Notes from a Long Break

Over the past few weeks I’ve reflected on myself—body, mind, and priorities—and rethought what I want to do with my life. It was a really good reset for my mid-thirties. Here are some self-reflections, learnings, and tools I’m going to carry forward in my toolbox, hopefully clearer and with a new perspective on how I should tackle things and how I should see them.

One note: I often change my mind, and I try to be open about it—not trapped inside my ego or by commitment bias. I hope I’m successful in that. What I’m writing here is my current opinion. In the future I might change it, I might not, but some of these points won’t hold as I progress, and I may change my mind.

I’ve heard many times that you shouldn’t compare yourself with others and only compare yourself with your past. I don’t think the comparison is fully wrong, but I do think it’s rarely fair. We usually take a snapshot of an outcome and pretend it tells the whole story. Two data points - mine and someone else’s - maybe from LinkedIn or social media, plotted on one axis. But life is multi-dimensional. Those points don’t show real progress. So constant external comparison isn’t useful.

For me, comparison is just a tool. How we use it matters. The only fair comparison is internal, because I can look at myself in more dimensions. External comparison is useful now and then, just to see other possibilities, directions, and what a normal pace might look like and then back to my own path. In the short term I compare myself with myself to check effort and direction.

Outcomes also won’t be linear. Progress is often S-shaped: a plateau, a leap, then another plateau. Sometimes we compare someone’s leap with our plateau and draw the wrong conclusion.

None of us have the same resources, the same starting point, or full control. Perception can be manipulated, and snapshots lie. Luck also matters more than we think. I’m grateful I wasn’t “successful” earlier; if I’d been even 5–10% more successful at certain points, I might not be here now. Steve Jobs said you can only connect the dots backward and I feel that. Sometimes where we are looks like a local minimum, but it’s the best path to reach the next summit.

So this is how I use comparison now: frequent (or semi-frequent) internal reflection to make sure I’m learning and not settling into mediocrity, and infrequent external checks to explore other ways and, if needed, adjust my course.

In different snapshots of my life I’ve felt ahead, and later behind. That’s fine. The choices we have are based on where we stand now. Comparison, used wisely, helps with direction. Progress, not position, is what matters.

A longer horizon changes how I invest. I’m in my mid-30s, and over the past few months I’ve been using CalAI, an AI based calorie counter, to get healthier, and it’s been amazing. I think we’re a generation that might see a leap in life expectancy.

I’ve looked into the numbers: developing countries have had gradual increases in life expectancy and many developed countries look plateaued. But with the progress humanity is making, advances with AI and other technologies, I wouldn’t be surprised if we see a jump in average life expectancy in the next few years. We might be the generation for whom 100 is normal, maybe even 150. If that happens, then being in my mid-30s is like being thirty at the end of my teenage years.

I’m grateful for what I have, and I believe it’s better to put effort now into building a very good foundation. I see myself as someone at the end of my teenage years because I believe average life expectancy will be higher. Of course I can’t say how long I’ll live, who knows, but on average, it doesn’t feel “late”.

This phase shift in how life may arrive before the official numbers move. The people stretching life expectancy are busy building; they’re not doing what people in their 20s, 30s, or 40s used to do. We keep comparing ourselves to the previous generation’s “normal”, but there are frontier people pushing it, and they’re the new normal.

Lifespan will stretch, and phases will be redefined. If someone in their 40s says they’re having a midlife crisis, it might not be midlife anymore, maybe it’s a quarter-life. Things are changing. Just imagine being in your 80s and still feeling mid way; that’s the point.

If I might be around for a long time, the game is different. My 30s feel like early middle chapters, not late ones. That pushes me to invest in things that compound. It also makes it easier to ignore the performative race. I don’t need everything now; I need the right direction and habits that still make sense twenty years from now.

Personal note: in two months I lost about 12 kg. I didn’t do anything extreme. I tightened food choices, used an image-recognition app to log meals quickly, and started running.

We like to look into the future in a straight line, but that’s not how it plays out. In the short term, we put in effort and see nothing and then, all of a sudden, there’s a leap, and then another plateau. Health, money, career, business. None of it is a single linear line. My life has many S-shaped curves: long quiet stretches, sudden jumps, then flats again. Compounding is a big reason. Inputs stack silently until they hit a threshold, then there’s a gear change.

This view helps me avoid two traps: overestimating what I can do in the short term and underestimating what I can do in the long term. We get discouraged because we expect smooth returns. We don’t see quick results and we quit one turn before the corner. I’ve done it; I’ve also seen step-changes. When I migrated countries, my salary jumped after a long flat line and then it flattened again until the next step. That’s normal.

My fix is two clocks. The short clock judges inputs: what I controlled this week. Show up, tighten the process, keep the basics. The long clock judges outcomes: what shows up over years.

That mix keeps me steady when the graph is flat and humble when it suddenly spikes.

These ten weeks of reflection helped me see one thing: I used to label things I tried that didn’t work as failure. Now I see them as iterations. “Failure” carries a full-stop feeling; “iteration” is just another try.

A handy frame here is the secretary problem (optimal stopping). You split the search into two phases: explore and exploit. In the classic version, if you have 100 candidates, you sample roughly 37 first to learn the landscape and set a bar; then you pick the next one who clearly beats that bar. It’s a metaphor, not a rule, but it cleanly separates learning mode from choosing mode.

With a longer horizon, it makes sense to keep exploration alive. Early, many iterations beat early “success” that traps you in a local maximum. If I could rewind, I’d iterate faster: change jobs more often early on, try different industries, learn more stacks, write more, publish more. Treat attempts as experiments, not verdicts. I was lucky to find work I like, but going forward I’m applying this lens intentionally. No failure, no absolute right or wrong, just exploration before settling.

“Failure” feels like a full stop. “Iteration” feels like a comma. If I run 100 experiments and none hit, I’m not a failure; I’m 100 data points closer. You don’t need to be right every time; you need to be right once in the right place. So test quickly, learn the shape, then go deep on the one that meets the bar.

“Fake it till you make it” only works when the gap is small. If the gap is too big, the weight of the claim breaks you. My rule of thumb: 10–20% stretch beyond current ability- big enough to force growth, small enough to be honest. “CEO of Tesla tomorrow” is fantasy. “Run a 5-person team” is a stretch that forces new skills and likely holds. Same in fitness, language, writing. one level up, not ten. Small, honest stretches, repeated, beat big performances that crash.

The other thing here is commitment bias and how easily it ruins progress. I’ve seen people (and myself) protect ego or a past opinion, insist a path is “right,” and push longer than needed, even when their gut and everyone else says it won’t work. Projects keep going just to defend an identity. That’s sunk cost dressed up as conviction.

During this break I spoke with a lot of old friends and met new ones. Many are successful, running their own thing or operating at a high level in great companies and a common thread came up: they were in the right place at the right time, and they openly admit luck played a strong role. There’s no playbook you can follow step-by-step to guarantee the same outcome. In fact, some tried to repeat their earlier win and couldn’t, even with more skills, more resources, and a bigger toolbox. It’s a multivariable equation we can’t fully optimise.

So luck matters, more than we think. But while you can’t control luck, you can increase your luck surface area. Play more, in public. You don’t have to be right every time; you need to be right once in the right place. The odds tick up as you keep showing up. I think the curve is S-shaped: tiny gains for a while, then a leap, then a plateau. Be humble when the leap comes and honest that luck helped. I was lucky early on to end up in computer science; another choice (electronics) might have led somewhere very different.

There’s a catch: opening more doors also invites more noise. You need two things at once: expand surface area and build a better filter. Be more open and public, but filter hard so the signal-to-noise ratio doesn’t drown you. Life is a lot of filtering: what matters, who matters, what to ignore. If your filter is too strict, you say no to everything. If your surface area is huge but you don’t filter, you get lost in distractions. Do both.

Most people, when honest, say their success was both skill and luck, with timing doing more work than they admitted at first. You can’t order luck to show up, but you can give it more doors to walk through. Material success also matters—it often buys the time and space for the virtuous or spiritual kind to grow.

People talk about consistency and why it matters. Show up, same time, same process, again and again. I like that. If you can turn learning into a habit, running into a habit, reflection or meditation into a habit, that’s invaluable. But for some of us, staying perfectly consistent is harder. That’s where persistence matters more.

I’ve seen many successful people who aren’t clockwork consistent; they work in sprints. Two relentless weeks, then a quieter stretch and repeat. Life is volatile. Sometimes you need to deliver fast; sometimes you need recovery. Adapting to the flow can produce the same result, or better, than rigid consistency.

My rule to myself: if I can schedule it and make it a habit, do that. If I can’t, then persist until the result arrives. It’s not just “discipline.” Persistence is trying again after sundown, even if the day didn’t go to plan.

When I started aiming for a 5 km run, I couldn’t run five minutes straight. But I stayed persistent, run, walk, run; up and down, and I finished the 5 km. Over time, persistence created more consistency. That’s why, for me, persistence beats perfect consistency.

I found running late, and it changed more than my legs. Ten minutes can tilt a day. When I run, my head clears. Problems don’t solve themselves; I become the person who can solve them. Alongside writing, reading, lifting, and honest conversations, running is now a reliable reset button.

I’ve said this before on LinkedIn: the computer surface is tilting from keyboard to voice and video. More dimensions, more streams, more multimodality. I’ve been using a lot of AI tools to write code and automate things and, most of the time, I don’t type anymore. It’s not about a great mechanical keyboard; it’s about a decent microphone. I can imagine interfaces with fewer buttons, more space to talk or show, camera-on interactions by default. Text becomes the accessibility layer. Even this blog post: I’m dictating it, then asking the computer to grammar-fix and polish it. I often double-tap the function key and start talking to the computer. If I had to bet on the future of the computer, it’s voice first. In the short term, voice and phone-like interactions will play a bigger role until we find an even better interface. Voice agents and AI agents will be more vocal and more visual; they’ll see, and they’ll generate better video in real time. That’s how we’ll use computers.

I’ve been testing small AI habits: logging food with photos, capturing tasks by voice while walking, turning a messy whiteboard photo into structured notes. When friction dropped, the behaviour stuck. That nudged a belief: the way we use computers is tilting toward voice and video, with text as support. Language is our native API. We’ll speak or show; the system will structure. Software will feel less like a place you go and more like a conversation you start.

I don’t claim these are universal truths. They’re what I learned on a long, needed pause—and what I’m taking back with me.