Disclaimer

Index

Content Types

Artist Profiles (2)

Audio (4)

Audio Papers (4)

Editorial (1)

Essays (5)

Interviews (6)

Scores (5)

Series (1)

Text Poems (5)

Contributors (57)

  • Allanah Stewart
  • Allanah Stewart is an artist from Aotearoa/New Zealand, currently living in Melbourne, Australia. As well as her work in various experimental music projects, she is the presenter of a monthly podcast radio programme called Enquiring Minds, hosted by Noods radio, which explores old and new, lesser known and well known sounds that loosely fit under the banner of experimental music.

  • Jen Callaway
  • Jen Callaway is a Melbourne musician, sound and performance artist, photographer, and community services worker raised in various parts of Tasmania. Current projects include bands Is There a Hotline?, Propolis, Snacks and Hi God People; and upcoming film Here at the End, by Campbell Walker, as actor/co-writer.

  • Isha Ram Das
  • Isha Ram Das is a composer and sound artist primarily concerned with ecologies of environment and culture. He works with experimental sound techniques to produce performances, installations and recordings. He was the 2019 recipient of the Lionel Gell Award for Composition, and has scored feature-length films and nationally-touring theatre installations. He has performed at institutions such as the Sydney Opera House; Black Dot Gallery, Melbourne; Institute of Modern Art, Brisbane; Metro Arts, Brisbane; Museum of Contemporary Art, Sydney; and Boxcopy, Brisbane.

  • Dylan Robinson
  • Dylan Robinson is a xwélméxw (Stó:lō) writer, artist, scholar and curator, He is Canada Research Chair in Indigenous Arts, and associate professor at Queen’s University in Kingston, Ontario. He is author of Hungry Listening, Resonant Theory for Indigenous Sound Studies, published by University of Minnesota Press.

  • Megan Cope
  • Megan Cope is a Quandamooka woman (North Stradbroke Island) in South East Queensland. Her site-specific sculptural installations, video work and paintings investigate issues relating to identity, the environment and mapping practices. Cope’s work often resists prescribed notions of Aboriginality and becomes psychogeographies across various material outcomes that challenge the grand narrative of ‘Australia’ as well as our sense of time and ownership in a settler colonial state.

  • Sean Baxter
  • Australian musician Sean Baxter died on 15 March 2020. Part of Melbourne's improv scene, he is described by musician Anthony Pateras as possessing “a unique aesthetic vision and intellectual depth, mixing highbrow philosophical concepts with punk sensibilities in how he lived, spoke and played. He was pure energy.”

    Drumkit and percussionist, Sean was an Australian improviser who forged an international reputation as a bold explorer of percussive possibilities both as a soloist and through his work with the acclaimed avant-garde trio, Pateras/Baxter/Brown. Focusing on the use of extended techniques applied to the conventional drum kit, he utilised an arsenal of metallic junk and other percussive detritus to expand the sonic palette of the percussion tradition. In addition to Pateras/Baxter/Brown, he was involved in many collaborations and was drummer for groups The Throwaways, Bucketrider, Lazy, SxSxSx and Terminal Hz.

  • Thomas Ragnar
  • Thomas Ragnar is an artist based in Singapore. His work is often underpinned by collaborations, affinities and research with experiential methodologies.

  • Alessandro Bosetti
  • Alessandro Bosetti is an Italian composer, performer and sound artist, currently based in Marseille. His work delves into the musicality of spoken language, utilising misunderstandings, translations and interviews as compositional tools. His works for voice and electronics blur the line between electro-acoustic composition, aural writing and performance.

  • Lin Chi-Wei
  • Lin Chi-Wei is a legend of Taiwanese sonic art, whose practice incorporates folklore culture, noise, ritual, and audience participation.

  • Mat Dryhurst
  • Mat Dryhurst is an artist who releases music and artworks solo and in conjunction with Holly Herndon and the record label PAN. Dryhurst developed the decentralised publishing framework Saga, which enables creators to claim ownership of each space in which their work appears online, and a number of audio plays that derive their narrative from the personal information of listeners. He lectures on issues of music, technology, and ideology at NYU, and advises the blockchain-based platform co-operative Resonate.is.

  • Sean Dockray
  • Sean Dockray is an artist, writer, and programmer living in Melbourne whose work explores the politics of technology, with a particular emphasis on artificial intelligences and the algorithmic web. He is also the founding director of the Los Angeles non-profit Telic Arts Exchange, and initiator of knowledge-sharing platforms, The Public School and Aaaaarg.

  • Emile Frankel
  • Author of Hearing the Cloud (Zero Books), Emile Frankel is a writer and composer researching the changing conditions of online listening. In his spare time he runs the Sci-Fi and critical fantasy publisher Formling.

  • Bridget Chappell
  • Bridget Chappell is a raver and theory bro currently living on the unceded nations of the Latji Latji and Nyeri Nyeri people. They make music as Hextape and organise parties in drains, observatories, and other natural amphitheatres. They founded and run Sound School, work with young musicians behind bars, and make experimental sound technologies to challenge police sirens.

  • Holly Herndon
  • Holly Herndon experiments at the outer reaches of dance music and pop. Born in Tennessee, Herndon spent her formative years in Berlin’s techno scene and repatriated to San Francisco, where she completed her PhD at Stanford’s Center for Computer Research in Music and Acoustics. Her albums include Platform (2015) and Proto (2019).

  • Candice Hopkins
  • Candice Hopkins is a curator, writer and researcher interested in history, art and indigeneity, and their intersections. Originally from Whitehorse, Yukon Territory, Hopkins is a citizen of Carcross/Tagish First Nation. She was senior curator for the 2019 Toronto Biennial of Art, and worked on the curatorial teams for the Canadian Pavilion at the 58th Venice Biennale, and documenta 14.

  • Raven Chacon
  • Raven Chacon is a composer, performer and artist from Fort Defiance, Navajo Nation. His work ranges from chamber music to experimental noise, to large scale installations, produced solo and with the Indigenous art collective Postcommodity. At California Institute of the Arts, Chacon studied with James Tenney, Morton Subotnick, Michael Pisaro and Wadada Leo Smith developing a compositional language steeped in both the modernist avant-garde and Indigenous cosmologies and subjectivities. He has written for ensembles, musicians and non-musicians, and for social and educational situations, and toured the world as a noise artist.

  • Lisa Lerkenfeldt
  • Lisa Lerkenfeldt is a multi-disciplinary artist working in sound, gesture and performance. Central to her practice is languages of improvisation and intimacy with technology. Traces of a personal discipline and form of graphic notation are introduced in the online exhibition 14 Gestures. The associated recorded work Collagen (Shelter Press, 2020) disrupts the role of the common hair comb through gesture and sound.

  • Haroon Mirza
  • Haroon Mirza is an artist who intertwines his practice with the role of composer. Mirza considers electricity his main medium and creates atmospheric environments through the linking together of light, sound, music, videos and elements of architecture. Regularly showing internationally in group and solo exhibitions, Mirza’s work has also been included in the 7th Shenzhen Sculpture Biennale, China (2012) and the 54th Venice Biennale, Italy (2011), where he was awarded the Silver Lion.

  • 33EMYBW
  • Shanghai native 33EMYBW (Wu Shanmin) has been an active member in the Chinese music scene for over a decade. She has also performed at CTM and Sinotronics in Germany, China Drifting Festival in Switzerland, and SXSW. Her 2018 album Golem, released on SVBKVLT, was met with critical acclaim and voted one of the best electronic albums of 2018 by Bandcamp. In 2019 she released DONG2 EP under Merrie Records Beijing, and will premiere her sophomore album Arthropods (SVBKVLT) at Unsound 2019.

  • Alexander Garsden
  • Alexander Garsden is a Melbourne-based composer, guitarist and electroacoustic musician, working across multiple exploratory musical disciplines. Recent work includes commissions from the Sydney Symphony Orchestra, the Melbourne Symphony Orchestra, Speak Percussion, Michael Kieran Harvey and Eugene Ughetti; alongside performances with artists including Tetuzi Akiyama (Japan), Oren Ambarchi, Radu Malfatti (Austria), Julia Reidy, David Stackenäs (Sweden), and with Erkki Veltheim and Rohan Drape. From 2014 to 2019 Garsden was Co-Director of the INLAND Concert Series. He has taught through RMIT University and the University of Melbourne.

  • Annika Kristensen
  • Annika Kristensen is Senior Curator at the Australian Centre for Contemporary Art, Melbourne.

  • Arben Dzika
  • Arben Dzika is an artist whose practice involves working with various media including, but not limited to: sound, image, word, and performance. His work primarily seeks to reflect on, interrogate, and play with technologies, systems, and human senses. Within his practice, he works as a producer and DJ under the moniker, Dilae.

  • Audrey Schmidt
  • Audrey Schmidt is a writer and editor based in Melbourne, Australia. She is a regular contributor to Memo Review, co-editor the third issue of Dissect Journal, and has written for various publications including Art Monthly, Art + Australia and un Magazine. She co-founded Minority Report with Adam Hammad in 2018 and released one online issue that was available until the domain expired in 2019. Audrey sits on the FYTA (GR) Board of Advisors.

  • Autumn Royal
  • Autumn Royal is a poet, researcher, and teacher based in Narrm/Melbourne. Autumn’s current research examines elegiac expression in contemporary poetry. Autumn is the interviews editor for Cordite Poetry Review, and author of the poetry collections She Woke & Rose (Cordite Books, 2016) and Liquidation (Incendium Radical Library, 2019). Her third collection of poetry is forthcoming with Giramondo Publishing in 2021.

  • Bianca Winataputri
  • Bianca Winataputri is a Melbourne-based independent curator and writer researching contemporary practice in Southeast Asia, and relationships between individuals and collectives in relation to history, globalisation, identity and community building. Currently working at Regional Arts Victoria, Bianca was previously Assistant Curator of Contemporary Art at the NGA. She holds a BA (University of Melbourne), and BA Honours from the ANU where she received the Janet Wilkie Prize for Art. In 2018 Bianca was selected for 4A Centre for Contemporary Asian Art’s Curators’ Intensive.

  • Brian Hioe
  • Brian Hioe was one of the founding editors of New Bloom, an online magazine covering activism and youth politics in Taiwan and the Asia Pacific, founded in 2014 in the wake of the Sunflower Movement. Hioe is a freelance writer on social movements and politics, as well as an occasional translator.

  • Chi Tran
  • Chi Tran is a writer, editor, and an artist who makes poems that may be text, video, object, sound, or drawing. Chi is primarily interested in working with language as a means of coming-to-terms. Their work has been published by Incendium Radical Library Press, Cordite Poetry Review, Australian Poetry and Liminal Magazine and exhibited at galleries including Firstdraft, Sydney; Punk Café, Melbourne; and ACCA, Melbourne. In 2019, as a recipient of The Ian Potter Cultural Trust Fund, Chi spent three months in New York developing their practice with renowned poets including Mei-mei Berssenbrugge, Fred Moten, and Jackie Wang.

  • Chun Yin Rainbow Chan
  • Chun Yin Rainbow Chan is a Hong Kong–Australian artist, living in Sydney. Working across music, performance and installation, Rainbow is interested in the copy and how the ways in which it can disrupt Western notions of ownership. Central to Rainbow's work is the circulation of knock-off objects, sounds and images in global media. Her work positions the counterfeit as a complex sign that shapes new myths, values and contemporary commodity production.

  • Dale Gorfinkel
  • Dale Gorfinkel is a musician-artist whose stylefree improvisational approach informs his performances, instrument-building, and kinetic sound installations. Aiming to reflect an awareness of the dynamic nature of culture and the value of listening as a mode of knowing people and places, Dale is interested in bringing creative communities together and shifting perceived boundaries. Current projects include Prophets, Sounds Like Movement, and Music Yared as well as facilitating Art Day South, an inclusive arts studio with Arts Access Victoria.

  • Danni Zuvela
  • Danni Zuvela is a curator and writer based in Melbourne and the Gold Coast. Her research is informed by interests in feminism, activism, ecology, language and performance. With Joel Stern, Danni has led Liquid Architecture as Artistic Director, and continues to develop curatorial projects for the organisation.

  • Eric Avery
  • Eric Avery is a Ngiyampaa, Yuin, Bandjalang and Gumbangirr artist. As part of his practice Eric plays the violin, dances and composes music. Working with his family’s custodial songs he seeks to revive and continue on an age old legacy – continuing the tradition of singing in his tribe – utilising his talents to combine and create an experience of his peoples culture.

  • Fjorn Butler
  • Fjorn Butler is an artist, researcher, and event organiser. As an artist, she works primarily in sound and performance under the name Papaphilia. As a researcher, she interrogates how biological discourses are used in neoliberal/colonial governance structures to shape the political. Fjorn's research informs her writing on sound-poetics and the challenges this framework poses to anglophone notions of property. She is also co-director of Future Tense and co-curator of Writing and Concepts.

  • Freya Schack-Arnott
  • Freya Schack-Arnott is an Australian/Danish cellist who enjoys a multi-faceted career as a soloist and ensemble performer of classical and contemporary repertoire, curator and improviser within experimental music, electronics, popular and cross-disciplinary art forms. Schack-Arnott regularly performs with Australia's leading new music ensembles, including ELISION Ensemble (as core member) and Ensemble Offspring. Her curatorial roles include co-curator/founder of the regular 'Opus Now' music series and previous curator of the NOW Now festival and Rosenberg Museum.

  • Gooooose
  • Gooooose (Han Han) is an electronic music producer, visual artist and software developer based in Shanghai, China. His current releases include They (D Force, 2017), Dong 1 (D Force, 2018), Pro Rata (ANTE-RASA, 2019). Gooooose's 2019 SVBKVLT–released RUSTED SILICON received positive reviews from media including boomkat, Resident Advisor, Dusted Magazine, and The Wire. Gooooose has performed live at CTM (Berlin, 2018), Nyege Nyege (Kampala, 2019), Soft Centre (Sydney, 2019), Unsound (Kraków, 2019) and Recombinant (San Francisco, 2019).

  • Harmony Holiday
  • Harmony Holiday is a writer, dancer, archivist, director, and the author of four collections of poetry, Negro League Baseball, Go Find Your Father/A Famous Blues, Hollywood Forever, and A Jazz Funeral for Uncle Tom. She lives in New York and Los Angeles.

  • James Rushford
  • James Rushford is an Australian composer-performer who holds a doctorate from the California Institute of the Arts, and was a 2018 fellow at Academy Schloss Solitude in Stuttgart. His work is drawn from a familiarity with specific concrète, improvised, avant-garde and collagist languages. Currently, his work deals with the aesthetic concept of musical shadow. James has been commissioned as a composer by ensembles including the BBC Scottish Symphony Orchestra (Glasgow), and Melbourne Symphony Orchestra, and regularly performs in Australia and internationally.

  • Jessica Aszodi
  • Jessica Aszodi is an Australian-born, London-based vocalist who has premiered many new pieces, performed work that has lain dormant for centuries, and sung roles ranging from standard operatic repertoire to artistic collaborations. She has been a soloist with ensembles including ICE; the Melbourne, Sydney and Adelaide Symphony Orchestras; and San Diego and Chicago Symphony Orchestras’ chamber series. Aszodi can be heard on numerous recordings and has sung in festivals around the world. She holds a Doctorate of Musical Arts from the Queensland Conservatorium, an MFA from the University of California, and is co-director of the Resonant Bodies Festival (Australia), and artistic associate of BIFEM.

  • KT Spit
  • Kt Spit (Katie Collins) is an artist and musician based in Narrm (Melbourne). Lyrically and visually her work explores subcultural narratives and challenges dominant representations of loss, grief, and true love. In 2015 Kt independently released her debut album, Combluotion, and in 2019 will release a visual album entitled Kill the King.

  • Immy Chuah and The Convoy
  • The Convoy conjure illustrious soundscapes from the abyss of chaos, revealing hidden worlds of the imagination as the performance takes form and infuses with subjective experience. Using instruments of sound, light and smell, The Convoy enchant space with themes of tension, evolution, entropy and regeneration. Sensorial immersion transports audiences through highly dynamic environments that shift and blend into one single, breathing moment. As entity, rather than singular, Immy Chuah is a guest within The Convoy on unceded land.

  • Sam Peterson
  • Sarah McCauley
  • Sarah McCauley is a Melbourne-based music producer, editor and writer.

  • Neil Morris
  • Neil Morris is a Yorta Yorta, Dja Dja Wurrung man. He is well known in Narrm/Birraranga for his musical project DRMNGNOW, a project built on subject matter tackling the colonial nature of the Australian construct and how that affects contemporary society upon this land. The work is unapologetic, clear, and deeply poetic. It hints toward Morris's extensive experience as a spoken word artist in Narrm since 2015. Morris's work is triumphant in the face of severe adversity often imbued in a quite fortified melancholy, a powerful marker of the survival of First Nations peoples in the now.

  • Natasha Tontey
  • Natasha Tontey is an artist and graphic designer based in Yogyakarta. She is interested in exploring the concept of fiction as a method of speculative thinking. Through her artistic practice she investigates the idea of how fear, horror, and terror could be manifested in order to control the public and how fictional accounts of the history and myth surrounding ‘manufactured fear’ might operate as a method of speculative fiction that determines expectations for the future.

  • Mat Spisbah
  • Mat Spisbah is a New Media curator with a unique portfolio of programming that seeks to integrate non-traditional artistic methods and emerging technologies. Having lived in Hong Kong for 14 years, he is connected to the region’s art and culture, and has created professional networks with artists, curators, galleries, promoters and industry professionals across Australasia. Portfolio highlights include the debut Australian performances of north Asian artists including: Howie Lee, Rui Ho, Meuko Meuko, Pan Daijing, Alex Zhang Hungtai, Tzusing, and Gabber Modus Operandi.

  • Mandy Nicholson
  • Mandy Nicholson is a Wurundjeri-willam (Wurundjeri-baluk patriline) artist and Traditional Custodian of Melbourne and surrounds. Mandy also has connections to the Dja Dja wurrung and Ngurai illam wurrung language groups of the Central/Eastern Kulin Nation. Mandy gained a Bachelor of Arts (Honours) in Aboriginal Archaeology in 2011, worked for the Victorian Aboriginal Corporation for Languages for six years and is now a PhD candidate studying how Aboriginal people connect to Country, Off Country.

  • Lucreccia Quintanilla
  • Lucreccia Quintanilla is an artist, writer, DJ and PhD candidate researcher at Monash University. Her writing and art have been published and exhibited both within Australia and internationally. Quintanilla’s practice is a collaborative one that manifests into outcomes within galleries and also as events and performances outside of that context. She regularly speaks at panels and symposiums on themes within her research, has received grants for her projects and residencies, and has taught at university level.

  • Amanda Stewart
  • Amanda Stewart is a poet, author, and vocal artist. She has created a diverse range of publications, performances, film and radio productions in Australia, Europe, Japan, and the USA, working in literature, new music, broadcasting, theatre, dance, and new media environments. Amanda collaborated with Chris Mann for many years in the Australian ensemble, Machine For Making Sense (with Jim Denley, Rik Rue, and Stevie Wishart), as well as in other contexts. Her poem ‘ta’ was written in honour of Chris Mann’s extraordinary vision and work.

  • Holly Childs
  • Holly Childs is an artist and writer. Her research involves filtering stories of computation through frames of ecology, earth, memory, poetry, and light. She is the author of two books: No Limit (Hologram, Melbourne) and Danklands (Arcadia Missa, London), and she collaborates with Gediminas Žygus on ‘Hydrangea’. She is currently writing her third book, What Causes Flowers Not to Bloom?.

  • Ivy Alvarez
  • Ivy Alvarez’s poetry collections include The Everyday English Dictionary, Disturbance, and Mortal. Her latest is Diaspora: Volume L (Paloma Press, 2019). A Fellow of MacDowell Colony (US), and Hawthornden (UK), her work is widely published and anthologised (twice in Best Australian Poems), with poems translated into Russian, Spanish, Japanese and Korean. Born in the Philippines and raised in Australia, she lived in Wales for almost a decade, before arriving in New Zealand in 2014.

  • Nick Ashwood
  • Nick Ashwood is a guitarist, composer, improviser and performer from Nipaluna/Tasmania now residing in Sydney. His focuses have been exploring deep listening, harmonic space and the possibilities of the steel-string acoustic guitar by means of preparations, just intonation, objects and bowing.

  • Johnny Chang
  • Berlin-based composer-performer Johnny Chang engages in extended explorations surrounding the relationships of sound/listening and the in-between areas of improvisation, composition and performance. Johnny is part of the Wandelweiser composers collective and currently collaborates with: Catherine Lamb (Viola Torros project), Mike Majkowski (illogical harmonies), Phill Niblock, Samuel Dunscombe, Derek Shirley and others.

  • Megan Alice Clune
  • Megan Alice Clune shifts between musician, composer and artist. Primarily, her work explores both the concept and aesthetics of ambient music through sound installation, collaboration and performance. Megan is the founding member of the Alaska Orchestra, and has presented work and undertaken residencies across Australia, Asia, Europe and North America, including the Bang on a Can Summer Music Festival (MA), Next Wave Festival, Underbelly Arts Festival, Performa 15 (NYC) and VividLIVE at the Sydney Opera House.

  • Andrew Fedorovitch
  • Andrew Fedorovitch is compos mentis.
 Andrew Fedorovitch embodies professionalism in every aspect of his life, including music.

  • Shota
  • Shota is an artist working in Australia. He makes sound-based works for varying contexts. He has had the opportunity to collaborate with a multitude of artists from varying disciplines. Shota is currently an honours student who is associated with the Plant ecophysiology and Ecosystem processes lab at the University of Sydney.

  • Sonya Holowell
  • Sonya Holowell is a Dharawal woman, vocalist, composer and writer working across new and experimental genres. The contexts for her work, and the forms they take, are diverse and deeply questioning. Her practice comprises interdisciplinary collaboration, improvisation, multi-form writing and conceptual composition. She is also a workshop facilitator; a curator of the Now Now Festival; lecturer in experimental vocal practice; and a co-founder/editor of online arts publication ADSR Zine.

  • Alexandra Spence
  • Alexandra Spence is an artist and musician living on Gadigal country in Sydney, Australia. She makes installations, compositions and performances based on (everyday) sound and listening. Through her practice she attempts to reimagine the intricate relationships between the listener, the object, and the surrounding environment as a kind of communion or conversation. She has a current, near-spiritual, obsession with the animation of material and object through sound. Alex has performed and presented work on radio, in concerts, festivals, symposiums and galleries worldwide, and has two releases: Waking, She Heard The Fluttering, with Room40, and Immaterial, with Longform Editions.

  • MP Hopkins
  • MP Hopkins is an artist based in Sydney, Australia, that is concerned with how to record voices that are not really there, and ways to make voices that are there not sound like voices. He makes audio recordings, performances, and texts.

Artist Title | 00:00 / 00:00 (Play)

Disclaimer

Index


Inhuman Intelligence:
Holly Herndon and Mat Dryhurst
in conversation with Sean Dockray

(play) (pause)
00:00 / 00:00

In framing this discussion between Holly Herndon, Mat Dryhurst, and Sean Dockray, I’d like to draw attention to what could be described as an evolution in the methodology and project of ‘shaping the future’. Holly and Mat’s album Platform, from 2015, made a call to imagine a better future. At this time ‘better’ was negatively bracketed against 90s cyberpunk dystopian imagery, reactionary club music and stale future visions which dominated contemporary art and experimental music. Nuanced, creative and kind, Holly and Mat’s Platform instead offered an aesthetic of a non-nostalgic futurism. Their message described togetherness rather than fragmentation.

Platform was made in conversation with the complex geopolitics of the time. Collaborators Metahaven defined this moment as one marked by the ‘black transparency’ of mass surveillance and ‘sousveillance’, and as the operation of a ‘reality stranger than fiction’. In 2015, the desire and conscious decision to try and ‘shape’ the future, via aesthetic and imaginative means, was influenced by the successes of chaotic online movements. Underpinned by philosophies of ‘causality’ (see hyperstition and accelerationism) — of one word having an affect on another word, of one sound having an affect on another sound, like a snowball gathering size — music seemed pedestaled, in this world, to say something important through its unique and interpretative language of play, of symbolism translated from screen to speaker, of conceptual feeling distilled like a form of alchemy from the hand into the song. Via the same method of kitsch meme-magic, and via the same method sold in The Secret’s sigilism of a child’s red bike drawing sticky-taped above a bed — of politics functioning like a law of attraction — the practical agency of music was understood here to work as a form of ‘magic’.

Of course it is magic to listen to a song you love alone in your bedroom, or to dance amongst others in a club, and the political potential of both of these listening conditions remains just as powerful today. Still, in charting the evolution of Holly and Mat’s method and message, from a distant past when the internet was ‘too complicated for governments and corporations to understand’, to today’s world where Anonymous hackers fail to make so much as dent in corporate run police states, the practical necessity of having a hand to play in the creation of new technology is absolutely essential. Holly and Mat here sidestep futurity and aesthetic causality by focusing on two forms of technological intervention. The first technology is old-school and doesn’t require any machines. The technology is called collectivity, and you can use it, as Holly and Mat do on their new album Proto, by simply singing together with other people in one space, by sharing the making of music joyously on stage, by attributing the work of others ethically, and by speaking openly and honestly about the processes of your creation. The second technological intervention feeds from this collectivity by having a machine-learning algorithm learn from interdependent data. In the following interview you will read about the whys and hows of this approach. Having evolved from ‘net-concrete’, the processes of machine-learning are used here to outclass the ‘magic’ of causality and the escapism once found in searching for an ‘exit’ or an ‘outsideness’. Technology in Proto, to my mind, becomes instead a worming inward to the future. This is music of an inwardly focused, closed and walled world, where economic, symbolic and systemic change is up for grabs.

Mat once played me an early version of Spawn and it reminded me of a needle being traced around an unearthed clay pot, a voice and its vibration trapped in a solid form like early recording practices of wet wax and hand-held cone. In today’s moment of algorithmic and network exploration, the analogy to early tape practice is apt, but I would add another image to describe this conceptual moment of an early and prototypical graphism. Experiments in the musical use of machine-learning are similar to the adaptation of figurative pictures which eventually became the first forms of writing. The invention of writing foretold a legacy of empire, of memorialising money and gain. Today, at a similar moment marked by the development of new language, the use of machine-learning is an attempt to reclaim technology from empire, as charted in the following interview. Maintaining an air of artistic rarity and mystique, as Mat describes, is probably a good economic decision in this industry, but you won’t find opaqueness when talking to Holly and Mat. Holly and Mat are honest and open. All is up for discussion here. I hope you enjoy reading the following conversation as much as I did.

*Minor edits have been made for the sake of clarity, I’ve also done my best to provide a lattice of references and links along the way.

— Emile Frankel, June 2020


SD

So glad you’re here. The overview for the event says that we’ll be talking about artificial intelligence and machine-learning technologies, focusing particularly on Spawn. I thought we’d just go straight there, if you could talk about Spawn, the development of Spawn and what Spawn is?

HH

Spawn is what we call our AI baby. That’s the headline grabbing part, like, ‘Oooh, AI baby!’ But what it really is, is a metaphor for some experiments that Mat, Jules LaPlace and I were doing with machine-learning and music. Three years ago we were given a grant by the German government to celebrate the death-a-versary of Beethoven.

SD

We’re really glad he’s dead!

HH

2020 is the 250th death-a-versary of Beethoven. The concept of the project was ‘what would the Beethovens of today be doing?’ I don’t have the hubris to say that I am the Beethoven of today, but they were giving out small grants for artists to pursue research projects that don’t really fit into ‘normal’ practice, or, who might not have had the economic incentive to go down that path. It was really nice to be given the resources to hire artists, to buy some hardware, and to hire Jules. Jules is a developer that we met while we were living in California. He’s a badass, crack developer who is also a musician. Being able to share a musical language and to have him understand what some of our goals were earlier on was really great. So we just started messing around with creating our own dataset, cobbling together different software that we found on GitHub. For the first six months, it basically sounded like shit. It was really awful. We almost gave up, but then we had a couple of breakthroughs. We started using SampleRNN, which is a recurrent neural-network technique. That’s when things started to sound interesting for the first time. This can be heard in the track on Proto called ‘Birth’, and that’s when we started to hear the logic of the network.

MD

I think you’re right, in saying that SampleRNN was a breakthrough moment, but actually, the bulk of the work we did on the record didn’t involve SampleRNN at all.

HH

That’s true.

MD

I only make that distinction specifically because the correlation between this technology, and what we classically understand as sampling technology, is something we’ve tried to step away from. Has anyone seen those 24 hour death metal YouTube things? Right? That’s SampleRNN. The basic idea is that you can train a network on a piece of sound and then it will try to anticipate what comes next on the basis of what came before. It creates a pretty interesting illusion of development, or generation. We were desperately trying to get away from the idea of just ‘recreating’ something, as impressive as that technique is. The bulk of the interesting stuff, that I would consider to be Spawn, is more to do with training neural-networks on a corpus of sounds that were created by us, our collaborators, and audience members from different performances that we were doing.

HH

When we started the project, we didn’t really have a clear idea of what neural-networks could do. We really wanted to just get our hands dirty with it, so that we could build a conceptual framework and understand our own stance on the technology. When we started researching it, we found that a lot of people were using it … Like, you guys have seen Infinite Bach, where you train a neural-network on an existing composer’s corpus, turn it into MIDI data, and then you can create Bach forever? Or music that sounds like Bach, forever. There are a lot of examples of machine-learning projects like that. A lot of examples where algorithms are hoovering up anonymous audio or anonymous MIDI files online, and training on these sources. So from a very early point we decided that we wanted to focus on audio material because we didn’t want to abstract audio into MIDI files. There were many aesthetic and political reasons for that. The other thing that we wanted to do was to create our own datasets. We didn’t want to use anonymous datasets, or use really large datasets, such as from universities. Those were two early constraints that we put on the project.

SD

How long did it take for you to get to that realisation of wanting to use custom data sets?

MD

From pretty early on. Most of the neural-net software libraries that were being released were focusing explicitly on MIDI. To the extent that when you speak to the vast majority of developers or most companies, and when they think about machine-learning and music, they think about ingesting MIDI and spitting out MIDI. That goes back to the early 80s.

HH

Which is why music, from a very specific period of time, is often used as an example for machine-learning. It’s flattered by the score, such as Beethoven, for example. This is when the score was at its apex, when the score was often even valued above the live performance. But over the last hundred years or so there are many kinds of musical parameters that aren’t very well captured in MIDI data. And those are a lot of the things that I care about, like context, timbre, etc … It became really clear that even though you’re not physically sampling, this process was similar to sampling. You’re essentially sampling ‘the vibe’ of something so heavily that it felt like a legal issue.

SD

What I think that you’re pointing out is an issue concerning the reanimation of the dead. Taking this archive, and continually refreshing it — or not even refreshing, but propping it up and keeping it alive. Going back to Spawn, you’re very quick to insist that Spawn functions as a metaphor. I’m interested in that, this need to assert that Spawn is a metaphor and not the thing itself.

HH

Well, I’m a little bit allergic to some of the kitsch around AI narratives. I was worried that by using a ‘baby’ metaphor, people would think we were actually having this baby. With all the femme bots and stuff that we see in this emergent culture, I didn’t want to perpetuate this further. That’s why I was trying to be clear, and I was trying to explain that Spawn is not just one, fixed thing. It’s multiple things, multiple processes.

MD

We’re being polite in using ‘machine-learning’ and not bringing up the AI question. In most of the proxy wars happening around AI, the conversation exists in the realm of metaphor. Most people have no idea about what is technically going on. That’s always the tricky thing in our case. It’s different with us individually, but collectively, we have similar experiences. From one angle, we’re putting out a popular record that we understand is going to be pushed out, and hopefully teenagers will care enough to come to festivals. That’s the thing that pays the bills. And then there’s this other side, your academic facing one, where Holly’s writing her PhD thesis. So you literally can’t bullshit, right? There are credibility issues, where many people — who I will not name — feel completely free to talk utter shit about AI and basically lie to people. I mean, fraudulent claims around the uses of AI music. Like, we can’t do that. You, [Holly] specifically can’t do that, because your thesis advisor at Stanford said so.

HH

But, also, I don’t want to do that!

SD

Could you talk a little bit about why you’ve given Spawn a gender?

HH

Yeah, this has come up a lot. I think the reason is that the first training I did with Spawn was with my voice. The first utterances I heard from Spawn were my voice, and I identify as she. I heard myself reflected back to myself. Some people have taken issue with this because of the history of female digital assistants being treated poorly and then that translating into the real world, into the human world, of treating female assistants poorly. I think that’s a real issue. But also, this project is about trying to create a counter narrative, another vision for this. So I didn’t feel like I had to be shackled by that version.

SD

Treating female assistants poorly makes me think about the way we treat the living. Just as a segue into the way we train our machine-learning models, I’d like to get you to comment on the particular ways that Spawn was trained, in contrast to something in the vein of Susan Bennett. If you read how she was treated during the training of Siri, Susan Bennett was working four hours a day. That doesn’t sound like very long, but it’s a really long time to be reciting nonsense sentences, cleanly and clearly, five days a week, for a solid month, in order to produce the voice of SIRI. Then she was more or less tossed aside. She only recovered years later, for various reasons. But the way that Spawn is trained obviously operates according to a different kind of logic. I want to ask you about both the public and the private performances of training Spawn. It seems as if there were different modalities.

HH

Yes. That’s another reason why we use the baby metaphor. We see her as being raised by a community of vocalists that we’re working with. Our ensemble. The idea was not to use random data, but to actually hire people, to name them and pay them for their time. But this gets more complicated as we do public performances. I think we always try to be respectful of people’s privacy.

MD

The first time we tried this, two years ago, was with a piece called Deep Belief. It was a theatre piece, and the whole idea was that we had icons who were all ensemble members that performed characters. The characters were symbolic, or representative of a different vision of where technology might go.There’s also a big distinction between recording the individual voice of somebody that you can identify and recording the abstract voices of thousands of people. But in terms of anyone who we sat down with and said, ‘we want your voice to participate in this record’, they are credited and were paid for their time.

HH

That was also the stoner thought: that as soon as something becomes captured in a medium, a photo, a video, it becomes machine legible. It can become part of a training corpus that can create something from its own logic. And that seems really obvious, but once you come to that conclusion it’s like ‘oh my god’. Everything could be modelled! Even from this talk, if somebody was recording it, our voices could be modelled. Then you start to get paranoid. You go down the wormhole, of what does it mean to have a voice? What is it? What is vocal sovereignty like? Like, traveling in Australia, I hear this amazing Australian accent that you all share. Voice is a shared communal thing, but you perform it as an individual. So it opened up a Pandora’s box of what it meant to have this audio fingerprint.

SD

You bring up a fear which is quickly being realised: that your voice can be synthesised from you. It takes as little as five seconds of audio to do. That is definitely terrifying. It brings up this question of consent, which doesn’t normally enter into our thinking. On the topic of ‘giving up’ data, something that is not often thought about is the terms of the reanimation of data. The reanimation of the dead.

HH

We don’t have proper legislation around it. I don’t know what it’s like here but, for example, the prison system in the United States is awful. The Intercept released an article a couple years ago, about how prisoners were caught in an ‘opt in only’ contract. If you wanted to use the phones in the private prison your voice would be modelled and the company would then have this model. Then in any future investigations, they could be scraping, trying to listen for your voice. Not only the voice of the incarcerated person but whoever they were talking to, who was not incarcerated, on the other end of the line. Things like that are happening in real-time, and sometimes the laws don’t catch up in time.

MD

You can clearly imagine future employers acquiring that data set, to screen convicted felons for future employment opportunities. It is distressing. And, further, on the consent issue the late Whitney Houston is going on tour today and presumably never personally gave consent to tour posthumously. There’s a whole lecture we give on this too, on Lil Pump and XXXTentacion. They were two SoundCloud trap stars who both sadly died during the past couple of years. Even though they didn’t like each other when they were alive, their management and family have given consent on their behalf to collaborate together posthumously. Isn’t James Dean being reanimated for a movie? … So bleeding edge.

HH

This is why I keep using the word necrophilia, artistic necrophilia. This is a term that Miles Davis used, when he criticised sampling when hip hop culture started. He thought that each generation should create a new sound that reflects their current conditions and, arguably, hip hop did exactly that. So I think that was a little short sighted but it’s also quite prophetic, when you look at these Tupac hologram shows, or the Whitney tour, and so on. From a lot of things, like voice modelling, or ‘vibe stealing’, there are things that you can achieve through machine-learning which make us question what we should do with our shared archive. Maybe this is something that’ll be completely passé in the future, like the way that we look at playing an old film. Maybe that would’ve been hideous to someone two hundred years ago, but it’s just something that we have to deal with legislatively. Together.

MD

And there’s no going backwards, right? That’s part of the reason why we’ve both been given pause. Questioning some of the ubiquitous logics of sampling, and just general entitlement to media, that has been the countercultural, commonplace mode since the 1990s. That’s the cultural mode that I grew up with, this libertarian, individualist ideal, that everything is basically mine to play with. So long as I’m not punching down, whatever that means, or whoever arbitrates that.

SD

It’s like, negative land.

HH

And it’s not to erase an entire history of music that we love. Of course, Pierre Schaeffer liberated sound from its origins, and all these things are really interesting, but they are also time specific. I mean, when you liberate a sound from its context, you also liberate it from its author. From cultural and religious content, from whatever context surrounds it. When we look at the history of sampling, there are many egregious examples, such as the Enigma song ‘Return to Innocence’, where they take Taiwanese aboriginal music and make a euro-trance song out of it. Luckily, they were taken to court, and then the musicians were reimbursed, in some ways. But that’s the kind of logic on display here. If we don’t deal with it now — and we’ve done a really shitty job of dealing with sampling — we only have two options. There’s freedom of information for everything, or hellacious digital rights management from the major label side. You don’t want to choose either of those. Nobody wants to be a sample cop. But what is the third mode?

SD

To ask that question more specifically to you, Mat, perhaps you could speak a little bit about Saga and the lessons that you’ve drawn from it. It hasn’t been updated in the last four years, so it feels like something that you’ve left behind?

MD

For those who don’t know, years ago I worked on a project called Saga. It’s basically a decentralised publishing framework. The idea is that when an artist, publisher, or anyone who wants to publish something publishes their work online, Saga gives you the tools to alter your content in every discrete location where it exists online. For example, you post the track to SoundCloud, the track disseminates across the web, people embed it. In this example, what happens today if you took exception to, say, Mercedes Benz embedding your track on their blog for free? You would either have the choice to take your track down everywhere or leave it up. The idea with Saga was that you could have little switches. You could say, ‘No, Mercedes Benz, I don’t want you to have it but this kid here can have it.’ Then, because that railroad has been built, you could additionally publish code. For example, ‘No, my song, video or text is free everywhere until 200 people view it in one location’ and then afterwards people have to pay money to view it. Part of the idea at the time was looking at how advertising runs the web. If you make something, put a bunch of time into it, then publish it online, there isn’t really a mechanism that allows for you to be compensated for that. And then what if you find yourself on an Intel blog and, all of a sudden, they’re driving all this traffic to their blog and ostensibly using what you did. There are even more egregious examples. Imagine finding your artwork on a white nationalist’s blog. What would you do in those circumstances? What kind of recourse do you have, when it’s all just out there? Saga was an early attempt to think through some of these issues. Why is it that we’ve accepted this, in the digital realm? This common sense, realism idea that once it’s out there, it’s just out there. Who does that benefit?

Saga was, and is, now a dead project. But the original idea was about allowing people to have control of their work in a digital space. Just as easily as if it were a piece of physical work that I would choose to put in a room. There’s a long story as to why we aren’t working on Saga anymore. But it does tie into this idea of not wanting to create impenetrable DRM, as you said, or to go back to a time of wanting to hoard information, but trying to think of what a middle-ground might be. I’m happy to have an argument with someone about this, but freedom of information narratives — while they might sound very seductive — ultimately helped Google. Freedom of information led to the creation of large entities who can, under the subcultural, countercultural cool of sampling, or the dynamic of moving fast and breaking things, feel at liberty to take whatever they want. That’s the origin story of Google. For better or worse. Google indexed the web without asking any party and created a new map upon which they could sell services that nobody else could offer. That connection is still in music and sound communities. It’s difficult, I’ve enjoyed a lot of sampling music, I’m really happy that this music existed and so forth. But that kind of countercultural tool is ultimately the same kind of logic that has been used to create Silicon Valley and its impenetrable companies. But when you point to sample music and say, ‘well, are you doing anything different?’, many artists that were quick to criticise Silicon Valley might back down. It’s a much longer conversation.

SD

What we’ve been discussing connects to the right to die, the right to be forgotten on the web. Just because something exists it doesn’t mean that it has to exist in perpetuity. Obviously, perpetuity is in the interests of Google, but how does this connect to the right to silence, or to the right to not have to say anything? In AI, the generative potential of these neural-network models is that you’re able to reanimate the dead, that you’re able to create a model from past performances, from an archive as a dataset. But what happens when you generate something ‘new’ from data which originated in a community? Who, really, does that information belong to? It didn’t originate just in your genius brain. That originated in a community, so it should go back to a community…

Music is one very audible, or sensible way that we experience this. But on an everyday level, after we pass away, our data is going to continue to have some agency, through some neural-network or another. Applying this to generative models, we see people who work with AI saying stuff like, ‘Oh, I generated this new thing, and it sounds weird. Isn’t that really cool?’ … Perhaps you could comment on that, on the ways that working with AI restructures the way you work with other people, with your collaborators. The way it generates new limitations and affordances?

HH

We see AI as a new way to collaborate and as a contemporary coordination mechanism. That’s why we created a thread between early vocal music, vocal folk music and using the voice with neural-networks. We were thinking about early vocal folk as a technology, as human coordination technology. People used nasal singing, in the upper parts of their faces, to make sound travel really far across mountains, to scare predators, and these kinds of things — as a coordination technology. I’m always trying to think about how we can make the computer a performance instrument and Spawn isn’t really there yet. We’re trying to get there but, for me, it’s always about how I can use this work in order to coordinate with other people so that we can play together. Specifically with this ensemble, I wanted it to augment our communication with each other. For us, it’s really important that we’re able to have these euphoric singing moments with our ensemble. We don’t want working with a neural-network to become an awkward or heavy process, it shouldn’t be a block. Instead it should be helping, facilitating.

MD

When we started working on Proto, before there was an ensemble, we thought to ourselves, if this record is going to mean anything then the technology has to be an ambient thing that witnesses real human activity. What you were saying about performance, Holly, is that it ultimately kills or slays real human activity. Most discussions around machine-learning and music are dry demonstrations that ultimately demonstrate more of what we’re lacking than what we’ve gained. For example, Amazon did some heinous generated audio thing. This is, presumably, state of the art technology, because Amazon has a lot of money to throw at this stuff. But when you witness it, it’s like … I don’t know anybody who would think, ‘wow, the future of music is going to be incredible!’ It shows the opposite. It demonstrates … a bedroom noodling thing, whatever.

HH

There’s nothing wrong with bedroom noodling!

MD

No, no, sorry. I’m not saying …

HH

What he’s saying is the generic identikit, like, ‘Rock kit, with beat’.

MD

Yeah, exactly. That crap.

HH

We all love a bedroom noodle.

MD

… I’m fascinated by the economics of all art. Once you understand it, everything becomes much clearer. Small economic distinctions and decisions, like how much something costs, who gets the visas, who runs it, everything. There’s already a habit of cost cutting and efficiencies around the festival circuits in the art market. Where demo art does do really well, particularly digital demo art where you can send a file and it can appear in every biennale instantaneously. That exists in music, too, but is also somehow more impoverishing. So part of the grand gesture that might be appreciated by this room, more than most conversations in the world is, how do we make the most irrational economic decisions on this topic? Because, actually, flying a bunch of people around the world to sing is really economically irrational for someone in our position. But it’s possibly the best thing you could do, when talking about AI or automation, particularly in an electronic music context … The point is to try to think about how technology can actually augment relationships and not exploit people in the process. This is what we have spent the most time deliberating about.

HH

I think that’s an early kernel of why we put together a vocal ensemble.

SD

One way that I’ve thought about similar things is in observing the way that a GPT2, which, if you don’t know, is a language model by Open AI (Elon Musk and Microsoft also funded it). You type in a sentence and the AI generates the next few sentences. It’s remarkably good. You read GPT2 and you think, ‘Oh yeah, that’s human-like, maybe I could have written that.’ And then you see people sharing these sentences, copying and pasting them, and everyone using it. You keep hitting ‘generate’, ‘generate’, ‘generate’. And there’s a disposability to every single artefact that’s generated. I was thinking about this disposability of labour when we become data that is reanimated through models like these. When the labour of writing doesn’t exist anymore and there’s no human labor and no real cost (except an environmental cost), when humans are cut out entirely … there’s a total devaluation of life. It seems like what your approach has done, at least in my reading of it, is to confront that disposability of AI through a reclamation of the training process.

MD

Yeah. Every time these machines get more efficient, it becomes easier to make meaningless art. It takes human beings to give things meaning, right? The thing that sounds a bit different, or maybe has a story associated with it, or a community associated with it, will be more valuable and scarce and rare as a property. It will stand out in a sea of nothingness. Fetishising how easy it is to generate fairly competent, meaningless art isn’t really that cool. Unless you are a developer at a research organisation, whose task is to try and render the most convincing piece of meaningless art as possible. And there’s nothing wrong with that, that’s an impressive accomplishment, but that’s not the art that I care about. Why would I care about that?

HH

The ability to ‘steal a vibe’ is something I care a lot about. Don’t steal my vibe! I think the speed at which you can ‘steal someone’s vibe’ is going to become lightning fast. It’s something that Peli Grietzer writes about. If you guys don’t know who he is, I would check out his work. He wrote an article for Glass Bead a couple of years ago called A Theory of Vibe, and it explains how a neural-network can efficiently analyse and understand the vibe of something and reproduce it without really understanding where that vibe might have come from. As someone who cares about ‘the bedroom noodler’, who comes up with something that is unique to them and then has their ‘vibe’ taken away immediately … that’s what I get concerned about, in a scary dystopian way.

SD

We’ve been talking mostly about AI in the production process. Could I use this opportunity to nudge the conversation to discussing the consumption and distribution process, and the role of AI in ‘vibe matching’, auto-generation, and stealing within platforms like Spotify? Within the act of listening, how is listening guided by these silent actors?

MD

I really don’t like Spotify. I mean, in trying to say something smart, the parallel is the entitlement found in a libertarian, individualistic sense. The fact that you can take something, decontextualise it, and all of a sudden it’s yours makes me really heavily critical of DJing as a practice, even though the history is venerable and wonderful. I have many friends who are DJs, they are wonderful. But actually, I had an argument with someone –– who I won’t name because I respect him –– on Twitter. But what he was saying was, ‘When I take a piece of music and I play it in my DJ set, it then becomes mine.’ And I was like, ‘Well, what do you do to it?’ Because to me, it just sounds like Usher, or whatever. What have you done to this track so that all of a sudden it’s yours? This is just entitlement, right? It’s like, a Lockean, settler colonialism idea. Saying, ‘I’m here, and I’m making value out of this piece of media right. It’s mine!’ … That’s a really long, deep liberal argument. A very profound argument …

Again, in relation to this 90s countercultural moment you should all go and watch the interview with Chuck D and Lars Ulrich, about Napster, on MTV. It’s fascinating. It’s amazing. And Lars Ulrich was actually so spot on, even though he comes across like a cock. You really get this feeling of this 90s moment. Where there was tech, where everything was possible. It was going to be the hip hop moment. It was going to be the remix moment. And how that opened this back door, to companies saying ‘look, just trust us! We’re gonna re-order everything!’ And then of course Spotify comes along with this act of decontextualising something, removing any semblance of information about where it came from, who supported it, the location. By design, Spotify alienates that piece of media from anything, any amount of labor or love that went into it. By design. Spotify are not stupid, they know exactly what they’re doing. That platform is designed by hundreds of people. And when you listen to a track you can’t even find the label. You can’t find the city. As little attempt as possible is made to root or situate this media. And this is ultimately an act of displacement and stealing. What hubris and entitlement, not only of the people working on those systems, who make these playlists (and we’ll have a race to the death of some random anonymous people who probably get paid a tonne of money to run these big playlists), but also of the individual who’s listening. There’s a degree of entitlement where the most common argument to justify Spotify is that it’s better than piracy, but that’s not really an argument because there is a degree of individual responsibility there.

Again, I don’t want to be Lars Ulrich, and tell everyone that you need to spend all this money that you can’t earn, which is because of the digital economy, on albums that don’t exist anymore, or whatever. But there is a root, an original sin there, of entitlement. And we can nerd about this for a long time. I often talk a lot about Tim Berners-Lee or Al Gore, and the original sin of HTML being an entitled act, and lament the fact that I can reference anything online, on my special page, and that there doesn’t need to be back attribution to where that came from. HTML allowed for that possibility. It is that 80s, 90s libertarian Californian ideology that led to that. Anyway, my long way of saying that Spotify sucks.

SD

I know that giving credit to a web of people who are involved in the generation of an idea is something you talked a lot about with Platform. Because the metaphor, the language of Platform is already about the fact of the platform as something that gives visibility and enables others. The role of the platform is to enable the behaviour of other people, for other people to work on, and build atop what you have created. Then you move to Proto, where AI is a discursive terrain. You mentioned paying the people who give their voices to Spawn, but in light of what Mat was just talking about, on algorithmic populism that drives taste on these platforms, how do you go about crediting a voice in a crowd? Again, what does giving credit mean?

HH

It’s difficult, because I feel like the whole industry is set up for the genius individual, meditating on the mountain. Everybody wants me to be a blissed out synth queen who’s coming up with all the ideas myself. If I push back against that image, then it’s like I’m pulling back the curtain and losing some mystique. It’s happened so often: I’ll have an interview where I’ll talk at length about the people I work with and then you read the interview, and it’s just like ‘me, me, me, me, me, me, me, me’! They just cut it out! You have to actively ask people, ‘can you please print this’, ‘can you please mention people by name’, ‘can you post people’s websites’, things like that. In this attention economy, we’re incentivised to just push ourselves as individuals as much as possible. It’s a constant balancing act and a bit of a struggle. I’m also not somebody who feels super comfortable being the centre of attention all of the time.

SD

In your interview Re-Engineering Hegemony you talk about what post-capitalist music might mean, or what it might look like. You reference what I think is an ongoing fantasy of automation, which is that if we manage to displace enough work to technology, then we’re freed up to make the most interesting decisions, or have different relationships with each other?

HH

I think that’s what we tried to do metaphorically on the stage. Now that we have all these intelligent light systems and projectors what can we do onstage instead of just standing there? It means that we can be joyful together, hug each other, love each other, enjoy each other, and ultimately be more human together, because we have time that’s been freed up to do that. That’s the ideal.

SD

You often talk about new fantasies, that as a musician you’re able to enact these new fantasies. I feel like what you’re describing in these shows, where there are a lot of different people involved — different modes of listening and attention, different modes of spectatorship, modes of production — is the performance of new fantasies.

HH

A lot of that came from touring a lot and being part of an electronic music community. There was a lot of dystopian fantasy around at the time. There’s such a history of dystopianism in electronic music. Of course, I come from that lineage too. A lot of that stuff is really amazing. But I was looking for something that was presenting an alternative, like asking for a seat at the table, as the music economy was changing so much. I wanted the musicians around me to also be making demands for what they wanted, instead of just saying ‘this sucks’ and just being unhappy with the situation, revelling in the situation. It’s actually much harder to come up with a vision of what you want. I felt this dystopian approach was giving up power to those who were already in power, by just ceding control. My practice, as someone who is obsessed with vocal processing, means that from the very beginning music is a fantasy moment. I’m transcending the physical limitations of my voice. I’m an okay singer, but I’m not, like, an awesome singer. But from the beginning, I was like: how can I do something that’s better than myself, bigger than myself, or more fantastical than my physical body?

MD

I’ll speak from maybe a different economic angle. The post-capitalist question is really interesting to me. How can you make an idyllic scenario that is actually visible? Because that’s a difficult question, more than making an idyllic utopian experiment in a vacuum, how do you make something that is resolute and resilient enough to be seen by masses of people? For those who don’t know, Alex Williams, who conducted the Re-Engineering Hegemony interview with us, wrote Inventing the Future with Nick Srnicek. He was a leftist accelerationist type; it’s funny because Alex and Nick were the poster book of the leftist accelerationists. This brings me back to that cultural moment, the idea and search for a post-capitalist music. At the time there was a proxy war that was happening discursively between Dis Magazine and PC Music. These were what I would consider, right accelerationist approaches to making music, with attitudes like, ‘Run it into the wall! Nike! McDonald’s!’ Out of such hyper-pop, accelerated advertising music, how can you make a critique? And again, I like a lot of those people. But you have to have an opinion, right? Then there’s the other side of it where you have to think, okay, how do you not dismiss the potential of new networks and dismiss all media as being corrupt and cynical, which a lot of it is? How do you embrace some of this stuff? To try and carve something out that might be positive, even if that could ultimately be a futile exercise. That was where Platform came from, from an attempt to open up this conversation. Like, hey, platform co-ops. And there’s all this cool stuff that you could do that’s maybe less libidinal!

HH

But that approach is much less functional in a music economy.

MD

And that’s a problem. It’s really difficult here, because as you said before, Holly, you try and credit all these people that you work with. And then when you read the article, it’s just about ‘me’. But my counter argument to this is, if it wasn’t just about you, would we even be sitting here? Would we have a career? That’s the challenge, because culture is like a sieve. You think you can just make this beautiful thing and push it through but then some things die, they get abstracted. Ultimately, the economics of the industry will dictate what products, or what people end up stepping through.

HH

Particularly when you actually care about your collaborators.

MD

And particularly because we’re academic-adjacent, so we’re really just corrupted the entire time, right? We’re hypocrites. Everything we do, we’re hypocrites. The most difficult thing to do is to find some way of having some of those embers survive, as they pass through.

HH

You have to Trojan-horse it.

SD

And I can also see the appeal of universal basic income, which is something that’s come up a few times around your work. What you really want is not to develop the perfect system, so that everyone is remunerated precisely for the contribution that they put in. That kind of overzealous accounting, it’s horrible, you just want everyone to not have to suffer in order to live. The basic communist creed: everyone gets what they need, according to their needs. As simple as that.

MD

That’s also the challenge when it comes to these machine-learning systems on that bigger level. Is it really socialism, or is it barbarism in this scenario? Because these things work with a lot of information. You don’t have the local farmer’s market version of big AI if you want to solve cancer, you have to have the data of the entire country working together.

HH

I’ve been thinking a lot about what it means to have a protagonist lead vocals on a stage. What it means to be a front woman, because now I have an ensemble of people that I’m working with. Is it possible to play that role, but to also pass that baton over, and have someone else on the frontline for a little bit? That’s something that we play with. It’s not fully baked, but it’s something that we’re experimenting with.

SD

I don’t know if it’s the same thing but when I was younger I tried to set up projects so that power didn’t exist. But then I came to the realisation that power is always there, whether you want it to be or not.

HH

It’s funny that you bring that up. In some of the early experiments that we did, we wanted to try out ‘non-hierarchical’ jam sessions. And they were the worst jam sessions we ever had. Turns out, you do need somebody to intervene. But it was actually important that we went through that process. Of being like, ‘no, I have to be the leader here’, but that doesn’t mean I have to be a shitty leader or an asshole. You can actually be a kind leader, while having power dynamics that aren’t so hierarchical or so stable.

MD

Society pushes you. Power exists ambiently, whichever way you deal with it.

HH

This is the thing that really turned me off classical, neoclassical, contemporary ensemble music work. I noticed that if I didn’t exert myself as a super dominant composer, the ‘you’re doing it wrong’ composer, it wouldn’t work. There’s this S&M dynamic between performers. Sometimes, they want to be told ‘no, you’re fucking it up! Work harder!’ But that’s not how I like to work with people. I had a couple of experiences where I had to hire a conductor to do that. But I couldn’t do that myself! I just couldn’t embody that. It’s really about finding the right kind of communication balance within ensembles. The writing of the music is much less difficult.

SD

Any questions from the audience?

Crowd member 1

In talking about file stealing and Spotify sucking, would you say that this will occur in a not-so-distant future? If the end game for platforms such as Spotify is to just have automated, generated music by stealing ‘vibes’, how far away do you think that possibility is?

MD

It’s quite far away. The ‘vibe’ question is a very real question. Right now the technology exists to take the vibe of a composition and render it into MIDI. So, rendered into the impoverished abstraction of MIDI, discerning the ‘vibe’ of a Kendrick Lamar beat will happen, but we’re quite far away from that. Piano playlists do really well on Spotify. And yeah, go home, make a neural-network to spew out piano playlists on Spotify, and you might make some money. But I think what you’re talking about is those automated systems taking over, which we’re far away from. And, actually, that’s a nice thing. In the short term, in our lifetime or career span, there are more opportunities to do cool work in the augmentation space, rather than in the replacement space …

HH

I disagree a little bit. I think that when it comes to genre-specific music, we’re not that far away from being able to replace it. But, if you think that composition is more than just the statistical analysis of notes, as in, it’s meaning making — and a lot of that meaning making is the context around the music: the concept around it, who’s making it, and where it’s being made — that’s the squishy, messy stuff, that a neural-network is really bad at analysing. But in terms of being able to create a generic house beat, I think we’re really close to that. And I think a lot of people are satisfied with that as their music. But a lot of people are also not satisfied with that. I think Hans Zimmer is going to be fucked … But I also think that’s okay!

MD

I think his assistants are totally going to be fucked. Yeah, you’re right. In genre specific experience rendered into MIDI, and the actual compositional process? Yeah, we’re screwed. But in the timbre-like sonic elements? It’s close. But actually making it mean something? Nah. Then again, there are a hundred thousand people going to the most boring techno festival in the world. But who am I to say what people give meaning to?

Crowd member 2

My question is that when you listen to Spawn, do you still feel like you’ve created that? Do you feel distance or do you still feel that connection, of the music being yours?

HH

It’s like when you write something for a human ensemble and they interpret it for you. It’s never 100% entirely yours because somebody interpreted it through their physical body. But you’re still the composer, because you wrote it. That’s how I feel about using a voice model. Spawn is performing a line we’ve given her. ‘Birth’ would be the most improvisatory track, where there’s a line between composition and improvisation. That’s when things start to get really blurry.

MD

I’ve seen some representations of the record, and you can’t control what people write about it, but people talk about it as ‘generative composition’. And that would be misleading, because a lot of people would also think that oh, yeah, the AI just came up with this music. No! That Jlin track was the most laborious thing on the record. It wasn’t a process where you say, ‘Spawn, just make some music!’ This actually took so freaking long. The actual print component at the end took a couple of minutes. But to actually get the system working, and to tweak it to the point …

HH

Spawn will perform something you’ve given her and sometimes she’ll do something unexpected. But, in a way, that’s what all human performers do.

SD

One place I didn’t go is the experience that I’ve had in working with AI, too, in thinking about the childlike aspect to it. On the one hand, it’s the most advanced technology, it’s the pinnacle of all these advancements, of research and computation altogether. And at the same time, if you actually experience working with AI, you feel that it’s a little child, learning how to talk. They do it all wrong but, still, you can tell that they’re really smart, intelligent.

HH

That’s why we wanted to work specifically with audio instead of MIDI material. With MIDI, you create a ‘clean score’ that you push through a Native Instrument suite or something. The outcome can appear really perfect. With audio, you can really hear the imperfections; where it’s still dumb, you know, she’s not there yet. These imperfections are really powerful and really silly at the same time.

SD

I think you are recognising that we’re at this moment where culturally, historically, we haven’t yet realised that all these dystopian or utopian narratives are in unfixed emergence. We’re still in a process, and there are still openings to intervene.

MD

That was the idea of calling it Proto –– the precedent before something happens. We joke a lot about artificial intelligence because a lot of the time there is a primitive thing that’s occurring. And there’s a pristine representation of that, you know, it comes up with a score, and then you give it to the chamber ensemble of Singapore to interpret the score, and it sounds incredible! And you’re like, I’m pretty sure it didn’t come out like that.

HH

But that also ignores all of the incredible human intelligence that goes into performing that thing.

MD

Exactly, the artifice there also obfuscates the human labor that’s going into training the freaking thing. The differences in output are so minuscule that, unless you get really deep into it, you’re not really going to appreciate it. But the practitioner community is quite open and shares a lot of stuff. This reminds me of the tinkerer generation, or nerding-out over early tape experiments. Like, this could be really crazy in 20 years time. Really crazy.

HH

But if we tried to make it sound as if it was ‘really crazy in 20 years time’ and super polished … I don’t know. I think it’s more interesting to show it for what it is today. And on that note, we should wrap it up. I’m hungry!

Contributor/s

Holly Herndon experiments at the outer reaches of dance music and pop. Born in Tennessee, Herndon spent her formative years in Berlin’s techno scene and repatriated to San Francisco, where she completed her PhD at Stanford’s Center for Computer Research in Music and Acoustics. Her albums include Platform (2015) and Proto (2019).

Mat Dryhurst is an artist who releases music and artworks solo and in conjunction with Holly Herndon and the record label PAN. Dryhurst developed the decentralised publishing framework Saga, which enables creators to claim ownership of each space in which their work appears online, and a number of audio plays that derive their narrative from the personal information of listeners. He lectures on issues of music, technology, and ideology at NYU, and advises the blockchain-based platform co-operative Resonate.is.

Sean Dockray is an artist, writer, and programmer living in Melbourne whose work explores the politics of technology, with a particular emphasis on artificial intelligences and the algorithmic web. He is also the founding director of the Los Angeles non-profit Telic Arts Exchange, and initiator of knowledge-sharing platforms, The Public School and Aaaaarg.

Author of Hearing the Cloud (Zero Books), Emile Frankel is a writer and composer researching the changing conditions of online listening. In his spare time he runs the Sci-Fi and critical fantasy publisher Formling.

Acknowledgements

This conversation was transcribed by Hannah Wu from the Liquid Architecture event, ‘Inhuman Intelligence’, in January 2020, presented with Melbourne Recital Centre.

Notes

    Disclaimer is a journal for new thinking and writing on listening and sound. Published by Liquid Architecture • Disclaimer is a journal for new thinking and writing on listening and sound. Published by Liquid Architecture

    Podcasts