Last Thursday (May 18), music industry executives and potential investors flocked to Los Angeles's El Rey Theatre for a glimpse of the potential future of the music-tech industry. Inside was a demo day concluding the first-ever Techstars Music startup accelerator program.
Headed by Bob Moczydlowsky, the former head of Twitter Music, the program invited 11 startups to work on their products from out of Techstars' LA office over the course of three months. Each company received a $120,000 investment, access to a hundreds-large pool of mentors (including members of program partners Sony Music Entertainment, Warner Music Group, Sonos, and Harmonix), and lifetime support in exchange for six percent equity. The incubation period was, as Moczydlowsky told Billboard, "designed to be a year's worth of work in three months."
Those three months led up to the event at the El Rey Theatre, where each startup had roughly ten minutes to pitch themselves in hopes that someone would believe enough in their vision to provide financial support. During the demo day, the companies offered a variety of ideas: one, Hurdl, created a wearable, text-activated LED device that collects and analyzes data about a live event's attendees; and another, Shimmur, aims to improve how celebrities and influencers interact with their fans through a content-upvote system. The most prominent theme of the day, however, was how companies are innovating the ways in which people make, experience, and monetize music using artificial intelligence.
Artificial intelligence (AI) is the science of making machines complete tasks and solve problems that would require intelligence if performed by humans. The concept of AI has been around since about 1950, but it has become an increasing reality in the new millennium: with this technology, people can ask Siri to pull up the weather forecast on their iPhones, play competition-based video games without a human opponent, and drive self-parking cars. AI robots even write news articles.
Artificial intelligence has also been changing the music industry, from how songs are made to what tracks consumers listen to. Last year, British songwriter and producer Alex Da Kid used IBM's Watson Beat—a cognitive technology that "understands music and lets artists change the sound of a song based on the mood they want to express"—to co-produce his single "Not Easy," which debuted at No. 6 on the Billboard Rock Digital Song Sales charts. Also last year, researchers at Sony shared the first pop song composed by AI, the Beatles-inspired "Daddy's Car." In the streaming world, platforms like Spotify are utilizing AI to recommend to users music they might like based on their listening habits.
Two of the startups in Techstars Music, Amper Music and PopGun, use AI to create custom music content, though each for different contexts. Amper co-founder and CEO Drew Silverstein, a film composer, envisioned the product as a solution for creating affordable, royalty-free audio for projects such as news reports, wedding videos, and sports highlight reels, whereas an alternative option would be using pre-written stock music. All a user has to do is determine which kind of style, mood, and length they want, and let technology do the rest—no prior music expertise needed.
Brisbane-based company PopGun, on the other hand, is using deep learning to build a "superhuman musician" that can write and produce original pop songs in collaboration with human artists. During his presentation, co-founder Jack Nolan introduced the audience to Alice, an AI that learned to listen to and understand music purely based on its sound (the raw frequencies that make up the song). He then showed a video in which a pianist "played" with Alice, who was able to anticipate the pianist's next notes following a few riffs. With more practice, PopGun anticipate Alice will have the ability to jam with a human artist as if she were a studio musician.
Stockholm, Sweden-based startup Pacemaker's use of artificial intelligence means that DJs no longer have to choose between staying behind the decks and venturing out on the dancefloor. The DJ app (the first of its kind with a Spotify connection) allows users to mix tracks from an iPhone, iPad, or Apple Watch using the streaming platform's vast library. Additionally, Pacemaker's Autopilot feature allows anyone to be a DJ without knowing how to beat-match: it analyzes a user's playlist, places tracks in the best playable order, and creates a seamless mix. Users can also use the AI technology to create and share professional-sounding DJ mixes using Pacemaker's own Metamix format, which only saves information on how the mixtape was created, and not any song files. Since all tracks come from Spotify, artists generate track streams, thus making DJ mixes an actual source of revenue.
Another startup, Syncspot, focuses on how to use AI to monetize music and music-adjacent ventures. The marketplace for cross promotions (when two brands team up to promote each other's products, as with the McDonald's Happy Meal and the Super Mario toy it comes with) was created by the Jalil brothers, one of whom (Jubair) once bankrupted his agency when executing a cross-promotions deal with a meager budget, he recalled during his pitch. With Syncspot, Jubair explained, AI technology does all the work that he did—brand matchmaking, planning campaign strategy, and reporting—without the time and cost of human labor. Examples of brands they've worked with include Heineken, Unilever, and Coachella.
Many of these startups have good intentions: to save users valuable time and money. But they also re-open a long-discussed conversation regarding artificial intelligence: what's the likelihood that this technology will stop being used to work with people, and instead be used to replace people?
Speaking to THUMP after the presentation, pop-music AI startup PopGun's Nolan recalled that these concerns surfaced early on during conversations with industry people, causing his team to second-guess themselves. "They were like, 'You don't want an AI to write the music, that's scary; people won't like it; you'll freak people out,'" he said. Despite these technological advancements, he believes that humans will always fundamental to the production process.
"[Our AI] can create music on its own, but we just don't think's as exciting," he said. "Sure, it's useful for people who just need a generic tune, but for us, the more exciting thing is that creative partnership and back-and-forth you get with people. It's about giving artists the freedom to experiment with it. We don't think the future is going to be either AIs or artists; it's going to be AIs and artists."
Krystal Rodriguez is on Twitter.