It couldn’t have been a place and time other than the Bay Area, California in the 1970s—an era immersed by openness and experimentation across every domain of life: economic, social, political, philosophical, technological, and, most certainly, artistic. But the foundation was laid a decade earlier, when, as the well-known story goes, a pair of composers on a mission to discover new paths to musical composition recruited an instrument builder to design a new kind of electronic instrument. As influential as the early electronic music studios of Europe, yet idiosyncratically distinct in both style and culture, the San Francisco Tape Music Center was founded in 1962 in the attic studio of the San Francisco Conservatory of Music by Ramon Sender and Morton Subotnick. As it grew (and as it shifted locations), what began as a modest tape-recording setup soon expanded into a fully equipped experimental studio. In 1963, with funds from a Rockefeller Foundation grant, Sender and Subotnick commissioned Don Buchla to design and build what became the Buchla 100 series—his first modular synthesizer—opening up an unparalleled toolset for experimental electronic composition, and so the Center, not unlike the Village studio of Louis and Bebe Baron, quickly became a magnet for adventurous artists.

Conceived as both a collective workshop and performance space, SFTMC offered artists and composers an environment in which they could both explore new technology for their musical ideas, as well as showcase them to an audience of peers and curious listeners. Its core members and facilitators included Pauline Oliveros, new-media artist Tony Martin, and engineer William Maginnis, each serving as director at various stages. The range of collaborators extended further to other composers such as Terry Riley, La Monte Young, Steve Reich, and dancer/choreographer Anna Halprin.

Despite its vitality as a hub for creative music and experimental research, by the mid-1960s, the SFTMC, which really was a self-organized DIY collective with scarce resources, became increasingly financially precarious. Fortunately, Mills College, located just across the Bay in Oakland, was looking to expand its contemporary music program. Its department head, William D. Wickes, offered to relocate the Center to Mills and provide full institutional support. Relocation came with a name change, and in 1966, it became the Center for Contemporary Music. Its first appointed director became Robert Ashley, and several members of SFTMC joined in as faculty. Around 1970, Ashley invited long-time collaborator Mumma to join, who began teaching several courses covering DIY circuitry, cybernetic composition, and live-electronic performance. This fostered a deep-seated ethos among the school's affiliates that building electronic circuits was an active part of a compositional process.

1970s: Orchestras Of Silicone

Although technically CCM was now part of a proper academic institution, the core countercultural ethos was intently preserved, along with a relatively fluid structure. This was notably articulated in the centre's motto: “if you’re not weird, get out!”

Among the first generation of artists to pass this threshold were Maggi Payne, David Behrman, Terry Riley, Alvin Curran, Frederic Rzewski, Paul DeMarinis, and crucially for the story of network music, Jim Horton. Horton was an eccentric musically-oriented polymath, who initially enrolled to study composition under Ashley and Mumma, and by the mid-1970s he was actively developing his own ideas about music created with then newly available microcomputers, exploring topics of non-human agency, and artificial intelligence—and envisioning an ensemble of fully autonomous machines—the Silicon Orchestra. As Tim Perkis recalls in his dedication to Horton in ZKM’s 2021 monograph, The Hub:

“Without Jim Horton there would not have been the Hub… [he] saw early in the 1970s the potential of the newly available, inexpensive microcomputers to deeply change existing electronic music practice, pushing experimental music into the realm of artificial intelligence.”

×

As reported by his friends and colleagues, what truly set Horton apart was not only his imagination but the conviction and belief in the importance of the work he and his associates were engaged in. To him, this music was not just entertainment, but serious research. Horton was an early adopter of the KIM-1 (above), a single-board computer produced by MOS Technology, Inc., which stood out for its affordable price, and extendability. Others followed soon, and it became a standard piece of equipment among CCM composers. This prompted the creation of a regularly meeting group, where the composers would get together to show the pieces created on the computer. In 1977, Horton proposed to create his "silicon orchestra." Since cybernetic circuits and generative modular synthesizer patches had already been actively implemented, it was only natural to extend the connection to computers.

This catalyzed into a series of collaborations. First was a duo where Horton and Rich Gold interlinked two KIM-1 machines together running separate programs: Jim performed an algorithmic composition based on Leonhard Euler's harmonic theories, and Rich ran an artificial language program which he had written. Then John Bischoff joined, followed by David Behrman, and in November of 1978, the group gave their first performance under the name the League of Automatic Music Composers. Two years later, the group went through reformation, with Behrman and Gold moving to different projects, and Tim Perkis joining instead, with LOAMC now a trio.

It is difficult to exaggerate how astonishingly limited these early computers were by today's standards: one kilobyte of memory, a one-megahertz processor, and a hexadecimal keypad for programming directly in machine code. Programs were stored on cassette tapes, loaded laboriously, and prone to failure. Since the internet hadn't been invented yet, the communication between machines was ingeniously wired via crude serial links, moving only a few hundred bytes per second—sufficient to exchange simple numerical values, but never sound itself. The group, however, figured out how to program the KIM-1 to be audio-reactive, responding to certain cues in the sounds it "listened to," an approach that became a mainstay in the group's practice. The aforementioned On The Ocean by Behrman is the direct result of these explorations.

The LOAMC went on exploring the network composition and performance together until 1983, when Horton's declining health posed a necessity to slow down at first, and then disband fully. However, Horton's vision of a silicon orchestra was far from being disintegrated, and instead it fueled ex-members of the League to start a new ensemble. And so, in 1986, John Bischoff and Tim Perkis were joined by CCM's newer cohort of composition students and then faculty, Chris Brown, Scot Gresham-Lancaster, Phil Stone, and Mark Trayle, to form The Hub.

1980s: Join The Hub

The Hub took the foundation for network music set by the League and transformed it into a more stable, technically coherent, and performative form. The League’s approach was generally more chaotic, resembling free improvisation; as Bischoff described, “It was as if each player was in a world of their own, but they were interconnected.” The Hub, from the outset, introduced more method and intentionality into the group’s practice. Two important alterations made this possible.

First, the group took its name from the function they ascribed to a dedicated central processing “hub,” which they employed to manage and route communication among the ensemble’s machines—initially via a pair of Synertek SYM-1 microcomputers linked through a modem, later through Opcode Studio 5 MIDI interfaces, and eventually using UDP-based OSC protocol. This infrastructure allowed the group to scale their performances reliably and sustain more complex forms of interactivity.

Second, since each member would develop distinct compositions, they needed a new way to communicate musical ideas. Traditional notation or abstract graphic scores were ineffective. Thus, they introduced the concept of a spec—a specification document outlining the rules, communication protocols, message types, and behavioral constraints under which their systems would operate. These specs functioned simultaneously as both score and system design: they defined what kinds of messages (control values, states) could be exchanged, how machines should respond to incoming data, and how internal processes would evolve over time. During rehearsals and performances, The Hub’s network became a shared computational space in which each member’s system executed its own program code while remaining synchronized through common specs. This allowed machines to communicate, influence one another, and negotiate musical behavior across the network.

In this sense, The Hub bridged two eras—the anarchic, homebrew experimentation of the 1970s and the incipient digital interconnectedness of the late 1980s. Not only did their practice extend the cybernetic ideals of feedback and self-organization into the digital, but in their pieces they also anticipated many of the concerns of network and generative art to come—distributed authorship, interdependence, agency, control, and the aesthetics of emergence—all infused with a characteristic humor and wit.

For instance, in Tim Perkis’ The Minister of Pitch (1988), each musician controls one specific parameter of all other players: one defines the collective timing, another shapes the global timbre, and another is responsible for all the pitches. This piece, being among their earliest, also granted another layer of freedom to the “minister,” allowing whoever was responsible for a given parameter to define sub-rules for how that control would be exercised—some could provide a palette of pitches for others to select from, while others could set an active key for all players in real time.

Another iconic piece by the group, Stuck Note (1994) by Scot Gresham-Lancaster, builds on a common glitch in the MIDI protocol, where sometimes the Note-Off message would get lost after a struck note, causing the pitch to drone indefinitely. In Stuck Note, each performer begins with a single sustained tone that persists throughout the piece, and the pitch cannot be changed. The only parameters that can vary are volume and timbre—yet neither can be adjusted by the player producing the sound. Instead, these controls are manipulated by others through the network, interlacing a dynamic web of fluctuating and, at times, unpredictable sonic interplay.

A peculiar example of playful inventiveness can be found in Mark Trayle’s Crybaby (1991). The composition, as Trayle wrote in the spec, was inspired by the wah-wah pedal. However, what stimulated the writing of the piece was not the sound of the pedal, but the sweeping mechanical motion of controlling that sound. As Trayle put it, “The thing about the wah-wah pedal is that you can’t get from A to C without going through B; it ain’t discrete (in more ways than one).” He proposed a circular loop structure, where a sound source from a CD player or radio feeds a daisy-chain of sound processors, each operated by fellow composers and tuned to effects with oscillating components—chorus, phaser, flanger, or delay. The control data flows unidirectionally through the network, with each performer receiving MIDI messages that determine the frequency and depth of their oscillation. Before passing the data to the next player, each must transform it according to a personal translation table, introducing subtle mutations at every step. Echoing the analog gesture of the wah-wah pedal rather abstractly, Trayle expanded and distributed a simple rocking motion across the ensemble and the network itself.

It would certainly be beyond the scope of this article to delve deeper into The Hub’s extensive repertoire. The ensemble—despite the tragic loss of Trayle in 2015 to illness—remains active today, continuing to build upon its own legacy, investigating the evolving landscape of music technology. I strongly recommend the ZKM publication The Hub: Pioneers of Network Music, mentioned earlier in this article, if you wish to explore the fascinating world of ensemble and network music in greater depth. And with that, we move on to the next destination in the evolution of distributed music.

1990: Machines Are Learning

With the rapid development of digital technology—the internet and personal computing in particular—from the 1990s onward, the conceptual foundations and practical applications of network music began to branch out into several distinct pathways: generative and algorithmic composition, laptop orchestras, live coding, interactive and multimedia art, and telematics. George E. Lewis, an American composer, trombonist, scholar, and longtime member of the Association for the Advancement of Creative Musicians (AACM), reframed jazz’s core practice of improvisation through the lens of machine agency. In so many ways, jazz, with its spectrum of freedoms granted to individual musicians—from standard to absolute—intrinsically embeds the emergent properties of networks, and thus Lewis’s shift toward improvising machines was anything but unnatural.

In his composition Voyager (1985–87), Lewis designed a proto–artificial intelligence software system that performed as an improvisation partner. Unlike modern machine-learning tools, Voyager was built on a hand-crafted algorithmic process, yet the system could listen to a human performer and generate its own improvised multi-voice responses—sometimes following the human, other times taking the lead, or fully diverging toward its own drift. Lewis continued to work in this direction, and in 2018 summarized his thoughts in "Why Do We Want Our Computers to Improvise?" published in The Oxford Handbook of Algorithmic Music, Lewis reflected The Hub's socio-musical philosophy, by speculating that interaction with computers surpasses the mere simulation, instead presenting a way to encounter an “other” and probe how interaction, learning, agency, and sociality work when machine partners are treated as nonhierarchical, conversational actors with their own subjectivities.

For Pauline Oliveros, who by that time had firmly established her practice of Deep Listening, human participation remained focal, however she saw a powerful potential of these new technologies to expand it. She began experimenting with telematic or long-distance performances early in the decade, well before high-bandwidth internet connections made such practices commonplace. The precedent for sonic telepresence was set a few decades earlier by Maryanne Amacher, where in City Links (1967) series she used telephone lines to route several sound sources into a single exhibition space. This of course also echoed Thaddeus Cahill's proto-synth Telharmonium (c. 1900). Oliveros, in this sense, brought interaction into long-distance sonic transmission. Oliveros’s first documented telematic performance took place in 1991, linking six artists in different U.S. cities through a video-telephone bridge—a system that combined multiple video-conferencing endpoints into a single shared session. The setup back then was, of course, highly susceptible to latency, yet Oliveros approached the signal lag not as a flaw but as a compositional feature, an extension of spatial and temporal listening across distance.

She scaled up the practice internationally in 1999 with Half a Dove in New York, Half a Dove in Buenos Aires, a telematic improvisation connecting Oliveros, trombonist Monique Buzzarté, and artist Kevin McCoy performing live in New York with the experimental band Reynols, streaming from Buenos Aires. Oliveros deepened her research into distributed collaboration further, and in the late 2000s, together with Chris Chafe and Jonas Braasch, took an effort to formalize the practice by founding The Telematic Circle—an interest group dedicated to developing techniques and methods for telepresent music performances. The Circle functioned as an ongoing collaborative framework connecting several institutions: the Deep Listening Institute, Stanford’s Center for Computer Research in Music and Acoustics (CCRMA), and RPI’s Experimental Media and Performing Arts Center (EMPAC). The group regularly organized improvisation sessions, workshops, and lectures, and experimented with high-fidelity network audio such as JackTrip (developed by Chafe), as well as other remote collaboration technologies.

That sense of musical intimacy at a distance, and the value of telematic tools that Oliveros, Amacher, The Hub, and others put so much attention into, took on a new level of significance during the 2020 COVID-19 pandemic, when musicians and audiences around the world suddenly had to turn to network performance as a primary mode of collaboration. This is exactly what Lewis discusses in his contribution to the ZKM monograph titled "The Hub's Telematic Socialities." Oliveros’s long-standing vision of “deep listening across distance” gained a renewed sense of urgency, inspiring a new generation of artists. Although Oliveros herself had passed away in 2016, her influence ran unequivocally through countless online improvisations, remote ensembles, and collective soundworks that emerged during lockdowns. This was perhaps most clearly emphasized in the online revival of her 1971 text score The Tuning Meditation, organized by IONE and the International Contemporary Ensemble (ICE) in April 2020.

2000s-2010s: Laptop As A Musical Instrument

While telematic research focused primarily on the distributed property of new network technologies, other explorations built atop the personalized, customizable, and expressive features they offered—computers as musical instruments. Laptop orchestras, commonly abbreviated as LOrks, emerged in the early 2000s, growing within the academic settings of contemporary music departments across American universities, and eventually spreading globally. They are perhaps one of the most tangible outcomes of network music philosophy. The first large-scale ensemble to form was PLOrk (Princeton Laptop Orchestra), founded in 2005 by Dan Trueman and Perry Cook at Princeton University, soon followed by SLOrk (Stanford Laptop Orchestra), founded in 2008 by Ge Wang and Dan Trueman at Stanford's CCRMA. These ensembles, whose purpose was as much artistic as pedagogical, took the configuration of traditional orchestras—distributed instrumentation, collective timing, and sectional organization—and adapted its structural logic to what promised to become the universal musical instrument: the laptop.

At a typical LOrk performance, musicians would be distributed across the stage as in a traditional orchestral setting, though in place of strings, woodwinds, brass, and percussion instruments, the focus of their interaction would be a laptop, perhaps coupled with a MIDI controller, biosensors, or a video game console interface. In the case of PLOrk—and a few others—each performer would often use a dedicated hemispherical speaker, usually DIYed from a repurposed IKEA salad bowl, to restore a sense of acoustic presence and individuality that is usually lost in conventional PA systems. Other laptop orchestras opt for more straightforward and unified setups: a stereo PA or a multichannel diffusion system.

In essence, what The Hub and other early network music ensembles defined as a creative practice, laptop orchestras transformed into a stable, institutionalized educational and artistic framework, as suitable for offbeat experimentation as for realizing particularly ambitious computer-music projects. Over the last two decades, the number of laptop orchestras has only increased—from CLOrk (Concordia Laptop Orchestra) to BiLE (Birmingham Laptop Ensemble), BSBLOrk (Brasília Laptop Orchestra), MLOrk (Milano Laptop Orchestra), and ELUNM (Laptop Ensemble of the National University of Music in Peru)—the LOrk is now a global phenomenon.

The creation of this new orchestra format also prompted composers to write new works. PLOrk's commissioned works included Pauline Oliveros' "Murphy Mixup" (2006), Brad Garton's "Idle Swamp" (2006), and Douglas Geers' "Sweep" (2008); SLOrk premiered "Mesetas" (2010) by Marisol Jiménez, and quite recently the first laptop opera The Furies: A Laptopera (2022), composed by Anne Hege.

In parallel to, and eventually fusing with, another distinct offshoot of network music took shape in the experimental computing communities of Europe and the UK in the early 2000s—live coding. It stemmed from a communal desire to make the very process of computer-music creation fully transparent. Thus, it requires a performer to write and modify code in real time to generate sound (and later visuals). Live coding exposes the thought process of the performer and aestheticizes the programming paradigm; and not unlike academic laptop orchestras, it embodies a model that is both educational and artistic.

The catalysts for live coding were open-source audio programming environments such as SuperCollider (1996) and ChucK (2003), which by design were well-suited for real-time interaction. Then, in 2004, a group of UK-based artists, composers, and programmers—Alex McLean, Nick Collins, and Adrian Ward—formalized and articulated live coding as a cultural movement by establishing TOPLAP (Temporary Organisation for the Promotion of Live Algorithm Programming), intentionally bringing together several different communities that had already formed around the practice. The number of live coding groups around the world began to grow, as did the palette of programming environments dedicated to the practice, including TidalCycles (2009), ixi lang (2005), Fluxus (2005), Impromptu / Extempore (2005 / 2010), and Sonic Pi (2013).

By the 2010s, the initially niche experimental practice had transformed into a vibrant community that seamlessly fused clubbing and geeking out, algorithm and rave—the algorave. Originating in the UK, the movement spread quickly across Europe, South America, and Asia, sparking the formation of hundreds of grassroots collectives around the world, as well as dedicated festivals and conferences. Algoraves typically take place in clubs, warehouses, or art spaces, where performers live-code dance-oriented rhythmic structures, projecting their code for the audience to see. The performances foreground process and spontaneity over lustrous polish—patterns and loops evolve, break, and reform in real time, in a manner that resembles both a DJ set and free improvisation. The movement marked a distinct cultural shift in electronic arts—where club and hacker cultures blended into a singular hybrid.

In the United States, live coding developed along somewhat different lines—less so, at least at first, in community-organized venues than within the academic environments of established music departments. The practice naturally fit into the already established paradigm of laptop orchestras, and several universities, including Stanford, Princeton, and McGill, integrated live coding into their ensembles and digital-arts curricula. Eventually, some overlap between the academic and self-organized algorave scenes began to emerge, with the spread of collectives across the country, including Livecode.NYC, AV CLUB SF, and livecode.boston.

And although it has not always been easy for the public to accept laptops as musical instruments, laptop orchestras, rogue live coding collectives, and, overall, a burst of active development of music-making software, and its users, solidified the powerfully creative potential of the machines. Laptop's ubiquity as a hub for music creation, not strictly an instrument but a whole studio, spilled into nearly all other domains of music creation, proving itself as a flexible, democratic, and affordable platform for all kinds of creative and functional work with sound. This ubiquity and universality of the tool, which now simultaneously could be used in nearly all life activities: work, social, artistic, bureaucratic, also removed some of the allure—as it was colloquially mentioned, after a day of office work one may not wish to stare at the screen, and picking up a guitar or plugging in a hardware synthesizer, precisely because of their limitations was a more preferred mode of interacting with music. That said, the growing power and convenience of portability of modern machines, with accompanying its continuous flow of inventive tools and features that are infinitely more costly and difficult in the analog realm, preserved the laptop's position as a universal musical instrument. Plus, with growing availability of embedded computing it is easier than ever to negotiate this relationship: one can start a process on a laptop, then conveniently migrate it to adapted physical interfaces.

Today, as artificial intelligence becomes such a dominant theme of this historical transition, it warrants a retrospective view of how the cybernetics-inspired approaches to machine agency and creativity map to the realization of modern AI. Over the decades, what began as experiments in feedback and distributed creative authorship evolved into vast ecosystems of creative code, telepresence, and machine collaboration. The feedback systems of the cybernetic art gave way to recursive datasets of machine learning. Yet while the infrastructures changed—from analog circuits to neural networks—the underlying questions of agency, authorship, and relationality remained.

Machine learning models arrived as the latest residents of this computational ecosystem, integrated into the same platforms that had proven so fertile for experimentation. Yet they brought a fundamentally different logic—one of prediction rather than emergence. This philosophical chasm, juxtaposing machines as collaborative agents and machines as predictive engines is where the story of cybernetic art, network music, and computational creativity in general stands today, and this calls for a closer examination. Which brings us, inevitably, to artificial intelligence.

IDM vs AIDM

I know, I know… here we go again with AI. If half a century ago artificial intelligence remained a distant dream—a radical idea, even a fantasy—today one can hardly go through a day without encountering this new presence in one form or another. It feels almost mundane to talk about it, yet it continues to reshape our lives in profound ways. New technologies have repeatedly brought radical change to the humanities, but this time the stakes appear even higher. It is important to underscore that artificial intelligence is unlike any technology that preceded it. Rather than being built and programmed in the traditional sense, it develops and evolves more akin to a brain—or another complex adaptive network. Even the architects of contemporary AI systems lack complete understanding or control over what unfolds within the hidden layers of these models, where correlations are formed and “decisions” emerge. Without a doubt, AI can do remarkable things, yet such power and opacity foster a legitimate sense of unease. For all its potential benefits, the disruption that evolving AI introduces spans the cultural, ecological, economic, and even existential.

Remaining within our own field, we might ask: what does this mean for music, art, and creativity? And what can the history of network music and cybernetic art offer in navigating these questions? What appears clear is that, in the familiar gap between idea and reality, the work and philosophy of the League of Automatic Music Composers, The Hub, and other frontier-dwelling artists rested on a distinctly different foundation from that of modern mainstream AI technologies. Although certain overlaps exist, these practices diverge into asymmetrical yet interrelated pathways, addressing fundamentally different inquiries. In one, intelligence and agency arise through open and closed feedback loops—through the continuous negotiation among humans, machines, and environments in co-regulation. In the other, they emerge through prediction and optimization of vast datasets, where systems evaluate probabilities to determine what is most likely to follow in a sequence of sounds, words, or images. In the former, agency and cooperation are entwined; authorship is distributed but still legible. In the latter, it becomes opaque and statistical, trading co-authorship for data curation and prompt engineering.

Artificial intelligence, within the paradigm of cybernetics, once presented the potential for—and encouraged—new shapes of music to emerge. By contrast, data-driven AI systems in their present form are generally tuned to produce only formulaic variations and mutations of existing mainstream music. Thus, fundamentally different conversations are required when we speak of George E. Lewis’s Voyager and that banjo deathcore R&B track we’ve all generated in an AI music app like Suno.

These days, when the discourse about the threat of AI to music is becoming overbearing, it becomes vital to delineate certain things—such as, for example, why we engage in music in the first place. Yet beyond the economic anxiety and technological fatigue lies a deeper artistic question: how do we inhabit these new tools without surrendering what makes the act of musicking meaningful? Contrary to the widely detested statement of the Suno executive that “people don’t enjoy making music,” the opposite has always been true: we make music precisely because we love it. Surely, the economic aspect matters deeply to working artists and musicians, and the flood of generated songs will have a significant impact on the music industry—its economy, and perhaps even more crucially, our perception of it.

If the artificial intelligence experiments of pioneering composers once added value and novelty to music, today the presence of AI often raises mistrust and devaluation of media. This is because, for many of us, the story behind the music—the tenuous path by which one’s experience condenses and transforms into an emotional piece of art—remains deeply meaningful. It is the imperfections, the detours, the frictions, the subtleties that make it special and worthwhile. That process is fleeting and dynamic, shifting within us in accord—or discord—with the changing world.

Perhaps that’s where history still guides us. The pioneers of network music faced different but comparable uncertainties—and they responded not by retreating, but by experimenting, listening, and inventing new modes of collaboration. One of their guiding forces was a blunt acceptance that there was no financial incentive whatsoever in the kind of music they were pursuing—and with that came a sort of creative freedom to test whatever wild ideas they happened to have. Don't get me wrong, in the asymmetrical arrangement of the world, one doesn't always have a prerogative to create for free. Yet, it is also true that we often create not because of, but despite. It often happens that powerful art emerges under conditions that might otherwise seem forbidding. There is a different kind of force that propels us to create; we learn to change and adapt our approaches, making space for that flow. And in such cases, creating authentically and originally becomes particularly valuable.

Still, as with every technological shift, curiosity and opposition go hand in hand—to criticize AI's encroachment on creative labor is not necessarily to reject it altogether, but to challenge the terms of engagement. What new modes of listening are opening with AI? What novel dimensions of exploring music does it present? AI tools offer a great deal of exciting musical possibilities; I personally love and use projects like the Fluid Corpus Manipulation (FluCoMa) toolkit, Data Knot, RAVE, and a few others. These surely follow the trajectory of inquiry set in a cybernetic approach to music. Machine learning in this case, at least as I see it, stimulates and inspires human creativity rather than overpowers it. And I am sure many of our readers have their own inventive ways of inviting AI into their creative processes that I am not even aware of. I reckon, the difference lies not in rejecting technology, but in defining our rules of engagement—whether as consumers of its repetitive output or as active participants of the creative process.

Returning to Attali’s notion of music as a prophetic force, it seems that our AI-driven moment occupies a volatile threshold, a sort of a hybrid state that is part hyper-repetition, part simulated composition, part, as some may suggest, something not yet named. The scaled-up predictive machinery of commercial AI models pushes the industrial logic of repetition to its extreme—generating a seemingly endless stream of derivatives from humanity's creative archive. And yet, there are also gestures toward a more improvisatory, decentralized ethos that resembles Attali’s Composition, however stripped of the political stakes he envisioned. So maybe this is the real site of artistic agency today where one gets to decide how to work with, within or against these parameters. In so many words, what I mean to say is: let's keep making music for the love of it—and, perhaps occasionally, as a small act of rebellion, let's keep it genuinely weird.