The Quiet Wins: Celebrating What Doesn't Scale
The tech industry is obsessed with scale. And honestly? I get it.
“Will it scale?” We ask this about everything. Systems, processes, side projects, even our own careers. We worship leverage. We dream of building something once and watching it multiply across thousands of users, millions of transactions, returns that feel almost infinite.
As a platform engineer, this is literally my job. I build infrastructure that helps hundreds of developers move faster. Scale is what I’m paid to create.
But here’s the thing that caught me off guard when I looked back at 2025: the stuff that actually moved my career forward? Almost none of it scaled. The investments that compounded were small, manual, and sometimes painfully inefficient. They didn’t make it to LinkedIn. They won’t show up in any performance review.
And yet, they mattered more than the flashy stuff.
The Case for Doing Things That Don’t Scale
There’s a story that gets passed around in startup circles so often it’s become legend.
Airbnb’s founders didn’t just build a website and hope for the best. They grabbed cameras, crisscrossed New York, and personally photographed strangers’ apartments to make each listing look decent. Stripe’s early team showed up at users’ offices and installed the software themselves. These weren’t scalable moves. They were the spark that made something bigger possible.
Paul Graham wrote about this back in 2013 (“Do Things That Don’t Scale”) [1]. The counterintuitive insight: the most successful startups didn’t succeed by building scalable systems first. They succeeded through brute-force, unglamorous, one-at-a-time effort.
The same pattern plays out in careers, though we don’t often talk about it. We get swept up chasing personal leverage — the viral post, the clever automation, the growing network. But beneath every real milestone, there’s a layer of hands-on work that nobody sees. Foundation gets built brick by brick. No shortcuts.
Cal Newport’s research backs this up. High-quality work equals time spent multiplied by intensity of focus [2]. Simple math, brutal implications. You can’t hack intensity. You can’t automate focus. And every interruption costs you. Cognitive scientist Sophie Leroy found that switching tasks leaves “attention residue” that takes roughly 23 minutes to clear. All those pings and notifications? They’re not just annoying. They’re quietly destroying our capacity for the deep work that actually compounds.
Same story with relationships. We’re told to grow our networks as large as possible, but Robin Dunbar’s anthropological research suggests humans can maintain only about 150 stable relationships — with maybe 5 truly close ones [3]. A study on XING found that users with around 157 connections reported the highest success rate for job offers. Not thousands. About 157.
Depth beats breadth. Every time.
Three Unscalable Investments from 2025
1. Building a Curated Learning System
Earlier this year, I noticed a pattern. I’d open YouTube to learn something specific (can be Kubernetes networking, Python design patterns, whatever) and 45 minutes later I’d be watching something completely unrelated. The algorithm had won. It optimizes for engagement, not for what I actually needed.
So I did something inefficient. I built my own system. Curated a list of YouTube channels and podcasts I trusted. Organized them in Notion. Committed to consuming only from this filtered list, ignoring algorithmic recommendations entirely.
Then I added a constraint that made it even less scalable: I had to find a way to apply what I learned. Not just watch. Not just take notes. Actually use it somewhere.
A few months ago, I watched a video on the Registry Pattern in Python. Not exactly thrilling stuff. But I was dealing with scripts at work that had grown into these sprawling if/else chains. Hard to read, nightmare to extend. The pattern clicked. I refactored the code. Cleaner, more maintainable, and it came directly from forcing myself through that inefficient loop: filter sources → consume intentionally → take notes → find application.
This system takes hours to maintain. It doesn’t scale. But learning things I actually use, instead of things an algorithm thinks will keep me watching? That compounds.
2. Going Beyond the Ticket
Platform engineering has a support problem. Developers open tickets, we resolve them, everyone moves on. Transactional. Efficient. Scales nicely. But the best work I did this year didn’t look like that.
Several times I got requests where the solution seemed obvious but the reasoning wasn’t clear. Instead of just executing, I took time to dig deeper. Why had the team made certain choices? What constraints were they working with? What had they already tried? What were they actually trying to achieve, beneath the surface of the request?
This took longer. Required conversations that weren’t strictly necessary. Meant asking questions that might make me look like I didn’t understand the task.
But it led to better outcomes. When I understood context, I could offer options they hadn’t considered. Point out trade-offs they might have missed. Help teams build better solutions, not just the solutions they initially asked for.
I also spent significant time helping individual teams with problems that weren’t platform-wide priorities. Improving Lambda build times for one team. Building an authentication system with OKTA integration for another. Creating a metrics database for historical test analysis for a third.
None of this showed up in platform-level metrics. These were bridges, built one at a time, for specific people with specific needs.
3. Reclaiming Attention
I have a confession. For months, I told myself I didn’t have time to learn Dutch (I moved to the Netherlands a year ago) or prepare for certifications. “No time” was my default excuse.
Then I looked at my phone’s weekly screen time report. Hours. More hours than I’m willing to admit. Spent scrolling content I couldn’t even remember five minutes later.
Worse: whenever I had to wait during work (running terraform plan, waiting for a pipeline, any pause longer than thirty seconds), I’d instinctively grab my phone. Then, when I switched back to the task, I’d forgotten half of what I was doing. The attention residue was real. I was paying that tax constantly.
So I made changes. Filtered my social media access. Blocked the algorithmic feeds designed to capture attention rather than deliver value. Created what I started calling “sacred” deep work blocks in the morning before work. Time for learning something new, no interruptions allowed.
I implemented shutdown rituals to separate work from personal life. Started tracking focus days the same way I track other metrics. Not to optimize, but to stay honest with myself about whether I’m actually investing in depth or just feeling busy.
This is ongoing. It’s not a habit I’ve built. It’s a daily negotiation with my own impulses. No automation, no hack, no system that does it for me. Manual. Repetitive. And the only thing that works.
The Paradox
Here’s what I keep coming back to.
I build highways at work. That’s my job (platform capabilities that let hundreds of developers move faster, deploy safer, operate with less friction). Scale is the whole point.
But my career wasn’t built on highways. It was built on bridges.
The one-on-one conversations that led to better solutions. The hours curating a learning system nobody else will ever use. The daily fight to protect focus in a world designed to fragment it. The choice to understand context instead of just closing tickets.
These bridges don’t scale. Can’t be automated. Require ongoing, manual investment.
And they’re the only things that actually moved the needle.
A Question for 2026
I’m not arguing against scale. Building highways matters. It’s how we create leverage, multiply impact, serve more people with less effort.
But I wonder if we’ve over-indexed on it. If we’ve gotten so obsessed with scalable wins that we’ve forgotten the unscalable foundations they rest on.
So here’s my question, for myself and for you: what quiet win will you invest in this year?
Not the thing that’ll look good on LinkedIn. Not the automation that saves hours. Not the network effect or the viral moment.
The bridge. The slow thing. The investment that won’t scale but might just compound into something that matters.
What’s yours?
References
[1] Graham, P. (2013). Do Things That Don’t Scale. paulgraham.com/ds.html
[2] Newport, C. (2016). Deep Work: Rules for Focused Success in a Distracted World. Grand Central Publishing. See also: Leroy, S. (2009). “Why Is It So Hard to Do My Work?” — research on attention residue.
[3] Dunbar, R. — Research on relationship limits. See also: XING network study on optimal connection counts for job success.