Making Products

Arduino was invented in 2002 to simplify product prototyping with breadboard circuitry and coding.

By standardising PCB hardware into a plug-and-play board, and wrapping the software in a user-friendly version of C++, it lowered the technical barriers to software and electronic engineering. Electronics became an accessible tool in the designer's toolkit.

While it was great for giving designers the ability to imagine how a product would work, there were obstacles between Arduino prototypes and the requirements for a shippable combination of hardware and software. The clunkiness of both hardware and code meant it was limited to experimentation or proving out a concept.

Although making hardware fit for manufacture sits comfortably inside the skills of the designer, shippable software still demanded collaboration with a coding specialist.

A significant friction in product development is the communication gap between designer and programmer: the vision of the product and its behaviour getting lost in the technical execution of the software, where logic may be prioritised over product experience.

Durrell highlighted his previous experiences with this problem when describing the development of his own products. He talked about the risk of working with programmers that might not "naturally understand the service" or the desired experience of the product.

Now he navigates around this problem with vibe coding, using AI to generate the bulk of the code, then testing the results and tweaking where needed.

He admits it's a process of trial and error but this is the kind of iteration designers are already adept at. Although the to-and-fro of creating the programme isn't instant, it's quick enough, and has the advantage of offering total control over every nuance of the product experience. This ensures that the final product is a pure reflection of the original design intent.

Using AI to translate ideas into code means the software and services living inside a product are no longer abstract requirements handed off to others, but tangible materials the designer can manipulate directly, themselves.

AI input doesn't stop at coding.

It can define the components and configuration of the Printed Circuit Board required to house the programme, effectively building the circuit for you. In combination with CAD, generative design algorithms can quickly iterate through structural variations to optimise for weight, strength, or aesthetics.

This potentially gives designers control over the entire design and manufacture process of a product, from the inside out, from concept to delivery.

Every design skill ever developed is deliberately intended to increase the resolution of an idea and make it easier to understand. Through visualisation and model making, designers bring others closer to the vision inside their heads.

AI short-cuts this even further, bringing designers — and their audience — closer, faster, to their own vision.