This Year’s Model, Designed by Machine

Generative Design

Whether it’s chess or designing a rocket, a lot of design is simply getting the numbers right, and that’s the kind of thing computers are good at.

From stone flints to nuclear reactors, humans have been designing things using imagination and trial and error for 2.6 million years (according to fossil records). We’ve long used machines to help us with the finer points but we still need to face unforseen variables, see what works and what doesn’t, and try different things to arrive at the design we think will perform best – and as examples from the leaning tower of Pisa to plastic clamshell packaging prove, we still get it wrong at times.

The practice of 3D modelling has been a great intermediate step. A designer working in 3D will build a virtual car, house or aeroplane on a computer, modifying and adjusting while searching for the best way to fabricate it in the real world, given the constraints of materials and budget.

Now we’re entering an age where we can just tell a computer what we want and let it come up with the best design by itself.

The examples are nearly endless. Imagine you’re designing a rocket to take astronauts into space. You know how much fuel you need to carry because of the weight of the vehicle, the speed it much reach to escape Earth’s gravity, the components, life support systems and computer circuitry that has to fit inside, how wide or pointy it has to be to minimise air resistance, and other factors that can number in the millions, depending on the project.

Click here to read the rest of this story