Get in touch
Design

What I've Learned Making Art with Computers

How to make something using machines.

For as long as I remember, I’ve been using my computer to help me make things. I’m a product designer here at Postlight, and I make art using digital tools and processes I develop over time. You can see some of my projects here. All this time sitting in front of glowing screens has made me appreciate what this medium has allowed me to create.

These machines have also been the source of much anguish and pain, trying to make the machine do the thing that I want it to do. Today I’d like tap back into those moments and share three ways that I’ve worked through that frustration and come out on top.

My first workstation, circa 2003

Don’t Over-Engineer

My moments of peak frustration have almost always been caused by over-complication. If you’re troubleshooting a programming task or trying to figure out how to , ask yourself: Does this Matter?

Sometimes the task does matter and it’s just a part of getting through the project. Other times, however, a technical task is a distraction to your end goal. Consider manual work flows to get a task done quicker. For instance, rather than spending the time writing (and troubleshooting) a script to efficiently do a menial task, it might be quicker to just do by hand.

This affects tooling as well. Consider a recent project of mine called Cards Against the Machine. I had this idea of using Cards Against Humanity cards as a data source for an art project, ideally incorporating some sort of machine learning library in the mix. Cards Against the Machine uses machine learning to computer-generate images based on the text of Cards Against Humanity cards.

I experimented with a lot of different techniques but got hit with blocker after blocker, to the point where it wasn’t really fun any more. I refocused, recognized that these arbitrary technical requirements were getting in the way of my work, and started using a tool called Runway to handle the heavy lifting.

Machine-generated art from Cards Against the Machine

Even though I didn’t “make” the models that Runway was loaded with, it didn’t matter. What mattered was the concept of using Cards Against the Machine data to make something interesting, and it worked!

Spinning wheels mean project death. Spending too much time on a technical task can risk making your project go from almost complete, to never finished.

Embrace the Aesthetic of Technical Limitation

Unsavory production artifacts found in today’s media will become sought after and romanticized when they’re no longer around. I’ve found that embracing these artifacts and limitations early can be useful in the ways you think about making things.

Consider Earl Sweatshirt’s “Grief”. If you listen closely to the beat, it sounds like a low quality MP3 you downloaded from KaZaA back in 2003. And that’s the point, it was a technical constraint that turned into stylistic tool for future work.

Rather than fighting against constraints in hardware, browser capabilities, or internet connection speeds, just make your project purposefully lower-fidelity.

You see this all the time in video games too. Low-poly games, once a source of frustration for console game designers in the late 90’s, are now widely used for their ease of rendering and frame rates on even the worst computers.

Minecraft 1.1 Update

Check out the latest Minecraft update. Jeb and Jon have been working diligently to bring you new features and bug fixes. This video highlights the new languages, sheep gaining wool, apples falling from trees and more. Hat Films Youtube: http://www.youtube.com/user/HaatFilms Hat Films Twitter: http://www.twitter.com/hat_films Hat Films Website http://www.hatventures.net

And if we go even further back in video game history, you’ll see this pattern on games like ZZT. Tim Sweeney released ZZT in 1991, at a time when more graphically advanced games like Zelda, Sonic, and (of course) Battletoads were on the market.

ZZT by Epic MegaGames

A short six minute clip from the popular action/adventure game released by Epic MegaGames in 1991. If the video has you in the mood to play, you can download the game at http://www.demu.org

Instead of competing with more graphically intensive games, he embraced the aesthetic of being technically limited. The simplicity of its level design and UI also allowed for less experienced, independent programmers to extend it as they see fit. As a result, the look and feel of ZZT is iconic and doesn’t feel like it’s falling short of an intended target.

Figuring out how to champion and own the limitations of your medium allows you to focus deeply on the core concept of your project than fretting about the details. Those old, dithered Instagram photos you took in 2013 on an iPhone 4 will be cool again, mark my words.

Recycle Your Work

Finally, plan for your projects to be built on top of each other. Everything you make, even concepts and ideas, should be recycled in future works.

One of my favorite creative recyclers was the architect Frank Lloyd Wright. Let’s take his Ennis House for instance, one of the several houses he built in Los Angeles in the 1920’s. At the time, Wright had been experimenting with using pre-cast concrete as decoration in his houses at a time when concrete wasn’t particularly interesting to architects.

Wright and his team of builders treated this technique (among others) as its own distinct project. They continuously perfected and iterated upon this approach for years in such a way that could be efficiently incorporated into future work.

By standing on the shoulders of your own work, you can leverage all of that time and energy spent figuring out a piece to what eventually will be a much larger and more interesting puzzle.


These are just a few things that I’ve learned over the years and I hope you find them useful! Good luck on your endeavors and don’t forget to have fun. 🙂

Luis Queral (@luismqueral) is a Product Designer at Postlight. Wanna talk machine-learning and computer art? Drop us a note: [email protected].