Everything around AI is all about “learning new things.” New capabilities, new workflows, new products, new models. Every social media feed and news article that even remotely touches AI evokes the feeling of being on a never-ending treadmill. A treadmill where someone is standing off to the side, slowly tapping the “speed up” button.

Rene Magritte knew the score

The real race isn’t about who can learn the fastest, but who can unlearn the fastest.

The learning race converges on the same stuff, making it easier to see. It’s about the prompts to use, the models to pick, and the tools to use. Sprinkle in a little “are you the person to do it first so you don’t get left behind?”

The race to unlearn is asymmetric, making it harder to see. Different people must unlearn different things. But that doesn’t make it any less important. The fact that it’s asymmetric probably makes it even more important.

There are many examples of this asymmetric unlearning. Maybe because I’ve spent the first half of my career in one world before transitioning to the other, the one that stands out to me is between the techie (engineer, designer, product manager, or anyone at a tech startup/large company) and the non-techie (artist, teacher, performer, craftsperson, social worker, researcher, plumber, etc.).

The non-techie has been told (and told themselves) their whole life that “they’re not good with computers” or “they can’t code.” That’s for other people. They worked with people or created things. It didn’t make a ton of money or come with prestige, but that’s where their talent and interests lie.

The techie has reaped money and prestige for years due to their rare skills. They could take an idea, transform it into software, and that idea-embodied-in-software could scale to the entire world over the internet and make them rich.

With AI and the AI-enabled products that are being built, there are some fundamental assumptions each of these groups have about what they’re capable of that they have to unlearn. These assumptions embedded in their respective identities. They’re “load-bearing,” as the AI kids would say.

The non-techie needs to unlearn the idea that they “can’t build.” That the world of software (and the ability to find and scale to the people that you’re building for) is off limits to them. AI tools may not be better than the top human developer or designer, but it can handle all the coding and design needed to realize the idea. Plus, decades of product development processes are available if they encounter a problem. They just need to ask.

The techie needs to unlearn that their differentiation comes from their ability to build software. For most of their careers (and for new CompSci graduates, their school years), they were unique because they could speak a language (code) that few could. Now everyone’s got access to a universal translator. There were aspects of the techie’s work where they applied their judgment and taste but mostly, it was enough to take someone else’s idea and translate it into software. This is not to denigrate people who can do this, by the way. It’s a hard-won and very impressive skill. Its value has just tanked. 

Since everyone can access tools to build whatever they want, differentiation will come from ideas, judgment, and taste. The skills that techies haven’t needed to practice (they could make money and prestige without them) are the same skills that non-techies have had to practice to set themselves apart enough to make a living.

I have no idea.

There’s no guarantee that anyone will win just because they’re in either group. The things that each group has to unlearn are related to identity, and changing beliefs related to identity is one of the hardest things for a human to do.

What I do know is the race is asymmetrical. They’re running on different terrain. And when you’ve got a race like this, you can learn a lot by looking at the differences in the terrain itself.

last metaphor, I promise

If I had to bet, I’d bet on the non-techie. 

Non-techies have been close to their problems for years. They’re the customers that the techies have been building for. Non-techies understand them because they live them. They’ve most likely hacked together a solution or waited for a company to build one. If their problem is hyperspecific, this never happened (no money in it) and they resigned to deal with it. They’re in a race to realize that they don’t have to wait anymore. Now they’re the ones with the power to solve their own problems.

The techies (excluding founders) have built a career and life around executing other people’s ideas. They’ve developed craft and skill around beautifully realizing (or managing others to realize) the solution to a unique problem. It’s been lucrative, so there’s a natural desire to hold on to this skill as an identity. First they need to let go of the identity of being someone who can make a living executing others’ ideas. Then they need to answer the question: what problem do I feel uniquely compelled to solve?

So what’s anyone to do? As usual, the kids show Bluey points to the answer:

the wisdom of a children's cartoon about talking dogs in Australia

Stop trying to run the race to learn. Start running the race to unlearn.