Rachel Schipull
Schip That Code

Schip That Code

Open-Ended Creativity

Photo by Li Zhang on Unsplash

Open-Ended Creativity

How limiting strict objectives can propel us forward

Rachel Schipull's photo
Rachel Schipull
·May 5, 2022·

6 min read

It’s easy to get lost in the sauce when it comes to creativity. When we’re building websites, apps, and other products for customers, there’s a constant push to always be creative.

Though some people may have an easier time grabbing ideas out of nowhere and building on them, I firmly believe that anyone can have a great idea. No matter who you are, there are some ways you can set yourself up for success.

In my next few blog posts, I’m going to be sharing a series on creativity, which will run the gamut from machine learning to popular music. I think you’ll start to see that there are some key components to how human minds create new ideas and some through lines we can draw between what makes us creative and how to protect that space.

Here’s what we’ll cover:

  1. The trouble with objectives guiding creativity.
  2. The real problem with brainstorming.
  3. The power of thinking to yourself.
  4. Living ideas and how to attract them.

It’s going to be four weeks that hopefully will change the way you think about how you think, and that might change the way you approach problems at work and in your personal coding projects.

Let’s get creative.


Now I wouldn’t even begin to imply I know everything about the science of creativity or even how creativity works. So I’m going to “bring in” some experts.

The first piece of primary research we’ll look at is this lecture by Kenneth Stanley called “Why Greatness Cannot be Planned”.

I recommend watching the whole thing for context (and to have your mind blown right out of your skull), however, life is short. So, here’s the abstract to save you some time, (with some notes and emphases from me):

“In artificial intelligence and elsewhere, it has long been assumed that the best way to achieve an ambitious outcome is to set it as an explicit objective and then to measure progress on the road to its achievement.

Basically, we have assumed that if we want to achieve things, we have to set goals.

Upending this conventional wisdom, a series of unusual experiments in machine learning has shown that, for a broad class of outcomes, the very act of setting objectives can block their achievement. More fundamentally, the same so-called “objective paradox” applies not only in computer algorithms but across many human endeavors: Often, to achieve our highest aspirations, we must be willing to abandon them.

Cue the sad trombone, it sounds like setting big objectives can be counterproductive!

As a corollary, collaboration can sometimes thwart innovation by tacitly forcing its participants into an objective-driven mindset. The moral is both sobering and liberating: We can potentially achieve more by following a non-objective yet still principled path, after throwing off the shackles of objectives, metrics, and mandated outcomes.”

Who hasn’t been in a meeting or group project that was actively thwarting their “innovation”? Woof.

Ok, let’s break this down.

A team of scientists set up an image “breeding” website so people could randomly find new interesting images by “breeding” computer generated images together.

They did this by not looking for anything in particular, just browsing and randomly breeding and picking images that looked promising. People were able to start with weird blobs and create cars, teapots, faces, etc.

However, in another similar experiment, they flipped it around and gave people the goal of finding specific items. They couldn’t do it. It was blobs forever. The objective ruined the creative magic of the experiment.

(We'll leave the blogs here for now, but if you want to try pic breeding yourself, by all means, give it a whirl: picbreeder.org)

The other main point is that closely held objectives prune out opportunities for diverse thought and divergence. When a team reaches consensus, it is hard to make a creative decision or promote an idea that goes against the grain. This is dangerous. Over time it really simplifies the ideas coming out of a team, and honestly, the product a company can create.

Even when creativity is encouraged, it’s common knowledge that creativity cannot be forced. Or at least I thought it was.


In my last job, I had a supervisor who set a meeting on the calendar for us all to be “brilliant.” She would say, “Ok tomorrow at 3pm, get ready to be brilliant.”

It struck me as the most inane thing, mostly because if I wasn’t already brilliant at 9am on a Thursday, I probably wasn’t going to be brilliant at 3pm on a Friday, especially not on-demand. These things don’t come and go.

This same supervisor set weekly staff “innovation” meetings for the whole team to brainstorm on topics that easily could have been handled offline. We’ll cover the pitfalls of brainstorming next week, but clearly my former manager thinks that you can plan creative breakthroughs and put genius on the calendar.

It’s no wonder that jobs that set expectations around creative objectives feel stifling. There is no room to blunder, make mistakes, do random things that might seem odd or off course, or deviate from the beaten path. So many amazing and important inventions were a result of mistakes or random "light bulb" moments. No one plans to be struck by a breakthrough, but it helps to not have it on the calendar.

When there is a demand to be creative within certain bounds, it is almost certainly going to backfire. Our minds aren’t built that way and nature isn’t built that way.

Most people do not have the luxury of leaving all goals and checkpoints behind, though. Although Dr. Stanley’s abstract says we should abandon them to “achieve our highest aspirations”, that’s highly unlikely if you are working FOR anyone or FOR anything (like money to live). However, when approaching a creative project, it can be powerful to leave things open ended.

Let's try a little thought experiment:

Imagine you work on a team that is expected to build an app. For the sake of the experiment, let's imagine this app is loosely related to the company's overall mission.

Instead of saying we need an app that does x, y, z, try asking some questions.

  • What would it be like if we built an app that did x?
  • What would happen if it also did y?
  • If it did x and y, could it also do z?
  • Or do we want it to do something else? Maybe q?
  • In fact, let’s try building x and see what happens along the way.

The conversation can be more open ended, but it is clear that something interesting and fruitful is happening.

A more open-ended approach to objectives also gives teams an opportunity to let individuals shine, and that brings us to our topic for next week: brainstorming.

Dr. Stanley has built an entire career around AI design focused on this open-ended approach. In fact, his company is called OpenAI. If you want to hear more about him and his work, I can recommend this podcast interview with him. And of coursehis twitter.

Finally a H/T to Cory Darby (Twitter, YouTube) for sharing Kenneth Stanley's work with me.

 
Share this