I was trained in real-time multi-agent systems before they were digital. Let me tell you about a thing I worked on, because that's the fastest way to explain how I think. When I was working on Amazon's first generative AI chatbot (Amazon Q), we did the obvious thing first: we watched people use it. Not demos, but real people, real tasks, real impatience. Very quickly, one complaint kept showing up, and it wasn't subtle. People were annoyed. Genuinely annoyed. They kept saying some version of: "Why do I have to tell this thing the same stuff over and over again?" At first glance, that sounds like a UI problem. Maybe better prompts. Maybe clearer affordances. Maybe a nicer way to surface past chats. But that explanation didn't survive even a little scrutiny. The frustration wasn't about finding old conversations. It was about something deeper: people expected the system to remember them. Not perfectly, not magically—but in the way a human-like intelligence would. That's when things got interesting. So I stopped thinking about screens entirely. There was no screen to show. The question wasn't, "How do we display memory?" It was, "What kind of memory do people expect when they believe they're interacting with something intelligent?" Humans don't replay transcripts. They form impressions. They remember preferences. They generalize. They forget irrelevant details but keep what matters. Once you frame it that way, simply storing chat logs and searching them later is obviously the wrong solution. That's data storage, not experience. No screenshots came out of that work. Nothing Dribbble-worthy. But what did come out of it was far more valuable: it directly inspired the creation of a core memory service that now underpins Amazon's AI products, services, and tools. That's the kind of work I love. The kind where the most important design decision happens before there's anything to design in the visual and UI realm. Where understanding what people expect from an interaction matters more than explaining what's already broken. And where the payoff is a product that quietly feels smarter, calmer, and more trustworthy. That's the craft of product design. If you ask me what I design, that's the answer. Not the screen, but the thing the screen is trying to make possible.