A walkable water mirror invites visitors to wander among lilies. Delicate strands of cable, rhythmically structuring the space, equip each flower with electronic components. Entangled in a dense network, these elements subtly modulate the lilies’ water supply in response to human movement: the footsteps of passersby send water waves that momentarily moisten the lily stems. Small scales continuously assess whether a lily has received water—if so, they trigger a brief lowering of its neighboring flowers, granting them the same access. Flickering numerical displays reflect the lilies’ shifting weight, echoing their gradual transformation.

Yet, the forest of copper filaments inevitably becomes the site of slow decay, a process that interaction can only postpone. For the semi-artificial neural network—emerging from the integration of the lilies—fundamentally differs from the architectures that underpin artificial intelligence in computer science. Unlike them, it carries a fate within itself. It is inseparably bound to the ephemeral life of the lilies it relies upon. Once the flowers—curiously reminiscent of the shape of a neuron—reach an irreversible stage, their motors stop, and their scales fall silent. The network gradually ebbs into stillness.