We are in 2048. Timo is a small artificially intelligent agent and the main character of Unexpected Scenery, a critical game by artist JooYoung Oh. The aim of the game is to help Timo get as much data as possible in order to train itself. The first stage introduces the mission. Overwhelmed by a familiar arcade aesthetic, we safely guide Timo, who grows and grows getting drunk on data until it reaches the next level. Here it discovers that in order to learn it must be continually compared with other data, day and night: this operation requires an intense, continuous human work. Timo, slightly more aware, wants to learn more and understand how this continuous connection is possible. The “wise rat”, a character from the game, is the guardian of the door to the next level and it suggests Timo goes down to the sea: the beginning of everything. But here we slip into a catastrophe: the submarine cables are engulfed by flames and sparks. They are the backbones on which the network runs, positioned at the bottom of the sea due to low temperatures, which, over time, have destroyed the ecosystem in a total short circuit of fire. Timo is disoriented: it was not ready to realize that its development is also the cause of this destruction. The latest level is in the depths of the ocean, the deepest layer of Timo’s deep learning process, inhabited only by giant squid, whose highly evolved biological neural networks have inspired important concepts in machine learning. They will be the key to the resolution of the whole learning trip. The path of the game draws a deep but refined learning process: the purpose of the game is not only to optimize efficiency and to let the agent swallow as much information as possible. The goal of depth here also seems to be in reflection and metacognition, so that Timo might also learn to make his own cognitive processes sustainable. Chiara Ciociola