift in strategy? Unfortunately, it started sounding less like it was a strategic decision, and more like the owners had reached their fund’s timeline to sell and were focused on finding an exit.

This pause in acquisitions was not a change in strategy as we had initially thought. They simply switched from buying companies to preparing to sell their own, reflecting a countdown timer running out.

The Opportunity Paradox

Why it matters: The gap between what someone claims they can do and what they have actually done can reveal hidden constraints or risks.

How it showed up: A company presented compelling growth opportunities in the CIM, highlighting growth for years to come: geographic expansion, introducing new services, integrating AI, and developing new technologies.

Everything sounded obvious, achievable, and straightforward. So then why has the owner not done it already? If these opportunities are as compelling as they are presented, then why sell now?

We couldn’t find a simple answer to these simple questions. If the current owner hasn’t done these already, maybe there’s a reason. Maybe they tried and it was too difficult, or it was too risky – if the current owner wasn’t able to do it, what would be different for a new owner? 

The Borrowed Advantage

Why it matters: A competitive advantage built on something you don't control isn’t really an advantage; It’s a dependency.

How it showed up: A software company looked like it could be a good acquisition. There were a couple minor issues like more churn than we would like and a big chunk of revenue that was non-recurring, but nothing fundamentally wrong about the business that couldn’t be improved. Until it turned out that they didn't actually own their platform – they relied entirely on putting together third-party software from different providers under their own brand.

Not owning the tech meant the business had no control over its product roadmap, and that its future was in the hands of the provider: They could make any change at any time and the business would either have to go with it or to find another provider. 

What was once considered the business’ competitive features were not really a competitive advantage, because their platform was available for any competitor to build if they only paid for the licensing fees. This company was just a middleman with fancy branding, using borrowed tools with no protection against competition.

My Takeaway

Not everything can be boiled down to just numbers. This experience opened my eyes to the qualitative factors of a transaction that go along with the quantitative. It was like having a flashlight in a dark room and then finding the light switch; You suddenly see the whole of the business and not just a small part of it. 

Each of the companies above looked good when only looking at their financials, but shining the light on their strategy, operations, and people painted a much different picture than what was being suggested from numbers alone. 

Next time you’re thinking of a big decision, whether in business or in life, remember that the most important information sometimes doesn’t show up where it’s obvious. Look beneath the numbers – what else is there that’s not showing up on the spreadsheet?

So that’s what Bill and Dave did with HP.

That small shift in encouragement created a gravitational pull: more engineers stayed local, which led to more companies sprouting, which brought in venture capitalists and talented engineers. Network effects started to form, with each successful company making it easier for the next one to thrive. Fairchild Semiconductor led to over 38 “Fairchildren” companies like Intel and AMD being founded by former employees. Today, the Bay Area is home to 49% of all Big Tech engineers and to trillion-dollar companies like Alphabet, Nvidia, and Apple.

This pattern of “clusters” can be seen around the world: Shenzhen transformed from a fishing village to the world's manufacturing hub. Paris is home to luxury brands that define global style. Toronto is a leading financial hub in North America and houses five of the world's most stable banks.

The same principle underlies each story, and the same principle applies to our daily habits.

My accidental discovery of the habit hub

In 2016 I decided to start weighing myself every morning. No tracking, no apps, and no big plan – just stepping on my scale to see the number. Soon I found myself paying more attention to nutrition labels and ingredients lists, which caused me to gravitate towards healthier options and to pay more attention to exercise.

As these habits solidified, I became curious to learn about weight management and how to build muscle, leading me down a rabbit hole of physiology, nutrition, and fat loss. Next thing I knew, I started roughly estimating my daily calories in my notes app. 

After a month, I figured I might as well track accurately if I was going to put in the effort anyway, so I made a Google Sheets to track my calories, fats, carbs, and protein. Since I was already weighing myself daily, why not add that to the spreadsheet too?

My habit hub was forming without me even realizing it. 

The power of spillover effects

As my awareness in the domain of “health” increased, I naturally started making better choices about my nutrition. As my nutrition habits improved, they spilled over into my exercise habits, transforming how I trained and created my workout plans. Better training demanded better recovery, which meant prioritizing sleep. Better sleep meant clearer thinking, which meant better discipline, food choices, and improved mental and physical performance.

Energy flows where the attention goes.

Once you start paying attention to one area of your life, that focused energy naturally spills into related areas. 

This phenomenon is called the spillover effect, and explains how one positive choice creates momentum for other positive choices. A 2015 study found that daily weighers lost 6.1kg more weight than occasional weighers and adopted 17.6 weight control behaviors vs 11.2. James Clear popularized “1% better every day”, emphasizing that small improvements lead to massive growth in the long run due to compounding: roughly 37x better in just one year.

This is because each new habit isn’t just a simple addition to the rest – they all strengthen and amplify each other. In my International Business class this summer, we learned that economists call this behaviour "increasing returns to scale": The more you invest in your habit ecosystem, the greater the return on each additional habit. Just as Silicon Valley firms benefit from proximity to each other, good habits in exercise benefit from proximity to good habits in diet.

Building your own habit hub

The key is to pick ONE habit so small and simple that it requires no willpower:

Health: Step on a scale every morning. Don't track it, just build awareness. Watch how this naturally evolves into curiosity about nutrition, hydration, sleep, and exercise.

Wealth: Check your bank account during your morning routine. This can spiral into budgeting, tracking spending patterns, and eventually investing and building wealth.

Career: Read one article (or paragraph, or sentence) in your field during lunch. This can progress to following thought leaders, joining communities, and building expertise.

Mindfulness: Take one deep breath when you sit in your chair. This little pause can turn into better awareness, emotional regulation, and decision-making. 

Relationships: Make eye contact and smile at one person daily. This small connection can evolve into genuine conversations and stronger social awareness. 

Commit just one month to one thing that’s important to you. Let the natural curiosity and awareness build, and notice what questions arise. Follow those questions.

Start small and let the network effects compound

Within three years of my first morning weigh-in, an interconnected system of habits had naturally formed around nutrition, exercise, sleep, reading, and journaling – and they can all be traced back to stepping on that scale.

You don't need a master plan, you don't need unwavering discipline, and you don’t need to have everything figured out. You just need to pick one tiny habit and trust the process. Your brain will start connecting the dots and related habits will feel natural to add because they support what you're already doing. Every transformation starts with a single habit.

Just pick something small, give it 30 days, and the new habits will find each other. 

My brain dump file started as a helpful tool but turned into an 1,836 word black hole (for comparison, that's longer than this post). This is an empty note on my phone's home screen to clear my mind of ideas, reminders, todos, etc as they arise – like a pocket notebook, just on your phone. The idea was to process this note at the end of each day and put things where they belong, like reminders in my calendar or ideas in their own notes, and get it down to zero for the next day.

The problem started when I got lazy and missed a few days of processing. The couple of items started collecting, and that opened the door to my short note turning into an endless pit of "stuff" over the next year:

• Reminders and todos

• Interview prep tips

• Notes from books, videos, etc

• Trading and investment ideas

• Random thoughts I had

There were some things I wanted to keep, but the thought of processing this massive file was crazy.

So I figured, why not give it to AI? I could get my note organized, and at the same time run an experiment to see how different models perform.

Giving it to AI

I pasted the note into Gemini, ChatGPT, Claude, Grok, and Perplexity. I gave them the same prompt to organize the content, and each model provided similar groupings just structured a bit differently.

For example, Grok provided subsections for Key Takeaways, Why it's Important, and Actionable Steps, whereas ChatGPT provided subgroups within the main group. Claude's response was my favourite, it was the most creatively written.

Here's a couple verbatim examples:

ChatGPT

• Systems & Frameworks

  • AI Coaching Tools

  • Knowledge Management

• Productivity & Habits

  • Evening & Transition Routines

  • Micro-Habits & Environment

• Career & Finance

  • Investment Banking Prep

  • Networking & Outreach

• Content & Personal Brand

  • Newsletter/Blog Series

  • LinkedIn Posts

• Personal Growth & Relationships

  • Self-Reflection

  • Relationship Practices

• Urgent & Immediate To-Dos

  • Home/Logistics

  • Outreach

Claude, for comparison

• 🎯 Core Life Philosophy & Insights

  • Relationship Dynamics

  • Success Framework

• 💼 Career Development (Investment Banking Focus)

  • Key Preparation Areas

  • Networking Strategy

• 🧠 AI & Productivity Systems

  • AI Context Framework (Valuable System)

  • Evening Optimization System

• 📈 Investment & Learning

  • Market Insights to Track

  • Continuous Learning Plan

• 🏃‍♂️ Health & Habits

  • Simple, Sustainable Changes

  • Habit Formation

• 🚨 Immediate Action Items

  • Urgent Todos

  • This Week

• 🤔 Questions for Further Reflection

  1. Why did you stop photography and podcasts? What changed?

  2. How can you apply your natural teaching instinct professionally?

• 💡 Standout Ideas Worth Developing

  1. AI parenting assistant concept - helping develop intellectual curiosity in children

  2. Monthly health/productivity newsletter - 5-10 implementable tips

Reading through each response was an interesting experience:

• It put everything into a 30,000 foot view, helping me reflect and organize my priorities

• I learned more about how different LLMs tackle this sort of problem

• It was like a time machine bringing back memories over the past year

Seeing these summaries made me wonder, what's the one habit I should start immediately?

Finding the habit

I asked each AI what maximally effective habit they would suggest for me, while making sure it's simple and easy to do.

This was also interesting – for example, every model with the exception of Gemini gave me a similar "reflection and planning" kind of habit, where each evening I reflect on something I learned or accomplished today and then plan for tomorrow.

Recommendations from each model

The following are verbatim suggestions from each model along with their reasoning:

ChatGPT: The 10-Minute Power Wrap-Up

• Goal: End each day by capturing insight, tracking one health metric, and planning one small “win” for tomorrow.

• Why it works

  • Micro-habit (10 min) keeps friction low.

  • Reflects & records your daily learning (career, finance, AI).

  • Reinforces health by a quick weigh or hydration check.

  • Drives momentum with one clear, small action next day.

Perplexity: The 10 Minute Evening Review

• Combines: Productivity, self-reflection, learning, relationship growth, and habit tracking.

• Leverages: Your desire for better evenings, knowledge management, self-improvement, and relationship mindfulness.

• Simple: Only 10 minutes, no special tools needed, and easy to stick with.

• High ROI: Sets up your next day, helps process your thoughts, and creates a sense of progress.

Gemini The 15-Minute Market Memo.

• This one habit is maximally effective because it directly leverages and combines four of your core goals into a single, easy-to-manage daily action:

  1. IB Career Prep: It forces you to stay current with market news, a non-negotiable for interviews.

  2. Skill Development: It hones your ability to quickly synthesize complex information and identify the key takeaway—a core skill in finance.

  3. Content Creation: Each daily memo becomes a building block for your future newsletter or LinkedIn posts.

  4. Habit Rebuilding: It reinstates the daily "research and learning" habit you valued from your high school days.

Grok: Daily Reflection and Planning (5-10 Minutes)

• What It Is:

  • At the end of each day, spend 5-10 minutes answering three simple questions:

    1. What did I accomplish today? (Celebrate wins, big or small.)

    2. What did I learn today? (Capture insights or lessons.)

    3. What are my top priorities for tomorrow? (Set clear intentions.)

• Why It’s Maximally Effective:

  • Combines Multiple Skills/Goals: This habit blends self-reflection, goal setting, productivity, and continuous learning—all in one go.

  • Simple and Easy: It takes less than 10 minutes, requires no special tools (just a notebook or notes app), and fits seamlessly into your evening routine.

  • Builds Momentum: By focusing on accomplishments and lessons, you’ll feel motivated. By planning tomorrow, you’ll wake up with clarity and purpose.

Claude: 🎯 The One Habit: 5-Minute Daily Voice Memo

• Every evening during your commute home (or transition time), record a 5-minute voice memo answering 3 simple questions:

  1. What did I learn today? (1-2 minutes)

  2. Who did I connect with and how? (1-2 minutes)

  3. What's one thing I want to remember or act on tomorrow? (1-2 minutes)

The five minute memo

Claude's pick won for me because it was the quickest and easiest, though I adapted it to do it at home vs on a commute (I thought it was pretty random it said during a commute vs at home or something, but I digress).

I also figured that since it was voice-based, it would have the highest rate of return because the amount of reflecting and "content creating" per unit time was more than writing or typing.

The drawback for me though is that it's less accessible and convenient in certain situations like if there are people around – it's like a verbal journal after all!

How to start in under 20 seconds

A way to make new routines last is by sticking them to a routine you already have. For me, this meant doing it when I journal before bed.

1. Open your phone's recorder or voice notes app

2. Take some time to think about the prompts:

   • What did I learn today?

   • Who did I connect with?

   • What do I want to remember or do tomorrow?

3. Press record

For your first memo you don't even have to think beforehand – just start talking and see where you end up. You can always put more thought into them later.

Once you're done, here is how to build the file we give to AI:

4. Create a new document in Google Drive or similar (note apps don't work after a few memos)

5. Copy your recording transcript and paste it at the top of the file

6. Add a header – a simple "memo 1" works

Repeat steps 5 and 6 after every new memo to build a running document of your transcripts, with newest memos at the top.

Analyze your memos

Take your file and give it to your LLM of choice, and make sure that the "pro" or "thinking" mode is on. Tell it that this document contains transcripts from your daily voice notes, then ask anything you're curious about.

Here is an example prompt you could use:

This file contains transcripts from my daily voice notes, answering these three questions: "what did I learn today?", "who did I connect with?" and "what do I want to remember or do tomorrow?". Analyze these transcripts and help me identify themes, action items, or things I should work on. Look for any blind spots, gather insights, and help me see things from a different perspective. Provide your initial analysis in concise bullet form and then ask me questions that will help you provide a customized analysis tailored to me and what's important to me.

You can ask it anything else you can think of:

• Which insights could become LinkedIn posts?

• What stories or examples can I prepare for [important event]?

• What limiting beliefs or blind spots appear?

• What patterns show up in my relationship interactions?

• Where am I growing vs staying stagnant?

A quick note: Google's Gemini has a 1,000,000 token context window, so it can analyze and hold on to about 750,000 words. For me, that's about three years of transcripts based on my current file of almost 32,000 words in 48 days.

For comparison as of July 11, 2025, ChatGPT and Claude both have context windows up to 200,000 tokens, or just over seven months of transcripts. Grok 3 is around 130,000 tokens for under five months of transcripts.

My first 48 days

The first three or four recordings felt a little weird. I've never been one to send voice notes or record myself or etc, so it didn't start feeling more natural until after a few days of practice.

After about a month, I noticed changes in my speaking where it had become more structured and I was using fewer filler words when I explained things.

A few other returns from this investment include:

• Learning about a new system of note-taking called the Zettelkasten method, which so far has been really cool and transformed the way I organize my notes in Obsidian.
  

• Inspiration for other areas to use AI. For example, I'm experimenting with automating parts of a weekly review session that I do, so that I spend less time "looking for stuff" and more time thinking about what to do with it. So far it's showing promise, but still needs more work.
  

• Interview practice, by asking it to give me stories from my memos that I can use for behavioural questions. It gave surprisingly detailed and relevant suggestions, and then put them into STAR responses for each question. I ended up getting the offer for an interview I was prepping for, and I think a big part of it was because of this habit.

Conclusion

Overall so far, this habit has contributed positively towards several of my goals. The thing I really like about it is that it advances multiple priorities at once, which ultimately was what I was looking for from this experiment.

Most days haven't been insightful, but every week or two something interesting comes up which has made this practice worth it – it only takes one thought or idea to turn into something big.

Try this sometime, you might be surprised at what you discover. If AI and growth resonate with you feel free to message me on LinkedIn – I'm always happy to swap ideas and learn from others.