Published in: Ines Kleesattel (ed.), Ruedi Widmer (ed.):Digitalization and the Cultural Public Sphere. Diaphanes, Zürich, 2018.
The Taste Machine
Flushing Avenue is an artery pulsing through Brooklyn, lined with warehouses, bodegas, and thrift shops, along with exclusive coffee shops, boutiques, and hip bars. In the midst of this potpourri, King Noodle serves East Asian street food and fruity, exotic drinks in bulbous glasses topped with tiny umbrellas. The service here matches Brooklyn’s image: likable and nonprofessional.
Two men and a woman are sitting at one of the tables; the two men are the authors of this text. Probably like most of the restaurant’s guests, all three of them are not from New York, not even from America – and yet they appear to fit in seamlessly. Their dress code and manners conform to trends that are subtly influenced by the idea of “New York City”: the image, the concept of this city is part of their childhood, their youth, their adult life. Without ever having been here before, they carried this place within – thanks to pop culture and the World Wide Web. Now that they are here, they fit right into this picture of the new Brooklyn, the Brooklyn of the creative class, of Generation Y, the Millennials, the ones who grew up into digital technology.
The conversation at the table has turned to music, drawn by the sound flowing from speakers in the corners of the dimly lit space: a blend of pop and crossover hits of the 2000s. Timbaland, Missy Elliott, Outkast – influential figures in the musical socialization of the generation gathered at King Noodle, representatives of the universal pop music taste of their time. Before Spotify, Facebook, and YouTube, this taste was shaped by MTV and radio. Other significant influences included music magazines, older siblings, and friends. The tone was set by America, and above all by New York: large, cool, urban. Before the trend towards unconditional individualization hit the mainstream, there were pop stars whose songs and lyrics everyone in the room knew. And so the three table companions not only share lively memories of growing up with R’n’B albums, early music exchange platforms, and slow internet connections, but occasionally sing along to the likes of 50 Cent, Destiny’s Child and Sean Paul. Today, people between 25 and 35 have diverse preferences, but the first decade of the century remains their common ground, and the predictable playlist at King Noodle meets expectations.
In response, the group at the table asks: Is this what it all comes down to, that I may like this and you prefer that, and yet we neo-Brooklynites somehow listen to the same thing the world over? And all this in a time characterized by a proliferating, fragmented music landscape that should in principle contain as many taste patterns as there are people? What does it mean to claim: Pop is going digital? What is changing in pop music and in the ears of its listeners if Spotify is in fact more than merely a means of transmission?
Between the East Village and Greenwich Village lies New York University (NYU). Hosting 55,000 students, it occupies an entire block. The largest private university of the United States is a city within a city, with its own restaurants, shops, museums, and its own radio station. Since 1949, WNYU has broadcast programs featuring all kinds of music, from hip hop to jazz and electronic music, supplemented by sports news, talks with artists, and readings.
It’s a Tuesday evening in May 2016. Time for the show Beats in Space, which is recorded in a basement studio at NYU. It is hosted by Tim Sweeney, who started the show as a student and is today an internationally recognized DJ. The concept of the show is in part responsible for making his now-considerable reputation. For each broadcast, Sweeney invites an artist to musically curate the second half of the two-hour slot. This evening, the guest is Marcus Lambkin, an old colleague of Sweeney’s from his time at the American electropunk label DFA, better known by his artist pseudonym Shit Robot. Over the years, Beats in Space achieved cult status, with many faithful listeners and returning guests. (This is Lambkin’s third time on the show.) Sweeney is renowned as a “tastemaker”; his radio program can arguably be understood a vehicle for music promotion. Conversations with guests center on their new albums or singles, sprinkled with personal anecdotes and shared memories. The listener gets the impression of a casual chat, not a coordinated critical interview. While DJs usually don’t like to give away their tracks, the music played on Beats in Space is available as an online track list. While Tim Sweeney thus functions as a “gatekeeper,” exercising sole control over the show’s content and confronting his audience with a fait accompli, what drives him is enabling the discovery of new music – for 18 years now, in the heart of the digital storm, largely avoiding algorithms.
An hour before the show, interviewed in a restaurant near the studio, Sweeney recalls the time before Spotify and YouTube. He still looks a bit like a student, wearing horn-rimmed glasses and a dark cardigan. His carefully parted hair lends the 36-year-old the air of a grown-up nerd: “Finding music for my radio show used to be going to the record store and digging up random records or talking to someone that works there and who knew what I liked. These days, a lot is happening online and YouTube suggests what to listen to next. This is less satisfying, compared to talking to someone or to the thrill of finding a record in the store coincidentally.”
While he acknowledges that the algorithmic approaches of YouTube or Spotify have had a noticeable influence on his role as radio host and DJ, Sweeney does not feel threatened by them: “If they have the access to enough data, then I think they could get something approximate to that. But at the same time, I find there is a boringness to it that they have not overcome yet with algorithms. I think they are not yet able to surprisingly connect the dots to some other musical world.” As a DJ, he considers the medium of radio to be on his side, since he values human individuality and intimacy: “In my show, you have these little interviews with the guests, this personal thing. You have the photographs that give it a personal touch, too, and it draws people in in a different way. Plus, I am on a college radio, so there is not going to be any algorithm that is going to take over.” And what about the mass media? “In terms of the BBC, for example, the algorithms could be more threatening to the DJs, because there it matters how many people listen and an algorithm could maybe foresee better what a lot of people want to hear.” It has to be noted that radio has also been relying on algorithms for some time. Take the music-planning software MusicMaster, which searches available titles according to rules determined by program editors to select songs for a show. These rules can mean, for example, that no two songs of any one genre may directly follow each other or that no song may be played more than ten times a week. But Tim Sweeney still trusts in the appeal of the human curator: “Radio definitely has a future. Look at Apple Music, they still have the personalities like Tim Westwood curating a channel. People still want to hear what these people play and discover new music through these personalities.” If Sweeney looks calmly towards the future, it is probably because the situation of his show on college radio is quite comfortable. Beats in Space is not subject to any pressures of commercial success. He does not get paid for the show and makes his income touring as a DJ in North America, Asia and Europe.
Questions of Identity
The office of John Seabrook, staff writer at The New Yorker, is in an old building on Franklin Street, just a street away from the Hudson River. The smoking doorman doubles as the liftboy. The ride in the elevator with missing ceiling panels is hardly glamorous. On the fifth floor, Seabrook awaits at the door to his office, looking a bit bleary-eyed. He greets his guests politely before ushering them into his refuge, a space of about fifteen square meters, containing a writing desk, a scuffed sofa, and bookshelves stacked with works by David Foster Wallace, a guitar, and treatises on hip hop and other pop cultural themes. Some guitar plectra are scattered across the floor; the walls are decorated with pictures of his wife and children. In a gap between the books sits a snow globe, the snow falling not on the traditional Christmas tree but on the solitary word “Fuck.” John Seabrook has worked at The New Yorker since 1993 and has published numerous books on the side. His most recent, The Song Machine: Inside the Hit Factory from 2015, as well as a long reportage on Spotify, published in The New Yorker in 2014, are the reason for this meeting. Seabrook’s texts contain an intensive analysis of music production and reception in the context of the technological innovations and related economic shifts of the past twenty years. The question to what extent taste in pop music can be created or at least directed – irrespective of whether by humans or machines – is central to this discussion.
Casually leaning back in his chair, hands crossed behind his head, Seabrook wades straight into the debate on taste: “People’s taste in music is more complex and less logical than an algorithm could foresee. The algorithms are ultimately based on the basic Amazon principle of: people who bought this book also bought this book, which works well with books, because you do not meet so many people who like commercial fiction like romance novels and then Jane Eyre or Jonathan Franzen. People tend to have relatively narrow taste when it comes to reading and some of the other arts, too. But with pop music you have people that like Kelly Clarkson and also the Sex Pistols and they also like jazz. So the algorithms face a steep challenge, because they try to model something as a logical system that is not necessarily logical. You also have to weigh in all sorts of identity issues and other social factors.” In Seabrook’s view, questions of identity are the big hurdle that algorithms have yet to overcome. He talks about his teenage son, whose taste in music is heavily influenced by his friends: “If you, as a teenager, need an algorithm to tell you what to listen to, then there is something wrong with you. My son’s taste is really more shaped by his friends than by the algorithms. It is easier probably for algorithms to work on people whose taste is already more established, to extend their knowledge of a certain genre for example.”
Again, radio seems the obvious point of comparison. In contrast to the individualism enabled and reinforced by Spotify, radio often has a leveling influence, argues Seabrook. Specific personal tastes are only rarely catered to: the same is broadcast to the many. Spotify and its ilk operate differently: “I used Spotify to write my last book and in the course of doing that I listened to a lot of pop music, like Ariana Grande, Katy Perry, or Britney Spears. But when I am not doing that I listen to hip hop like Young Thug. So I have two very different patterns of uses and the algorithm is trying to figure me out. It hits the sweet spot once in a while, but it is also way off a lot of times. Commercial radio on the other hand tries to reach a large amount of people, so it can’t be very finely tuned to anyone’s specific taste and tries to stay more in the middle.”
The continuous refinement and optimization of the algorithms powering Spotify, Soundcloud, or YouTube can also be read as an attack on the encrusted structures of the music industry. Labels, which Seabrook describes as fundamentally technophobic, are adapting to the new possibilities: “Certainly A&R [artists and repertoire] has changed from being something where you would have someone come in to play, or you go see them in a club, to a practice where you do the research on YouTube. And YouTube does the pre-selecting that you used to do. It is like market research, because you can already see how many views a band has, for example. In a sense, it is automating a bit of that process. It is not literally being done by machines, but it is networked in a way that makes it a mechanized process. Social media is a curator of talent. That is a pretty significant change, I think. ”
The conversation touches on the story behind Netflix’s breakout series House of Cards: the evaluation of user data reflected viewers’ preferred actors and narratives, thus providing the basis for the show’s concept and guiding the central decisions regarding cast and crew. Although the creatives selected to manage the project were then given maximal creative freedom and budget powers, their selection in the first place was driven by user data, thus, essentially, by the audience. Will this logic of data-driven, creative production soon apply to music, through services such as Spotify? Seabrook, still relaxed and attentive despite his mobile buzzing relentlessly, weighs in: “I do think we are heading towards automatic song making. And I also think the concept of an author in songwriting is disappearing, as data becomes more and more a part of the whole process.” In this model, potential sales play a decisive role even at the creative stage of the process. While this was always to an extent true of the music industry, the availability of big data considerably widens the scope and the impact of such practices.
DeForrest Brown, Jr., sits in a hip café with modern decor called Freehold on the Williamsburg shore of the East River. Brown is in his mid-twenties and works as a music journalist for Mixmag, an established British magazine for electronic dance music. He talks with excitement about his views on the role of the music journalist, the New York scene, and how music works in the United States. The bagel in front of him remains almost untouched until the end of the conversation.
“I am always talking to my boss, who is one of those old British ravers. I have to reroute to him how people get music here. Growing up in the South, there were no clubs.” For Brown, club music, when it mattered at all, was a soundtrack – for cartoons, later for car rides. “The cartoon network Adult Swim was my entry point to electronic music. They are known for putting music like Autechre on as bumpers in-between shows, which led to me being awake in my room at 3 a.m. in the morning, when I was not supposed to be awake, to listen to this music I never really heard before. ”
It was only later that Brown learned of other ways to listen to electronic music, via another TV station: Boiler Room, a broadcaster that emerged deep in the London underground scene, wanted to be nothing less than MTV updated for the 2010s, the one place to discover new, exciting music. Like Tim Sweeney’s radio show, it is a voice that appeals to many (almost two million fans on Facebook). At Boiler Room, too, there are gatekeepers who decide what is hot and what is not. A camera is pointed at a DJ or a live act with the audience behind them. Via livestream, YouTube, and other channels, this simple switch in perspective can bring a rave into bedrooms all over the world – including that of DeForrest Brown, Jr., in Alabama in the early years of this decade. Brown talks about how, after he discovered Boiler Room, he used to invite friends to his dorm to become part of the virtual extension of the club and dance together in front of the computer screen. These moments came closest to the club experience that seem so natural to his boss Nick DeCosemo, a British rave veteran and chief editor of Mixmag.
Brown emphasizes that it’s not about a showdown between humans and machines, but about finding niches for the production and reception of music for various participants: “The icon that stands in for the music does not really have to be a person as much as it is a delivery platform, context, or network.” The question is not who or what presents the music, but how. A non-human variant for example is Hatsune Miku, an avatar that sings computer-generated and algorithmically composed songs and became a world-wide megastar in the process. Brown, who developed an interest in analyzing techno-social systems through his exposure to electronic music, imagines a future in which systems do not compete as much as supplement each other: machines profiting from humans and vice versa.
On leaving the café, he notices a few acquaintances who have settled with a fleet of laptops at a neighboring table. They turn out to be colleagues who work for Vice. Vice, which began as a lifestyle magazine before turning into a global news and video factory, has, so he tells us, bought up an entire block in Brooklyn, kicking out three established local clubs. Brown goes up to their table to tell them about an event at the end of the week at the office of his employer. Mixmag regularly hosts parties and records them – a bit like Boiler Room but with a much smaller reach.
“A Question of Time"
The colleagues from Vice never show up at the party in the backroom of the Mixmag office, on the Brooklyn side of the Williamsburg Bridge. Instead, there is Matt Liston, an Internet acquaintance of Brown’s. They met exchanging ideas in forums for experimental electronic music, not least because Liston lives on the West Coast and the scene is not bound to a single place. Liston is currently in New York on business, followed by a trip to Europe. With his NON T-shirt and shaggy beard, he doesn’t really fit into the party. Too maladjusted, too bulky. He showed up as a favor to his online friend DeForrest. Liston spends most of the evening outside in front of the door, vaping.
Two days later, at Bushwick Public House on Myrtle Avenue. The place touts a small selection of coffee, beer, and culture; socket strips on the floor between tables mark it as laptop friendly. Liston sits at one of the round tables by the window, with a view of the passing M train on the steel overpass. But his gaze is fixed on the laptop screen in front of him. He wears clunky black headphones with shiny blue ornaments, typing briskly until he notices his guests. Liston, 24, works for the internet start-up ConsenSys; before that, he co-founded one of his own, called Augur. His field is future markets, meaning the trade in information about future events. So, sports bets and trade in financial derivatives? “Future markets are not tied to the financial sector,” replies Liston. In contrast to the examples mentioned, enrichment and speculation play only a secondary role in his field. It’s about aggregating information and disrupting ossified systems, he explains, and by way of example mentions the acquisition of information in conflict zones. The data obtained from crisis zones is often influenced by those in power or even faked altogether. If there were first-hand information from decentralized sources, relying on the assessments of people on the ground, the picture would be much less distorted.
The idea behind this is the blockchain: decentralized structures, direct access, no middle men. Liston sees applications for this idea in the realm of music, for example in licensing: “Imagine having a piece of music indexed somewhere. It is always there and it is hosted on decentralized hosting and whenever someone wants to use it they send money to a smart contract which then gives them the rights to use it. That is like automated content licensing and that could disrupt a lot of the fat part of conventional label work.” In the beginning, Spotify’s distribution relied on a peer-to-peer system, where data is transmitted laterally from user to user. In 2014, however, they restructured their service to transmit streamed data centrally from Spotify servers. Licensing also doesn’t take place automatically between producer and user, as Liston envisions, but still involves various intermediaries, including Spotify itself.
As a final consequence, the entire system might be redesigned to better fit individual needs: “I see the potential for super-customized art or music. Imagine walking into a space and then there is some meta-data about your preferences or your life and it starts to create totally personal experiences based on that.” The computer would thus recognize what appeals to an individual consumer and re-program a track while the artist sits at home sipping coffee? “I think it is going to take a while, but it is going to happen, that algorithms are going to have organic insights into complex information.” The technology, according to Liston, is already here. So far, it has failed only because the various technologies and systems cannot be connected to work together smoothly. There are two reasons for this: “The main thing is scaling, which is going to be taken care of by Moore’s law at some point. The other thing is adoption, because these technologies and systems rely on having lots of connecting points. The more complex an algorithm, the more important it is that the different parts work together. Until this happens, the possible outcome is hard to tell.” When that will be is largely determined by the influence of existing institutions. It is in the interest of labels, for example, to prop up the existing system and keep it going for as long as possible.
In any case, music would be just one of many applications for this technology. Liston talks enthusiastically about “smart cities.” So-called “smart contracts” would play a key role and handle a range of processes on their own. In addition, meta-data would drastically increase efficiency – beyond the delivery of individualized music experiences – in areas such as personal transport, finance, and even the judicial system. In these cities, blockchains and cryptocurrencies would function as the link between human and artificial intelligence, Liston suggests. Listening to his claims, real and virtual worlds begin to blur. The vision is promising, even alluring; but it also seems distant and not always coherent. What role, for example, remains to us humans when artificial intelligences are superior to our own? What reason remains for our existence, once our labor and our money have become insignificant? Will we become idle beings whose desires and appetites are satisfied by machines?
Brave New World of Music
At the inconspicuous entrance to the Trans-Pecos, a music venue about a mile east of Bushwick Public House, a large, lonely candle is flickering next to a chalk board announcing the program: an evening of avant-electronic music curated by musician Robert Aiki Aubrey Lowe, entrance fee five dollars. On stage on the ground floor sits Marc Kate, a multidisciplinary artist from San Francisco. He manipulates the machines installed on the table in front of him, creating heavy ambient sounds that drift through the space for half an hour. His performance is followed by a DJ who plays slow beat tracks; the atmosphere is relaxed. Then a rupture: creaky synthesizer sounds, an energetic singer who hurls her microphone and herself through the space, and ear-splitting feedback dominate the scene for the next 45 minutes. The reactions to this performance by the duo Wetware range from delight to irritation.
The element of surprise that came up in many of the conversations this week was in full swing at the Trans-Pecos. We might flippantly say that Robert Aiki Aubrey Lowe put together a bit of a mixed bag, a rollercoaster of an evening. But there were still points of connection between the various performers; the fluidity and diversity of the evening’s aesthetic corresponds to how many critics, including John Seabrook, describe contemporary pop music. Lowe himself remains in the background, observing the performances from a corner. In his own work as a musician, he is interested in the potential of the synthesizer as an instrument in its own right, and he has explored this in original and compelling ways. For example, he developed a technical solution to connect the human body or even plants to a synthesizer.
Under the impression of Lowe’s musical output and his evening of curated music, it is hard to imagine that machines will ever truly replace human creativity. It seems more likely that human and machine will continue to become ever more entangled, losing their distinguishing features and merging into a hybrid organism in the process. Antiquated roles and responsibilities will rupture and shift. The concepts of music production, reception, distribution, promotion, and presentation will change fundamentally. Occasionally, even the present-day streaming world of the early 21st century appears like the realization of a past utopia – for example, the music rooms invented by Edward Bellamy in his 1888 science fiction novel Looking Backward, which provide playlists to subscribers around the clock, delivering music on command in perfect quality, unlimited quantity, to serve any mood. The microstress of having to decide what music might be the best for a particular moment disappears.
The consumer desires a voice that speaks only to her, not to the many. Streaming services are fulfilling this wish in ever-more-consistent and precise ways – and they are discovering better ways to know what we like. As it becomes dematerialized and omnipresent, the system overcomes physical challenges such as limited product quantities or local availability, bringing the pre-industrial past into the repository that has become the world. Bushwick Dinner Music, Manhattan Was a Blank City, or Algorithmic Computer Music: all these might be playlists for such special moments or situations, compiled by a machine, by a human being, or in productive symbiosis. Welcome to the brave new world of music.