I’m trying to figure out AI’s purpose. I know I can use it for building software, but how? There are approaches like vibe coding or spec driven development. But that doesn’t answer my deeper questions. I can use AI to prototype or write a spec, but what should my relationship with this technology be?

I recently read Zen and the Art of Motorcycle Maintenance and it shook some things loose. Robert Pirsig wrote it in 1974. There’s nothing about AI. Nonetheless, Pirsig’s exploration of Quality and our relationship with technology resonated with me. In one section, Pirsig describes the importance of being stuck 1. He describes being unable to fix a motorcycle because of a stripped screw. Most of us dread these moments where we are blocked by something as seemingly inconsequential as a measly screw.

Yet these are the great moments of growth and discovery. Figuring out how to get the screw loose is what leads to greater understanding. You might even find a new way to unstick a screw that nobody ever considered before and end up patenting a billion dollar idea.

If your mind is truly , profoundly stuck, then you may be much better off than when it was loaded with ideas.

Stuckness shouldn’t be avoided. It’s the psychic predecessor of real understanding. An egoless acceptance of stuckness is a key to an understanding of all Quality.

We never stay stuck. It’s hard to become completely stuck on a problem and never become unstuck. Eventually you will find a way forward. Being stuck leads us to the disoveries that move us forward. Those are the discoveries that move everyone forward.

If we consider stuckness to be the “best possible situation”, then maybe the purpose of AI is to help us become stuck; it does not exist to avoid becoming stuck. Note that I haven’t described AI’s role in becoming unstuck. Avoiding stuckness and getting out of unstuckness are distinct activities.

Given the current obsession with using AI to become faster, how can I reasonably think that we should use AI to become stuck? Isn’t stuckness what makes us slow? I believe that stuckness is what makes us fast. We rarely get stuck on a problem that never presents itself again. And once a problem goes into production it becomes harder to resolve. The sooner we get stuck on the big questions, the earlier we resolve them, and the less time we spend on them later. I love using AI to find inconsistencies with an interface or discrepencies between different user expectations.

I haven’t toyed with this idea long enough to define its practice or write descriptions of which tasks best find stuckness and which ones are simply avoiding it. However, I do think that AI as the route to stuckness is the path to greater Quality.

  1. Pirsig, Robert, Zen and the Art of Motorcycle Maintenance, p 298, 1974. ↩