The persistent allure of generative music lies in its promise of perpetual mutation—a ceaseless flow of sound that, like river-washed stones, reveals unexpected forms through repetition and difference. The most elementary technique in this respect is not merely the loop, but the cultivated introduction of stochasticity: a weighted randomness woven into the fabric of code. In Brian Eno’s seminal experiments with Koan Pro, for instance, even a gentle swing in probability—say, shifting a MIDI note’s chance to sound from 40% to 55%—can open deltas of possibility, letting motifs emerge and recede on what feels like tidal whims rather than strict grids.

This generative approach is not about ceding control to the computer, but about shaping the riverbed through which the current will run. By modulating parameters such as event density or note direction within algorithms—whether inside Max/MSP, SuperCollider, or an Electronica sequencer—one composes with conditions, not outcomes. Avoidance of cyclic predictability is achieved by inventing systems prone to eddies: weighted Markov chains, irregular LFO modulations, quietly accumulating delays. The music flows, at once indifferent and responsive, treacherous yet oddly companionable.

Listening to these streams, awareness lingers on recurrence tinged with difference. The human pursuit is not mastery, but attunement: the slow realisation that surprise and structure need not be enemies. There is a humility to generative composition, an acceptance that sounds will reroute themselves, just as rising water finds new channels. This humility, I suspect, lends generative music its uncanny power—a reflection of a world where everything is subject both to accidental drift and to careful human shaping.