Suno v3 text-to-music generation (March 6th 2024). Nothing short of astonishing.

Listen to the evidence: "Suno v3 (Alpha) w. ReRites Lyrics" playlist.

To get a sense of the extraordinary evocative range of Suno v3 Alpha, please listen as you read. 

Playlist: Suno v3 (Alpha) w. ReRites Lyrics (Created: March 6, 2024) 

The 39 songs in the playlist (many brief, some unfinished) were all made on 26 March 2024. And because Suno has a very explicit copyright control policy, all lyrics are from my ReRites human+AI poetry paperback (Anteism, 2019). 

My conception of what AI-generated music is capable of has shifted substantially. 

Copyright Infringement & Deterioration in Taste

Here’s composer John de Sousa’s 1906 reaction to audio recording. It’s remarkable how it replicates the anti-AI controversies now: 

“SWEEPING across the country with the speed of a transient fashion in slang or Panama hats, political war cries or popular novels, comes now the mechanical device to sing for us a song or play for us a piano, in substitute for human skill, intelligence, and soul. …  I foresee a marked deterioration in American music and musical taste, an interruption in the musical development of the country, and a host of other injuries to music in its artistic manifestations, by virtue -- or rather by vice -- of the multiplication of the various music-reproducing machines. When I add to this that I myself and every other popular composer are victims of a serious infringement on our clear moral rights in our own work” -- The Menace of Mechanical Music, John Philip Sousa, Appleton's Magazine, Vol. 8 (1906), pp. 278-284.

Did audio recording destroy human music? Did photography destroy art? Did writing destroy memory as Plato famously thought?

Make Me Care

If you are not listening to the playlist and are an AI skeptic, I challenge you to listen to the final track on the playlist: Make Me Care. (Listen to all of it!) 

Cliches & Fringes

Suno replicates a diverse variety of the central cliche hook audio-worm tropes of the music industry (pop, edm, etc). But Suno v3 also is capable of exploring fringe genres. It replicates emotive dynamics, vocal cadences, song structures, melodic inflections, drop beats and hybridized forms. 

It can move from sensitive ballad to epic post-rock, syncopated candy, roots edgy raunch, and virtuosic elegiac transitions. 

In short, this is audio generation that inhabits a stunning uncanny valley terrain. Albeit for now with tinges of low-res grain and moments of inarticulate yelping, but grandeur, excitement and visceral joy may arise. Pop radio beware. 

Free Joy & Gentleness

Imagine a future where the joyousness of music is leaking from every moment and is available to all freely infinitely as consolation, stimulus and profound biometric-linked solace. Musical medicine. Musical learning, yearning, growth, collaborations. 

This could either decimate the traditional music industry or as Alpha Go provoked the Go community to develop totally new innovative group strategies and moves (the game of Go has changed radically), perhaps AI text-to-music will allow every child to play along with AI composers; perhaps music will expand rapidly into new forms and hybridized modulations. Perhaps the back-to-analog DIY play-music-yourself community momentum will accelerate, mutating into an inconceivable symphony of new forms. 

Just as walking did not disappear when cars appeared, the AI-music epoch might prove as paradoxical as the introduction of processed food into diets, causing an epidemic of audio-diabetes, yet also evoking its opposite: resonant organic food, audio-gym pilates, live-AI crowd-controlled concerts and collective listen-making sessions.

Enjoy the reverberations of this unprecedented resonant era! And huge respects to the Suno team.

~

Postscript-update: the next day (as if hungover from the creative flurry of last week), while generating new results, what was once astonishing, now seem to fit within proscribed banal boundaries, — the model refuses to produce anything beyond its base foundation pattern. Staleness reigns. The perimeters of the known cage potentials. And the grit and grain of the sound seem to resonate with neglect. Perhaps AI music will induce many to learn instruments and gather again to sing, as bodies. Perhaps this is simply how the creative process proceeds: fluctuating between inspiration and inertia, even with algorithmic augmentation.