It’s better to copy someone else.

It’s necessary to use open-source libraries that have already done half the job.

I’ve realized how idiotic it is to try to create a new product category.

Just like I’ve realized that building code from the ground up is dumb as fuck.

You will miss a huge number of edge cases and contextual knowledge that only comes from interacting with the problem over years. The lack of those details inevitably makes the user experience worse.

There is no realistic way to incorporate all those edge cases and localized ecosystem knowledge when you haven’t faced them yourself. Nor is there any rational way to predict them in advance.

You can’t predict edge cases through imagination alone. They emerge from actual usage.

Mature categories have already paid that learning cost collectively.

Open-source projects are a manifestation of thousands of hours of work by hundreds of people operating across UI, UX, APIs, and codebase architecture.

Hopefully, you understand what I mean when I say you cannot code things from the ground up, because code itself is not even 5% of what is required to build a good app.

The same applies to software products.

Even if you are the first person to build something in a category, you are operating in the dark. You have no idea what use cases people will demand or which flows they will prefer.

A lot of time gets wasted fighting unknowns.

Even with Rohkun, people expect so many features that it’s unclear whether it even makes sense.

That’s the core problem with building something novel, like a codebase architecture map provider for AI.

Now look at CRMs.

Do I even need to say anything?

They’ve been around for decades. The software is so mature that its features account for most edge cases, and we’re close to an optimal workflow for a large segment of users.

The smaller the team, the stronger the case for choosing a well-defined, well-explored problem.

Your differentiation should come from small, focused improvements, not from creating an entirely new category.

A single strong feature can be enough to get clients. You don’t need to invent a new category to win customers.

Good user experience and product quality are the result of accumulated collective knowledge, experience, and iteration.

Open-source projects and mature categories embed thousands of invisible decisions that you would otherwise have to rediscover the hard way.

Missing even a small subset of those decisions shows up immediately as friction, confusion, or churn.

When you try to create a new category, you also take on the burden of educating the market.

You’re here to make money, not to be a teacher.

That’s also why salespeople thrive in old, established industries like insurance, car sales, or ERP.

The problem is so well defined that they can develop a playbook and answer almost any objection.

The same cannot be said for novel technologies, which depend heavily on how receptive the customer already is.

You’ll have a much easier time selling a CRM through salespeople than trying to manufacture hype for a novel AI idea.

When you work on long-standing problems, you can reuse existing industry knowledge.

You don’t have to reinvent the go-to-market wheel.