When the Apple Watch App Store opened, a prominent "Get Started" section featured a line-up of usual suspects. The first app was Twitter. The next was The New York Times. No surprises there. Then came... a DJ app? It was an odd inclusion. How do you DJ on a watch?
You don't. You let AI do it for you.
Pacemaker, the unlikely recipient of Apple's coveted recommendation, arrived last year as a groundbreaking DJ app for the iPad. Its thoughtful interface and exclusive pipeline into Spotify's catalog made mixing music terrifically easy. But according to creator Jonas Norberg, not easy enough. "It's still too hard for the majority of people to create something that sounds good," he says.
The Apple Watch offered Norberg and his eight-person team a chance to do something even simpler. Indeed, the Stockholm outfit's new app scales DJing back to something so basic it you can't even really call it DJing.
The app relies almost entirely on a pair of algorithmic features. The first, MatchMachine, analyzes your playlists, evaluating factors like beats per minute and key to shuffle songs into the optimal sequence for mixing. Then, a patent-pending artificial intelligence engine orchestrates transitions between songs. Internally, the team refers to the AI DJ as Mållgan, after a character in a popular Swedish children's book. Officially, it's called "Autopilot."
From the watch, you can trigger Autopilot transitions and sprinkle in a handful of effects, including reverb, a whooshing wave of white noise, and a glitchy 8-bit-style filter (The app is free; the effects cost a few bucks each). The AI assists here, too. When you tap an effect from your watch, it isn't deployed instantly. Instead, the app scrutinizes the waveform of your track and mixes the effect in smoothly. "It's not good enough just to start the effect when you touch it," Norberg says. "You need curves---you need someone turning on the knob. Autopilot controls the knob."
As many have noted, Apple's new device forces developers to rethink assumptions about what apps should do. "The watch is really interesting because of the constraints. To do something with those limitations is extremely hard," Norberg says. Algoriddim, maker of another popular mixing app called Djay, took what might be considered a more obvious approach to crafting its Watch app, shrinking the standard two-deck layout and providing features like crossfading and looping. It does you might expect a miniature DJing app to do.
Pacemaker's creators consciously opted to give users less control. As they see it, the ideal experience for the wrist isn't being a mix master, but enjoying a straightforward, slightly more active form of music listening. "There's many people out there that want to do more than passively consume music," Norberg says. "Mixes are a great because you can mess around with them and make music more personal." Autopilot's meant to ensure it sounds good.
Norberg cites Paper, the popular iPad sketching app, as inspiration. One of its great innovations was the "Expressive Ink Engine," which interprets scrawls and smooths them into beautiful lines on screen. "Autopilot is our Ink Engine," Norberg says.
These examples point to an interesting trend. On mobile devices, where input is limited, algorithms and AI can serve as powerful mediators, smoothing rough edges both literally and figuratively. But as our devices shrink and input becomes ever tougher, this mediation becomes increasingly important. On the iPad's large screen, you can let users do, say, 90 percent of the driving and let algorithms fill in the rest. As the new Pacemaker app suggests, the opposite ratio might be best.
It's an approach we'll almost certainly continue to see developers experiment with. The smaller our devices become, the more chances there are for apps that act all on their own.
