“One student replaces three specialists” – and today?

Twenty years ago, the Reutlinger General-Anzeiger published an article about my thesis project. The headline read: “One student replaces three specialists.”

The project involved developing a fully automated laser cutting system for miniature springs made of hardened CuBe – a material thinner than a sheet of paper, delivered in coiled helix form. Tiny orientation holes were meant to guide the cutting process. That was the theory.

In practice, the challenges were significant:

• the material was extremely fragile

• feeding the parts was problematic

• positioning accuracy wasn’t reliable enough

• the laser cut was on the edge of technical feasibility

• and communication between the subsystems was still immature.

Several teams had already spent more than a year trying to crack the problem. No one seriously believed it could be solved any time soon – least of all by a fresh graduate with no industry experience. I mostly received sympathetic smiles rather than actual trust.

Well – perhaps they underestimated me.

At 14, I started experimenting with AMOS and Blitz Basic on my Amiga 500 – and I never stopped coding after that. (Yes, I also played Monkey Island, Bundesliga Manager, Shadow of the Beast and Defender of the Crown.)

At 16, I was working as a student in production at Federn Vogt in Reutlingen, measuring automotive components stamped by 400-ton Schuler presses running compound tools. One stroke every two seconds – a sound you don’t forget.

At 17, I worked at Ruwel in Pfullingen, producing multilayer circuit boards. I went through every department – etching, coating, tinning, drilling, printing – and got to experience all of it up close. 500-degree molten tin occasionally splashing out of the bath – a thrill for young and old.

As a student assistant at Bosch in Reutlingen, I operated second-generation SMD pick-and-place machines, already equipped with visual inspection systems. They watched the assembly process, and I watched them – industrial espionage of the friendly kind.

Maybe I wasn’t quite as green as one might assume of a mechanical engineering graduate. After all, I had also learned the basics of countless hands-on skills around the house, in the garden, and on cars, growing up alongside my father.

So yes, I figured it could be done: a Trumpf laser, a Mitsubishi controller, a smart Cognex camera, the necessary mechanics for feeding and ejecting parts, and a few piezo motors to position the product – all combined into a mechatronic “clockwork.”

A bit of youthful naivety didn’t hurt, either.

The result was convincing: a tiny component measuring 10.54 mm in length, cleanly separated between two 0.35 mm punched holes – later integrated into a new blood pressure monitor, documented in patent US2016367156A1.

And today?

It raises the question whether experiences like this will even matter in the future, given the pace of AI development. Or if they still make a difference at all.

If you think about it, the world today feels far removed from that time. Few people still tinker with their own cars. Most of the companies I mentioned no longer exist. And code? These days, a large language model writes it for you.

So – no experience, but AI as a replacement? Could that actually work? Maybe even faster and better? Perhaps without any freshly acquired engineering knowledge?

This is where the uncomfortable but necessary reflection begins.

Where does the human being remain when systems start to develop themselves?What is our role when machines become better specialists?Where do we find meaning when experience and intuition pale in comparison to computational power?

If we accept that this may already be our reality – or very soon will be – what advice do we pass on to our children?

Don’t just build machines – ask questions.Don’t just master tools – develop perspective.Don’t just optimize – seek meaning.

Last Christmas, I gave my son a fret saw. For his birthday, a programmable robot.Let’s see where that takes him.Maybe right into the spaces where we still need humans – the subtle gaps only people can navigate.

And as tempting as it is to rely on all these new tools and assistants – the truth is: not even the experts agree on how AI will evolve in the coming years. Some speak of approaching superintelligence, others point to limits in data, energy constraints, or a return to traditional modeling.

So what should we do?

Stay calm. Keep learning. And most importantly: go out. Touch things. Feel the problems. Make mistakes. Take responsibility. Don’t just simulate – experience. Because nothing replaces experience – not even the smartest algorithm.

So here’s my advice to anyone just starting out:

Don’t just rely on the next prompt.Don’t get stuck behind GPT. Seek what AI cannot give you: the real world, real people, real accountability.

The rest will follow.

And by the way – thanks to Mark Reich, who years later managed to get me the original digital print plates from that GEA article. A friendly gesture, no agenda – simply because he could.

I doubt a mechanical colleague with a chip in its head would have understood that moment.