The other side of building

I think at this point everyone can agree that using any type of machine learning tool a la Claude code or Codex will allow you to build whatever you want at 400% speed. All fine and dandy. But here’s something almost nobody is prepared to admit.

Each thing you build becomes its own product. With its own needs. Marketing, copy, design, website, animations, product-market fit, target audience.. In the beginning, when you’re just focused on building stuff, you tend to forget or ignore that. But once you start thinking about monetization, this is when it hits you.

All of that becomes a super long todolist.

The illusion

If you’re a designer, or at least a UX professional, you’ll likely think about this quite profoundly. Even more so about quality, and presentation as the calling cards of the product. And then, you’ll soon find out about legal compliance, privacy policy, terms and services, and GDPR. Basically everything it takes to have a fully fledged product that is something people will pay for.

All of that takes a great deal of effort, care, research, information, design, copy, marketing, whatsoever. Who is saying that having AI there to do the heavy lifting is going to make all of this a breeze is simply huffing paint or just plain talking out of their ass.

Having codex or Claude helps enormously. But that’s only half the battle! You still have to think about how you’re going to do all of that. Maintain all of that. Comply with all of that.

“But why would I want to use my own copy and my own strategy and my own…” you get the picture

Because its 1. your product and 2. because you are supposed to be the person in charge. Since you are legally responsible for what happens if something bad happens. Not your AI assistant. Not Claude, not Codex.

Then there’s also the fact that AI can give you a framework. Guidelines. Ideas. Nuggets. Simply using whatever it outputted without reading it, without reviewing it and writing it in your own style, does nothing for you or the product you are building.

Plus, there’s 1000 other people out there like you, that think that if they slap together some corporate jibber-jabber and have a gradient design they never have to touch again, that’s going to be enough.

Quality

How is your product going to stand out from those other products? Let’s set featureset and value aside for the moment. Since those are things you will NEVER be able to fully communicate out the gate when you launch. Or even after. Quality is often still mirrored by visual execution level, since that signals “This person is ready to pour in time to polish this product up”.

UX Is what keeps people coming back. But to get those people ? UI is the one doing the marketing pitch. It sucks, but that’s the truth.

Power overwhelming

If you got the reference, you probably played Starcraft too. All of this building power will come with the downside of having to follow up on everything generating extra work. And even for 1 person, that’s a LOT! You will need to prioritize, and learn to do it as if you were one of the ex bosses at one of the companies you worked or work for, and loathe so much, when they say, we’re not going to do this, but that.

Here’s what you have to understand, when you’re at a grassroots level, things look very simple. Deceptively so sometimes. Because you’re looking at things from a single lens. Your feature. Your product. Your work. But whoever is acting as a product manager probably also uses the business lens, the management lens (the one we usually hate because its the one that decides what gets to be built and what doesn't, based on the available resources - read people - and impact it will have on the product and potential earnings) and sometimes the research lens, if they’re really really good.

When you start building for yourself, you have to embody a whole team. Marketing, software design - read architecture, development, design, product management, support, and legal all rolled into one.

The other risk

If you’re the type of person that likely has ADHD or suffer greatly from starting a lot of projects at the same time, because you’re either enthusiastic or simply restless or love trying out things, you will quickly find out how this is going to be problematic.

This is where knowing what to build comes in. Abandoning projects that seem like a good idea, but will likely offer little in the ways of prospective business, might not be as interesting as ones with prospective business return.

How do I know this

I started building a plugin, wanted to add a favicon, didn’t have one, so I built a favicon editor, which turned into a full on graphic suite, but then figured I need some sort of project management tool that wont cost an arm and a leg, runs locally and is privacy focused, so i built it. At the same time, I was looking for a new job and entertaining the idea of offering my services as a freelancer, so, naturally, I built myself my portfolio website. With CMS, password protection on projects, database and what have you.

By this time, I had just about 3 major projects on my hands. Each one needing its own website. With its own backlog, and with a LOT of design debts on the content.

Some of it I used AI for, to add fill-in until such a time as I managed to replace it with my own text.

And that’s where I figured out how much work that is going to take.

One rabbit hole after the other. Building just creates more debt, more work, and more need for care. Some people and businesses have figured this out, and when they do releases they try to combine debt with new features in equal proportion to ensure that they don’t turn into a feature factory where the next release is going to be better than the last one, burying bugs somewhere in the backlog of oblivion where the software is no longer going to be supported. Because that’s one way to address it.

But it’s not THE way.