Paul Graham wrote about doing things that don't scale. The essay is about startups, but the principle applies everywhere. The best founders do unscalable work. They hand-hold customers. They personally call people who leave. They obsess over individual feedback loops.
Most people think this is a phase. You do unscalable stuff while you're small, then you automate it and become efficient. That's true for some things. But it misses the point. The unscalable work is often where the real learning happens.
Think about research. People write papers about their findings. They present at conferences. But the actual discovery process? That's full of unscalable work. You stare at data. You manually inspect samples. You try things that probably won't work. You build tools to look at one specific phenomenon more closely. None of this scales.
Most researchers would love to skip to the part where their insight has been automated, packaged, and published. But that's where they'd miss everything. The manual inspection of neurons revealed patterns that no algorithm would have caught. The tedious data labeling built intuition about the domain. The half-baked tools taught them how to ask better questions.
The unscalable work is the thinking work. Once you've thought deeply, then you can build tools. Not before.
Career building is the same. You could write one blog post and hope it reaches millions. Or you could personally reach out to the fifty people whose opinions matter in your field. Tell them about your work. Ask for their feedback. Send them your latest research. Maybe grab coffee with the people who are nearby.
Networking advice often frames this as a transactional thing. You meet people to build your network. But that gets the causation backwards. You meet people because you're genuinely interested in their work. You learn from them. The network is a side effect. And that approach, while unscalable, is how you actually build relationships that matter.
Writing applies too. A shallow blog post that tries to appeal to everyone reaches nobody. A detailed post about a specific thing that matters to ten readers will be read deeply. Those ten people will think about it. They'll share it with the people they respect. The idea spreads through referral, not through optimization.
But only if you write for the ten people who actually care, not the ten thousand who don't.
The scaling mentality is seductive. Everything in startup culture points toward growth. More users, more revenue, more impact. But in the early stages, scaling is often the enemy. You scale the wrong thing and suddenly you've optimized yourself into a corner.
You scale the interface before you understand what people actually want. You scale the product before it's good. You scale the marketing before you have anything worth marketing. This is how you end up with a startup that has millions of users but nobody who really loves it.
The founders who win are usually the ones who resist the scaling pressure. They do the unscalable thing. They learn from it. Only once they've learned deeply do they think about optimization and growth.
This applies broadly. You want to be good at teaching? Teach one person very carefully. You'll learn what works. You'll develop intuition about how people learn. Then you can write courses or tutorials that actually work.
You want to build a community? Start by having one-on-one conversations with people who are interested. Understand what they want. Understand the patterns in what they care about. Then you can build something that serves the community properly.
The unscalable stuff is boring. It's tedious. It doesn't look impressive on a resume. But it's where you actually get good at things. The people who are exceptional have usually done the unscalable work. They've paid their dues in the form of hours spent on things that don't multiply.
The counterintuitive truth is that the fastest path to scale is sometimes to reject the idea of scaling entirely for a while. Focus on depth. Build something truly good. Understand the problem deeply. Only then worry about leverage.