Our First Computer Generated Product Idea

The genesis of Novaline was a Kickstarter campaign for a computer generated idea. The hypothesis underlying our approach was that computers and humans working together would be better than either alone. We supposed that, when it comes to creativity, the computer would help break through mental filters and convention (one place where being context unaware actually helps).

The idea generator, Blender, simply takes a seed term as in input and outputs stimulus words meant to spark a creative idea. Our experience is that very simple mechanisms can spark creativity almost as easily (and sometimes more easily) than more sophisticated algorithms. In late 2016, all Blender did was scrape Google results, make a language model and find words near the median “distance” from the seed term. This was meant to strike a balance between novelty and relevance. As it turns out, human minds are adaptable enough that you can almost throw words out at random and find a creative idea from them in combination with a seed term. In fact, creativity researchers even have a name for this: forced association.

The seed term we used was “work light.” The “light” part came from the fact that Gavin had just finished a product involving lighting and knew a thing or two about it. The “work” part came because I had a hunch it would place us into a higher value category. We stuck that term into Blender, and out popped 10-20 phrases. One of them was “try combining a work light with a whiteboard.” We shortlisted this phrase, along with a few others, and exercised our powers of forced association. We came up with a backlit whiteboard for better visibility in large rooms, and a number of other things I can’t remember anymore. At some point, I ended up Googling the phrase “lightboard,” and up came lightboard.info from Michael Peshkin, an engineering professor at Northwestern. Just as it seems, the site provides information about lightboards, which are just transparent writing surfaces with light shining through them. You don’t even need the light, it turns out, but it does make the writing pop a bit more on camera, which is the primary medium for these devices. The point of a lightboard is that the viewer can see your face while you write, which some people prefer (though my brother, who has a degree in instructional design, tells me it hasn’t been shown to improve learning outcomes.)

On Prof. Peshkin’s website, we noticed some outfits selling lightboard systems for obscene amounts, and we thought we could do a desktop version for a reasonable price. We figured that since people take the time to build their own, we’d find some buyers for a pre-made system. Gavin designed the prototype, and we did a Kickstarter campaign. The Kickstarter campaign succeeded, but that’s not saying much, because our goal was $500. We had no clue what we were doing. We chose to go in a really odd direction for the video, and I suspect that hurt us (although that’s where the name Lightboard 14T came from). In addition, we originally decided to price and sell the Lightboard as a single unit: a base and a big piece of tempered glass. Tying them together increased costs a great deal. (In fact, when someone from Indonesia backed us, our initial excitement turned to dread when we discovered the costs of shipping a heavy piece of tempered glass across the world.) By the end of the campaign, it was obvious we should have allowed customers to buy the base alone and find their own glass. This is one of those discoveries that produces embarrassment in me. I feel as if I should have known all along that we should have done that, but it eluded us.

I’m understating the success of the Lightboard a bit. Long after the campaign ended, Gavin continued to receive requests for Lightboards, and he made a little bit of money on the side from them. It’s possible we could have built a small business on that product, but our motivations were elsewhere.

Even though we didn’t launch a blockbuster, we learned a lot about Kickstarter, the value of keeping things simple, the costs of international freight, and the art of using computers to brainstorm. Most importantly, Novaline was born.

Jason Bell