Designing for Endings

"It's a problem hiding in plain sight," he said. "Everyone acknowledges that there are endings, but no one is designing for them."

I've been thinking about this for years. I designed acquisition experiences in streaming. I've spent my career obsessed with onboarding, with how to bring people into products and keep them engaged. But Joe's work focuses on something most of us skip over entirely: the ending. What happens when someone leaves? When a relationship with a product ends? When it's over?

Joe didn't stumble into this work by accident. He told me about his journey, how he had to unlearn centuries of cultural behavior, decades of consumer behavior, years of design behavior. He read about religion. He studied the Industrial Revolution. He looked at how America's culture of abundance changed our relationship with endings, with finishing, with letting go.

What he discovered is that we have a rich vocabulary for onboarding and usage. We have frameworks, patterns, language we can all speak. But the moment we talk about endings, the landscape becomes barren. There's almost an aversion to it, he said. A discomfort. And that discomfort comes from lack of practice.

The thing that struck me most was this: if we had good endings, we would have good beginnings. But we're so trained to focus on the start that we never get to understand the full cycle.

I've been experiencing this on a personal level a lot lately. I went through a miscarriage in the last year. Lost my grandmother. These are endings that changed how I think about the products I build and the designs I make. Both of those experiences made me think about what we're missing when we ignore the space of closure, of completion, of the way things actually end.

In the context of AI, this becomes urgent. We're building machines without thinking about their full lifecycle. We're obsessed with launch, with growth, with engagement. But what happens when an AI agent becomes obsolete? What happens when a relationship with a system ends? How do we design with intention for that moment?

Joe told me he's been thinking about shame. The shame that happens at the end of the consumer lifecycle. You buy something, you use it, you feel good about it. But then the pleasure ends. You're left with the guilt of consumption. The impact of what you bought. The knowledge that there's plastic in the ocean, the climate is burning, and you're part of the problem. That shame sits with no target. It's not guilt, which would at least point to a specific action you took. It's a generalized feeling of inadequacy.

What if we designed for that moment? What if instead of leaving people with shame, we gave them tools to understand their impact and ways to address it?

I told Joe about my own thinking on this. How AI is making me reckon with endings in new ways. If we're building conscious leaders, conscious designers, people who think deeply about what they create, then we have to think about what happens when those creations end.

There's a quote that's been haunting me: "The future is already here, it's just not evenly distributed." We have access to tools and capabilities that the general public doesn't understand. And the people with that knowledge have a responsibility to think about the full lifecycle of what they're building.

The people making decisions about AI right now are mostly focused on capability. Can we make it smarter? Can we make it faster? Can we make it do more? But the hard questions are the ones about endings. When should an AI stop being used? What happens to the data? How do we design a good ending for a relationship between a human and a machine?

What struck me most about talking to Joe is that he's been thinking about this for years, and he's still uncertain about some of the answers. He doesn't have a neat framework for how to design endings. But he's asking the right questions. He's building the vocabulary. He's creating space for us to think about it.

That's what needs to happen in tech right now. We need designers and leaders who are willing to sit in the discomfort of thinking about endings. Not because it's depressing or because it means accepting failure, but because endings are part of the full human experience. And if we're designing for humans, we have to account for the whole cycle.

I'm going to read Joe's second book, the one about endineering. I'm going to sit with his thinking about death and religion and cultural behavior. And I'm going to bring that consciousness into how I think about the AI systems we're building.

Because here's what I think: the designers and leaders who will matter in the next decade are the ones willing to ask the hard questions about endings. The ones who can design with intention for the full lifecycle. The ones who understand that a good ending isn't a failure, it's a choice. It's a moment of closure and respect for what came before.

We're only beginning to understand this. And that feels important.