Harnessing AI to build better UI

A fascinating article (and set of demos) about how generative and improvisational AI techniques could help us invent better interfaces, and better ways of thinking for humans:

At its deepest, interface design means developing the fundamental primitives human beings think and create with…

We’ve described a third view, in which AIs actually change humanity, helping us invent new cognitive technologies, which expand the range of human thought. Perhaps one day those cognitive technologies will, in turn, speed up the development of AI, in a virtuous feedback cycle.

So not just computers quickly generating lots of options based on existing pieces, but helping us think up new ways to frame the questions, and build new tools to explore them. Your next design colleague could be a machine.

The biggest constraint in most product development is not technical capability or design skill–it’s willingness to talk with customers.

Design by inquiry: Can this be a question?

I recently started a new job with new brilliant, experienced colleagues, and it’s been difficult to make helpful contributions while I’m still learning about the problems we’re working on. Often when I propose a solution it turns out to be already considered and rejected, hopelessly naïve, or entirely misguided. And when asked for my opinion on the ideas of others, I sometimes freeze up and stammer out something noncommittal.

I’ve found that the most useful technique is to constantly ask myself, “Can this be a question?” Specifically, I take whatever proposal I was about to make, and turn it into a question for others.

For instance, if I think we should change a design element from blue to green, I might ask “How did we decide on blue for this?” or “How well is blue working here?” If that doesn’t lead anywhere, I could continue by asking “What are the goals of the color choices?” followed by “Would any other colors do that even better?”. Even if we don’t end up making it green, we’re likely to end up with some improvement, and I’m certain to learn something along the way.

Asking questions like this–something I call “design by inquiry”–has several benefits:

  1. It encourages others to share their thoughts and ideas, and puts them in a creative mode rather than a critical one. Often when people hear a strongly-presented idea they feel responsible for pointing out its flaws rather than building constructively on it. And design always benefits from more diverse perspectives.

  2. It gives the people with the most context–they’re asking the question, after all–the opportunity to answer it themselves. If an engineer comes to me with a problem, they’ve already started thinking about it. I’d like to hear what they’ve considered already, and what they feel might be best now.

  3. It can open up an overly-constrained problem to new opportunities. More often than not, the difficulty in design comes from solving the wrong problem, and restating the question gives everyone a chance to reframe the problem and make sure you’re still looking in the right direction.

In several ways, this is similar to the Socratic method, which is often employed in order to discredit a hypothesis or proposal, and sometimes characterized as “acting dumb”. However, design by inquiry comes from a place of open creativity and “actually being dumb”–as designers always are when starting a new project.

One of my mantras is “be the dumbest person in the room”–to make sure I’m always learning–and that means asking a lot of questions!

“Style is knowing what play you’re in” – John Gielgud

A list of human universals

Things that everybody does, in every culture worldwide: Human Universals

For example:

  • Dance
  • Jokes
  • Language
  • Laws
  • Planning for future
  • “Snakes, wariness around”
  • Tickling
  • “Weather control (attempts to)”

via Alan Kay, who says “Suppose you want to make a lot of money. Well, just take the top 20 human universals and build a technological amplifier for them.”

Learning the basics

I think learning should be about learning the basics in all the fields and learning them really well over and over. Life is mostly about applying the basics and only doing the advanced stuff in the things that you truly love and where you understand the basics inside out. – Naval Ravikant

When I taught an introductory design class at Stanford, I finished by telling the students, “That’s it! That’s all you’ll ever need to know about doing design. Now, go out into the world and spend the rest of your lives trying to actually do it.”

Get a grip on it

Innovation in packaging:

I love how the handles swing out to let you feel what it’s like to ride with them.

Jonathon Keats on Google Glass design

It’s not often one of my design heroes comments on something I worked on, so Jonathon Keats’ comments on how to redesign Google Glass were interesting to read.

To navigate this uncanny valley in time, the designer must either create something so futuristic in appearance that it arrives from beyond our collective vision of the future, or something that looks and feels like a natural extension of the present.

It mirrors one of my own new product design lessons learned from Glass (which admittedly wasn’t meant to be a huge consumer product)–either create something completely new to the world, or replace something that people already use. Simply being faster or better isn’t enough if people still need their existing solution.

Trusting the artist

Ahmet Ertegun – arguably for a long time the greatest record executive of them all – told me that unless you’re 100% sure the artist is wrong, go with their vision. – Jason Flom

True for designers as well, in my experience.

Design Sprint Kit

Great set of resources for running a sprint: Design Sprint Kit.

This is the highly, highly distilled and refined version of what designers at Google (especially Knapp, Kowitz, and Zeratsky) have been developing for years.