Tech x Design | Rendering design concepts with AI
In the book ‘Show your work’ by Austin Kleon I was introduced to the ‘Sturgeons Law’ thats says that “90% of everything is crap”. Then only 10% of my ideas is not crap. So in principle the faster I can make 100 concepts — the sooner I would have 10 that is not crap.
AI for rendering
I decided to test out vizcom to render out a design concept. I believe it is important to be a reflective AI user. What part of my creative work do I want to delegate to an AI?
For me early concept visualisations could be such a thing. It would be nice to save the time of modelling and rendering and use the time sketching. As the concept then matures I would properly like to regain control and do the modelling and rendering myself.
The concept used for this exploration was:
a lamp inspired by a sail
Sketches
For the sketches I chose to do some simple vector drawings using figma. I find vector sketches to be a fun way to work with concepts. They allow me to work with precise dimensions similiar to 3D, but with a speed similiar to analog marker renderings.
Vizcom renderings
I managed to get some renderings out that was pretty similar to what I had in mind. And the the realism of the concept is quite high. My biggest concern is that it might be too high? In 1997 Houde and Hill wrote one of my favourite academic papers about ‘ what does prototypes prototype’. One thing they stated was ‘know your audience’. I think that is as true for visualisations as prototypes.
Who is this rendering intended for? Is it for you or an internal team to discuss? Then it might be fine that the realism is high, but with some weird visual mistakes.
If the audience is an external client a high realism might be damaging. The client might expect the concept to be more finish than it is. Maybe you will spend time discussion the details instead of the overall concept.
More abstract sketches
My first sketches was a very direct interpretation of a sail. For the second round I would like to go for a more abstract version of a sail.
I decided to go for this abstracted sail design.
Then it was time for vizcom to its magic.
This time I wasn’t able to get vizcom to provide me with a rendering I was satisfied with. I wasn’t able to stir towards what I had in mind.
Conclusion
I am certain AI will find its way into the design process. What I am missing most is more control. What I like about vizcom is that it have the potential of allowing me to use sketches as input. This is much more natural than pure prompting. I am however not sure if AI is more efficient to convey a early stage concept than a sketch combined with a moodboard.
