Welcome to The Prompt

Amber Bravo , Director of Content & Editorial, Figma

AI promises to reinvigorate the way we design and build, but how we realize that potential is still an open question. By answering a series of prompts, leaders in this rapidly evolving field take stock of where we stand, what the limits are, and where we’re bound to grow.

At its most basic, prompt engineering is the art of getting the right answer by knowing how to ask the right questions. It is not dissimilar to being a good interviewer—or putting together a magazine, for that matter. Really any act, creative or otherwise, starts with a prompt. The real art lies in thoughtfully designing the input—to give the right context, framing, and guidance—so that a Large Language Model (LLM) can arrive at the most satisfying result. Because an AI’s brilliance is inert, it needs coaxing to shine. Lucky for us humans, our impulse to question is innate.

It is in the spirit of designing a well-crafted prompt that we set out to assemble this collection of essays and interviews. We wanted to learn from folks both inside and outside of Figma who are working across design, engineering, product development, and the built environment to bring a new “legibility to these complex systems,” as multidisciplinary designer and roboticist Madeline Gannon describes it. Through using the technology, they are finding better ways to prompt the technology. They are teaching the Als the mechanics of good design and exploring ways to solve big challenges across industries—from manufacturing to residential housing. At the same time, they’re setting their sights on more attainable horizons, focusing on the value that AI can reliably deliver right now. “We are in the telnet days of LLM and diffusion model interfaces. There are so many possibilities for new interaction models to explore,” says Figma co-founder and CEO Dylan Field.

On the topic of technology taking our jobs or replacing us—particularly in the realm of design and engineering—even the AIs are skeptical. Claude, Anthropic’s notoriously articulate LLM, is the first to admit it. “There may be some contexts or situational nuances that I could miss without having a truly human experiential base to draw from,” it says. This is where the human in the loop matters most and makes all the difference between what’s good and what’s great. It’s a humbler, more human vision for the AI era. As Kris Rasmussen, CTO at Figma, sees it, “There’s so much more to being an engineer than just outputting code. It’s also about knowing which problems to solve and how to solve them.”

Our ability to process information that an AI cannot—the subtleties of context, the nuance of style, and maybe most critically, the pull of intuition—makes it so that the more we use AI, the more we poke and prod it to produce better results, the more we understand its potential and its limitations. We also get clearer about our own strengths; we start to see our human capacity in stark relief, and the reality that this technology needs us just as much as we need it.

What is good design in the age of AI?

As AI transforms the product development process, long-held principles of the design craft are more salient than ever. Figma Vice President of Product Design, Noah Levin, and his team dig into what guides their work now that the baseline has changed.

Read more

Why are we so afraid of code as a commodity?

On the development side, argues Figma Chief Technology Officer Kris Rasmussen, it’s less about writing code more efficiently and more about the new problems engineers are uniquely positioned to solve. His team sounds off on the ways in which creativity and expertise can stretch in the wake of AI.

Read more

What is minimum viable data?

We’re so focused on what AI can produce that we often overlook what informs that output in the first place. Ovetta Sampson, Director of User Experience Machine Learning at Google, unpacks how high-quality, equitable data ultimately yields better tools.

Read more

More on The Prompt