The Ogre’s New Clothes
Creative systems at scale
I didn’t always work in tech. I used to make cartoons.
Towards the end of Shrek 3, I had finished all of my shots and had a few days of down time in what’s called a production gap. My team worked on the front lines of the film, plotting, carving and cutting our way through the story jungle.
The schedules were arranged so my team would roll onto the next film while the previous film was still moving through the trailing departments (Lighting, FX, Paint Fix, Completion, Post, etc.). Sometimes the schedules didn’t quite align and we would get an unplanned vacation or we’d jump onto other projects or we’d just get a few days off to relax.
Scheduling the gaps became an art unto itself: too wide of a gap and teams risked getting laid off; too short of a gap meant teams never got a chance to take a break, creatively recharge or mix with other members of the studio.
This particular gap was small, and most people just took a vacation. Instead of chilling, I wandered over to the Art Department and pitched in to help design and paint new costumes. I had studied painting in school and had designed costumes for the ballet, so I was excited to try designing virtual costumes for a CG film. It wasn’t quite the same as cutting clothes against the human sculpture of a professional ballet dancer, but it followed the same initial design process of sketch, block and then paint.
But here’s the thing… although the painters in the Art Department were highly skilled (honestly, they were awesome), they had a brutally inefficient workflow. They were Photoshopping old stills of characters, cutting the arms and legs off, and re-assembling them back in a new pose. Then they were cutting out the background and finally painting on top of all of that.
Some artists were faster and better than others, but it was grueling, thankless work. It took hours of prep work before you could even get to the creative and productive part.
So after painting a character or two, I had an idea: I wrote a series of scripts that grabbed a character, posed it using values from the character animation pose library, and then lit it with a set of canned lighting set-ups, and then rendered the characters both with and without clothes, broken out into layers with beautifully pixel-perfect alphas. The whole process could be triggered from the command line and all those hours of prep work were instantly gone—instead, you just got to work on the fun part.
The team was so happy that a few actually cried. Why? I worked smarter instead of harder to make everyone’s job more enjoyable. Why didn’t the artists do this for themselves? They were experts at Photoshop but never thought to try the 3D rendering system used by the animation team.
It was an organizational problem, a process problem, and an education problem. It took someone wandering across the aisle, who was comfortable in both worlds, to make a change.
I got to be the hero for a moment, not because I was the fastest or best painter, but because I worked across the organization to find a way to help all of the painters get faster and better.
After that, I got excited whenever I saw someone groan on the production floor. It meant that there was another tedious process in place that no one had bothered to solve. This meant another chance to help everyone get better.
Tip: Don’t get faster at using a broken system, instead create a better system so that everyone gets faster and better.
DreamWorks gave me a special award and a cash bonus for writing the system outside of my normal job, but looking back, it was my first glimpse into building creative systems at scale. My switch to Twitter and my love of Medium all make a little more sense now in retrospect.
Apply it to your job
There are two lessons here:
- If you find yourself doing something tedious more than once, don’t get faster at it, instead find a way to automate.
- The best projects are ones you assign yourself.
My most successful projects started as self-assigned efforts. After leaving DreamWorks and joining Twitter, the same held true: The Twitter Mirror, #FollowMe and Twitter Stories all started as self-assigned projects.
Assign yourself your dream project. Then ship it.
