Technological progress has always come from humanity’s grasp exceeding our reach; before we could build drones we needed to imagine flight. However, our successful flying machines bore little resemblance to their fictional counterparts. 

Design has not learned this lesson, and continues to try to wow users by recreating familiar aesthetics of futurity, rather than the outcomes depicted in this media. And whenever it does this (as with any aesthetics-first effort) the result is always a failure. The inevitable future disappears like so much smoke. 

I remember back in 2010 when the Xbox Kinect came out and influencers promptly announced that we would be entering the Minority Report future. This was immediately followed by a renewed interest in the gorilla arm effect, studied in old light-pen interfaces (but greatly amplified by gestural interfaces). This was hardly an issue with Kinect’s original game-related use case, but it would make any extended productivity user miserable. 

The Timeline

Transparent interfaces are another such usability disaster. We talked about Liquid Glass last week, but Chris Noessel called out the flaws of this design style way back in 2018. It looks cool on the screen, because its purpose is to look cool on the screen. When it comes to our devices, it sucks. And yet we already have copycats following Apple on the road towards a future that doesn’t give a damn about functionality, as long as it looks cool. 

Which brings us to AI. Well, let’s be honest — AI is just a marketing term for automation. Today AI means LLMs, and there is exactly one product that has achieved mass adoption: ChatGPT. The frenzy of investment into LLMs was kicked off with a chat-based demo for OpenAI’s latest transformer model. This form factor still benefits from the mirror effect to convince people that the LLM behind it is capable of far more than it actually is (until another product leverages that effect more, erm, effectively). 

Throughout most of our sci-fi media, characters interact with computers by talking to them. Star Trek. 2001: A Space Odyssey. The Jetsons. It’s a really cool way of telling a story. But as an interface, it sucks! It is the recourse of a lazy designer attempting to cash in on the general fuzzy feeling of The Future from our favorite TV shows. 

If you are going to be using LLMs in your product, forget the aesthetics. You need to be able to think of transformers as a normal technology. Not the prelude to general intelligence, but a printer or a database. And definitely don’t try to fake having an AI and mimic its aesthetics with human beings cast in the role of robots. 

Reading Material

While the future on our screens continued to delight, the future in our hands was already exhausting over a decade ago when Ian Bogost first reflected on this “yet another thing” phenomenon. 

The trouble with all this junk is that it lacks purpose. While people have found uses for tools marketed as AI, they are not designed to be purposeful. They show up in our lives through a corporate mandate, and we must expend energy on finding a place where they can fit. 

So it’s not that taking cues from fictional worlds is wrong. It’s just that taking solely aesthetics without understanding their purpose is bad design (regardless of where the aesthetic is from). There’s actually a great source of inspiration that I repeatedly turn to that is both very futuristic and highly usable — game design. 

Because serious game designers don’t talk about fun, just like serious UX designers don’t talk about delight. They talk about flow. They talk about user journeys. They talk about onboarding and building expertise and opportunity cost. And yes, all of this is done in the context of laser guns and space cowboys, but if you just try to mimic the floating blue hologram interfaces that the cowboys have, you’ll make, frankly, awful design. 

Good Questions

If you want to create experiences rather than spread aesthetics over a product like margarine over toast, take care to refer to Neil Postman’s seven questions for technological ideas (extracted here by Scott Berkun). They are all important, but as a problem designer, this one is my personal favorite: 

This goes double for when you are solving technology problems like “can we make this a chatbot?” or design problems like “what would impress other designers” — but creating user problems like “I need a guide to turn all of these "improvements" off.” 

Slow down, and fix things. 

—Pavel at the Product Picnic