Chengxi Yang, Yalai Pang and Zhehan Xu

Being recognized in the NY Product Design Awards is both a milestone and a motivator. It validates the craft and impact of the work, but more importantly, it pushes us to think bigger, take creative risks, and explore challenges that haven’t been solved yet. It’s a reminder to keep raising the bar while staying grounded in creating meaningful solutions for real people.

The award strengthened our credibility and accelerated our momentum. It validated that our design thinking and execution hold up, which translates to more confidence in balancing the essentials—clarity, usefulness, and restraint. It has also opened more space and support for us to push our AI–human blueprint forward. This recognition has deepened trust from collaborators, tightened our focus, and made it easier to stay aligned on building what truly matters.

Experimentation is how we think: we often start broad, exploring multiple conceptual models of an operating system—each with its own information architecture, interaction grammar, and mental model for AI. Only after stress-testing and learning do we converge on the model that best aligns clarity, capability, and trust. That divergence-then-convergence rhythm keeps our work honest and pushes the invention forward.

Twilight—those liminal minutes where edges soften and contrast sharpens. We drew our product branding from twilight at sunrise. Lume’s signature blue-orange gradient echoes that soft sunrise glow and inspired the name Lume—it symbolizes the dawn of a new era and visually communicates the intelligence of the AI technologies powering the system, while also evoking warmth and curiosity.

The design process is inherently messy and rarely linear. People imagine a tidy path from brief to solution, but real design loops. You diverge, test, discard, and reframe before converging. The strongest outcomes come from embracing that ambiguity—documenting assumptions, building fast proofs, and letting feedback reshape direction. It’s less a straight line and more a series of calibrated bets that gradually reveal the right answer.

We see client expectations and our ideas as complementary, not competing.

The design process is collaborative and participatory: when we invite clients into the work—clarify intentions, align on outcomes, and share designs early and often—we build on their input rather than fight it. That inclusion creates trust, surfaces constraints sooner, and strengthens the solution. Great work happens at the intersection of our creative initiative and the client’s business needs, where vision is translated into measurable impact.

Our biggest challenge was deciding how AI should exist in the operating system—ambient vs. invoked, how context is understood, and how that understanding is made legible to users in a generative, app‑less environment. We resolved it by prototyping multiple representation and behavior models in parallel, pressure‑testing them against multiple use cases, and converging on design principles guided us forward.

Chengxi: Go out. Touch grass, watch a movie, visit a museum. Play.

Zhehan: For me, creativity is both serendipitous and subconscious. When I hit a creative block, I’ve learned not to force it. I step away, take a walk with no destination, breathe, feel, and let things surface on their own. Ideas have a way of growing when I stop thinking too hard about them. I usually come back with fresh eyes and a stronger sense of the path forward.

Yalai: Sometimes I step away from the problem entirely. I’ve learned that giving my brain permission to wander always surfaces connections I couldn’t see when I was too close to the work. Other times, I actively seek inspiration from other fields to unlock new approaches.

Two anchors guide our work: designing trustworthy AI that’s transparent, predictable, and clearly augments human intellect; and our experience designing for commercial user needs for billions of people—clarity, reliability, and measurable outcomes over cleverness. That means legible states, user agency over automation, and solutions shaped by constraints that businesses actually face.

While the definition of success could be varied and personal, we encourage aspiring designers to put in the reps and cultivate taste. Do the work—daily. Ship, critique, refine. Taste comes from exposure and discernment: study great products, dissect why they work, and train your eye to prefer clarity over cleverness. Repetition builds skill; taste directs it.

Chengxi: Muriel Cooper. As a long admirer of her pioneering interaction work during computer’s nascence, I would love to see with the advancement of AI how she would continue on the mission to build responsive tools and systems and demolish the boundary between design and production.

Zhehan: Hella Jongerius. She blends traditional craft with modern technology, creating objects that feel deeply human. Working with her would inspire me to bring warmth, tactility, and subtlety into tech-human experience, push the boundaries of emotionally rich and meaningful design, take bold and disruptive creative risks, and always embrace the "new."

Yalai: Dieter Rams. His “less, but better” philosophy bridges my architecture background with digital design. He proved that thoughtful, principled design can make complex functionality feel effortless.

One question we wish people would ask is: “What vision of human attention guides Lume—and how does it honor agency?”, which sits at the heart of the project’s motivation. Our answer is that Lume treats attention as scarce and meaning‑seeking. By thinking in goals rather than apps, it aligns technology with intention, not impulse. Intelligence becomes a companion—transparent, emotionally aware, and accountable—so users remain the author of their actions while the system clears the path.