I don’t think I’ve met any product designers this summer who really think that prompt/chat interfaces are the right or best way to interact with these new models.
They make sense, though. Everyone can try it out because everyone can type into a text box. Plus, written words are probably the largest data set we have right now, so you’re going to get the best results there.
But people aren’t that good at writing and describing what they want to make.
GUIs exist because, at some point, the creators of early computers wanted people to be able to use them without knowing how to code or use the terminal.
That’s not exactly what this is, but it feels close. Right now, being a good ‘prompt engineer’ is a valuable skill, but should it be?
It’s early days, but I’m interested to see what is being built that isn’t chat boxes.
I have a couple of projects today where we are experimenting with video input as a way of giving instructions to an application, as well as voice. Imagine opening your browser, turning on your camera, and then how you behave changes what the app does.
Earlier in the summer, I was interested to see what Figma demoed at Config. Some of their interface solutions were really interesting and start to speak to what I wonder could be a fundamentally new thing…
Vibes
Today, everything and every setting in my design software is defined. Each color has a value—type sizes, fonts, widths, etc. Knowing how to set everything is literally the job. But they are all means to an end.
In the end, we (creatives) are in the business of delivering things that feel right. We can imagine a feeling, a vibe, and know how to deliver that.
That applies to the art direction for a campaign shoot, onboarding for an app, or the design of a new wine bar. It’s balancing a combination of vision and details.
But I have been thinking about interfaces that use AI to let me… adjust the feeling of something. Which felt weird to type out just now, but I think there is something there.
The tone-of-voice dial that Figma showed is close to what I’m thinking about. But you can imagine that being applied to lots of different variables.
In my head, I can picture some kind of dashboard with knobs and dials, and that making visual things might feel more like being an audio producer rather than a UI designer.
Or more like, maybe it’s both. Like this app that I’ve been playing around with this week—a combination of drawing and prompts. It’s crude, but it’s a tool and it works.
Interestingly, Procreate came out yesterday and said, “We’re never putting Gen AI in our products.” And lots of people clapped.
And truthfully, part of me was clapping along. I like opinionated products.
But another part of me wonders what the innovative artists of the past would think about that?
There are lots of questions here that don’t have good answers. But the more I think about it, the more I am excited about the kinds of interfaces, tools, and products we will make in the coming 5–10 years, and then what kind of art and creative projects we end up with.
