Many years ago, Brian Eno came up with a deck of cards called “Oblique Strategies.” Each card had a provocative statement on it. Examples:
- Is there something missing?
- Work at a different speed
- Use an unacceptable color
The idea was to hit a creative roadblock, draw a card, and hopefully it would be enough to inspire your creative mind into action. Brian Eno and many other artists have famously used these cards and techniques like it time and time again.
About 15 years ago, I discovered a book called Lateral Thinking, which talked about similar things. When you’re thinking in a black and white way, where one step leads to the next step in a process, it can be easy to get out of a rut simply by throwing a curveball, like:
- How would this product be used in a zombie apocalypse?
- What would Jay-Z do?
- How can we make the absolute worst version of this?
The same process in your brain kicks in. Rather than thinking about the obvious next step, you’re now layering in something new, which drives creativity. And one of the most beautiful things I’ve learned about this technique is that the breakthroughs don’t end up mattering for the edgecases like zombie apocalypses, Jay-Z alignment, or bad versions of products. They end up directly relating to the project itself, even if it takes a bit of outside-the-box thinking to get there.
Much of my career has centred around these sorts of techniques. I have a collection of approaches that I can use in different situations to get different results. I’ve taught classes using them, I’ve driven my own teams to be more creative, and I’ve written lots of articles, which have led to some books. This is how I stand out in a crowded field, so I’m always on the hunt for more of these strategies. Which is what is so exciting to me about chatGPT, and why I think so many people are missing the boat.
Today I was at a restaurant with my two 11-year old daughters and their friend. I asked if they wanted to play a game of D&D while we waited for our food. Already, that’s a pretty fun idea. I can think of many times in my life where I just sat, waiting for food, and if I hadn’t brought a book I’d just read the menu for entertainment. I would have loved it if the adults in my life were fun enough to have spun up a random game of D&D a time or two.
I could have run the game on my own, of course. I could have asked each girl what sort of adventurer they wanted to be, then I could have presented a scenario to them to react to. I’ve been a dungeon master (DM) before, and it’s tough! You need to be able to think on your feet while you build the story around the decisions each person makes. It’s possible, but it’s not easy. I’d say most people aren’t well-suited to it without some amount of practice.
But with chatGPT, I could go to the next level of creativity. Rather than coming up with the name of the village, the nature of the threat, and the vibe of the adventure, I could offload that to the AI and think at a higher level. I said something like this:
Hey, I have three 11 year old girls waiting for their food and we want to play a quick game of D&D. Please come up with an enjoyable adventure so they can all play together. Start by asking each of them one by one about their character, perhaps giving them multiple choice answers to choose from.
And away we went. For the next 30 minutes we suited up, addressed a threat, made decisions, and had a whale of a time. At the end, I even asked for a keepsake image of the adventurers and the large giant they befriended, who happened to look a lot like a dragon crossed with Ed Sheeran. (You had to be there) It was a truly magical experience. But I want to analyse why, because this is important. Let’s look at the guidance I gave chatGPT, and why it represents next level creativity.
- 3 people (create the characters)
- 11 years old (make it kid-friendly)
- Waiting on food (make it quick)
- Enjoyable + play together (rather than earnest and solo)
- Start by asking each one (to keep everyone engaged)
- Give them multiple choices (to avoid writer’s block)
I could have gone a different way, like:
chatGPT, tell me a joke
Or even if I knew I wanted to play D&D, I could have just said:
chatGPT, set up a game of D&D for three people.
But neither one would have hit the same quality bar that I was able to achieve. This is why I am surprised and a bit bemused when people make jokes about “pressing the creativity button” when writing off the potential of generative AI. When you strum a guitar, there are hundreds or thousands of different techniques you can employ. A poorly done strum sounds pretty much the same no matter who does it. But someone good at the guitar can make art, and the very best guitarists somehow find a way to channel themselves into it in a way where they sound like no one else. The same is true with effects, pedals, and music production techniques. Anyone can press an auto-tune button. But some producers are able to take the millions of different creative options and form them into something transcendent. Such is the case with generative AI.
Think back to 100 different restaurant visits over your lifetime, each with their own waiting time for the food. Imagine ranking each experience on a bell curve from forgettable to memorable, or from boring to awesome. For me, today’s experience might not have been the best I’ve ever had, but it was certainly in the top ten. So what does that mean? What can we learn?
First, the “floor” for great experiences will keep rising. If I picked my top 100 experiences waiting for food from the 80s until today, most of the best experiences would be in this decade, specifically the post-chatGPT era. I don’t know that many would make the top 100 before the age of the smartphone. Maybe a few would sneak in for exceptional reasons, like spotting a celebrity in at the Cheesecake Factory in 1989 or something. But the average experience would skew heavily towards smartphones, google, and now chatGPT enabled moments.
Second, that doesn’t mean we’ve gotten less creative. It’s because we’ve gotten much, much, more creative. And yes, that’s in huge part because of technological improvements. But so what? I don’t think a DM is failing at being creative when they look something up in the D&D Monster Manual, or double-check something in Google, or have an app to track magical effects during a scene. I think of that DM as well-prepared to bring their full creativity to bear on the game because they’re offloading the fiddly bits to technology. Some of the best artists in the world have leaned into new technologies, and that’s no small reason why they are considered great artists in hindsight. The act of creation requires a pioneering spirit. Staying still creatively isn’t brave or noble.
I would like to teach a class about this. Not about chatGPT per se, but more broadly about using lateral thinking and creativity-boosting techniques to create art. We’ve been given the greatest tool in human history for dreaming up new concepts and ideas. Not boring ones that emerge from pressing a creativity button, but truly novel and human insights that you can only get from pairing an open-minded human to a huge breakthrough in technology.
Postscript: when I finished this essay, I threw it into chatGPT to find grammatical issues. It found 1 typo, called out 1 error that’s not an error, and had 3 suggestions I discarded. Imagine if I had accepted all 5 because I think AI is always right, or rejected all 5 because I think AI is always garbage. Both would be objectively worse than the way I used it. Those who master this blend, like grammar checkers or Google before it, will create more, with higher quality and confidence, than those who don’t. It’s an exciting time.