For Arjan, the world has always been about systems.

A mechanical engineer by trade, he spent most of his early career as a serial entrepreneur, launching companies in 3D printing and personalized medicine. For years he was used to the slow, methodical grind of building a product from the ground up.

When artificial intelligence first entered his orbit, he was not immediately convinced. "I remember exactly when I first heard about AI. I had a business developer in one of my companies who was an early adopter and really enthusiastic about ChatGPT. I learned a bit from him and started to be curious. I was a little bit disappointed," he recalls. "I saw the opportunity, but it was not really at a good level yet. It was, let's say, an advanced Google at that point."

He continued using it for fun. Quick calculations. A clever assistant. He could not imagine it changing the world.

“

If you asked me three months ago, I'd say I can do 10 times more work. Now I'd say it's 20 or 30 times more, and it's still growing.

The turning point came as a challenge from a friend in data science who urged him to look closer at Claude. Arjan had a vision: a mental health assessment tool based on scientific evidence, designed to give doctors and patients a clear indication of burnout and mental well-being. As a way to test the model's true limits, he grabbed a complex academic paper, a scientific framework designed to help clinicians assess mental health indicators in patients. Traditionally, translating that methodology into a software tool meant weeks of manual data entry, custom UI design, back-end scaffolding, and significant overhead.

Arjan fed the paper directly into the model and asked it to build a functional digital tool.

"Immediately, I had quite a decent application, at least a minimum viable product. So then I was like, okay, let's finish this. That was the game changer for me. It actually made me a true application."

That single experiment unlocked a new trajectory. He began to see the friction between engineering teams and months of development collapse into a single afternoon. Even the tedious parts of entrepreneurship, hours spent drafting pitch decks, days spent on outreach, the gruelling process of transcribing customer interviews to find statistical trends, were gone. "If you wanted to validate a market through customer interviews, you had to record the meeting, manually plug the data into a management system, and synthesize conclusions. Now? You just record, transcribe, fill the matrix. Done."

Over the span of a year, the speed at which he could move compounded.

"If you asked me three months ago, I would say I am amazed, I can do 10 times more work than I could two months ago. But at this point, I would say it is more like 20 or 30 times more work, and it is still growing rapidly."

“

Be careful of not being used by the people who own the infrastructure. Don't play along with the game, but be ahead of the game.

With this new power came moments of caution. As an engineer, Arjan knows that any system is only as good as its control mechanism. In his view, the real threat to developers is not falling behind on speed. It is surrendering agency to the platforms hosting the models. He warns new users not to become blindly dependent on the tools, but to remain architects of their own processes.

"Be careful of not being used by the people who own the infrastructure. Don't play along with the game, but be ahead of the game."

He advocates for a disciplined, control-first approach: maintain a zero-knowledge stance on data privacy, keep data architecture strictly controlled internally, and build with modular flexibility so you can switch between models and providers as token prices fluctuate and capabilities shift.

"You need to be in control of what is happening. Work with a zero-knowledge approach. You can hire consultants to set up a process, but once it is done, you should not share the data flowing through it with those third parties. And stay flexible across LLMs. Token prices change, systems change, keep yourself agile."

“

For business owners, there is no way of not going into it anymore. The question is how to enter safely and in full control, not whether to go in at all.

As the technology continues to rewrite the rules of building software, Arjan sees no debate left to have. The window for hesitation has closed. The only question is how to navigate the new landscape without losing your autonomy.

"Generally speaking, for business owners, there is no way of not going into it anymore, because if you don't, you are outplayed by a competitor for sure. The question these days is, how do I enter this space safely and fully in control? Rather than, should I go in or not? That question is of the past."