September 6, 2026 • Author: Vitalik

Why slow?

Who’s slow? Me?

If you’re asking me if I’m slow, I swear I’m not.

I work on HandsOnMoney every day. Two to four hours a day. On weekends and holidays.

I work early mornings or late nights when everyone is asleep. I remove all the possible distractions and put myself into a flow mode. I become single with code and a problem I’m trying to solve.

I should admit though that I might be confusing busyness with productivity. I don’t ship a major feature every week. Heck, I’m happy if I can ship something major at least once a month.

What’s hard?

I spend around 50% of my time coding. Yet I do not think it’s coding that’s a bottleneck in shipping features faster.

An application’s code is like a legal contract. A wall of text made of precise language. Worked out through years and years of experience.

I program for 20 years and do professional software engineering for the last 15 years. So writing code is not a big deal. I think in code, so the process of typing is rather fast. It’s rare when I’m getting stuck seriously for more than an hour with code. Also Internet, and modern languages have made this process even easier.

Then why does it take so much time, and what is it that takes so much time?

It’s knowledge acquisition. With HandsOnMoney, I’m trying to understand how to merge two worlds. Nerds’ world of double-entry accounting and the busy-bee world of simple finance trackers.

Sure, there are definite answers to things that are set in stone - like the accounting equation. But there is no definite answer to things like displaying a net worth chart, guiding users through onboarding, or preventing them from making iCloud file conflicts.

I’ll be honest with you, that was and is the hardest challenge. And that’s where the majority of time goes. I make something that I think makes sense, and then it simply looks ugly, does not make sense, or feels wrong. And it takes at least a dozen iterations to make it at least more “right” than wrong.

And there is one more time sink - changing previous assumptions. In software engineering, it’s called refactoring. And surely it can be done with AI. Yet I do it manually.

Because these are my assumptions that need corrections; thus, I want to be in the driver’s seat to make sure that the new assumptions do not break the old ones and results actually make sense after I correct myself.

Outsource? Open source? AI?

So theoretically I can delegate those tasks to churn features faster:

Each approach has its own pros and cons. Hiring a team would test my leadership and management abilities and burn some of my personal savings.

Making it open source will build more trust and hopefully bring other humans to the table, but will risk a “bazaar of ideas”-style product and divert limited attention to community building.

Building with AI brings more questions than answers (how can I delegate decision-making to a machine that does not have my product taste, faster yet make its own decisions that rarely align with mine?).

All those practices would take decision-making from me. And that’s a core problem for a product at its infancy like HandsOnMoney. At this stage, I decided to be in the driver’s seat, drive slow, feel every bump, but at least drive where I want HandsOnMoney to be.

Will it be like this forever?

Of course not. I’m not dogmatic about coding with AI; I love working with people and am open to different ways of working.

Right now I’m open to ideas coming from users. Their way of using HandsOnMoney, their way of budgeting, sharing their data, etc.

It will take me a bit to reflect and generalize those ideas into product decisions. But I’m not yet ready for other people’s ideas (or AI ideas) implemented inside of HandsOnMoney.

But when I’m done with core mechanics. I’ll need a lot of help shipping HandsOnMoney to other platforms: Android, web, and desktop. And surely I’ll ask for help. For now, I’ll stick to what I do now: slowly acquiring knowledge of what’s working and what’s not.