When I started in digital design, my main intent was always the same: build useful things. Tools, software, products, services – or, to put it more broadly, solve problems worth solving with digital means. Some people say that's mostly done by now, because there's an app for almost everything. I don't agree.

Products and tools

If I look at what's really useful, I see two kinds of products. The first ones have good UX, they solve a problem in a way people actually want to use. The second ones have an excellent vendor lock in, based on a market position they established and grew over time – and in most cases rightfully so.

The rest are tools, and where there isn't even a tool, people find workarounds to organise themselves: Excel lists, pen and paper, WhatsApp groups. I don't mean that negatively, all of them do their job. They work, but they rarely go beyond that.

Why has it stayed that way?

Resources decided which problems got solved

Building really good software takes resources, and for a long time that was the limiting factor. At the same time, resources were the door opener. Whoever had them could solve problems – and earn a lot of money with it.

Turned around, this means that problems mostly got solved well where money could be made. Many of those are important problems, and the right ones to solve. But everything else was left to tools and workarounds, or stayed unsolved.

AI lowers the barrier, not the bar

With AI, this changed. Finally everyone can build software – at least theoretically.

But not everyone can draw a good picture, a good graphic or a beautiful and individual invitation for a birthday party, and AI didn't change that either. The same is true for software. Even with AI, not everyone can build good solutions for problems.

This is an important distinction, so let me be clear: I'm not talking about vibe coding or prototyping. I mean developing real software products, services and solutions that get rolled out. Until recently, that was something only people with a real software development background were able to do.

And let's be honest, all the WYSIWYG tools and later the AI vibe tools – from Frontpage to Dreamweaver, Wix, Framer, Figma Make, Lovable and Base44 – are nice. They help with websites or smaller things. But they don't get you to real productive software, even though they promise it.

Who can build now

What did change is the entry point. People who know how to build software with good UX and solve a real problem now face a much lower barrier. The same goes for people who are eager to learn and willing to dive deep.

There's one more ability that matters now, and that's using the possibilities of the AI models.

Of course one might say the bill will come, once the labs increase their prices. That's a fair point. But be aware that there's a huge leap forward with open weight and open source models, and a lot of interest in working with them.

I see it in a positive way, despite all the risks and dangers. Maybe it's like with the railways: many of the companies that built them went bankrupt, but the rails stayed. If you trust the benchmark tests, Opus 5 level coding abilities are really good, and whatever happens to the business models behind them, that level of capability won't simply disappear again.

What it takes

It's really exciting times to work in digital now. You can build whatever you want, without the limiting factor that decided for so long what got built and what didn't.

What you still need can't be handed to you by a model. Part of it is craft: a basic know-how of how software works, and experience. Part of it is judgement: taste, the ability to identify the problem in the first place, and to decide what a good solution looks like. And curiosity is what gets you in, especially if you don't come from development.

Money still matters, of course. Companies still need budgets, and as a private person you still need a problem that matters to you. But when building gets this accessible, the first question is no longer whether you can afford it, but whether the problem is worth the effort.

Imagine software for elderly people, for teams, for families, for integration or for neighbourhood help. Software for small and medium companies, or for large companies with specific workflows that today depend on tailored or licensed solutions. And a lot more for NGOs. That would be a real leap forward.

Problems worth solving have always been solved, the ones with a business case included. Now there are more and easier ways to solve the others as well. We can do that today, and I think we should – for me, that's very close to why I got into digital design in the first place.

Sources