The work is done with a combination of in-time rendering motion graphics and projection mapping. It sends sky pictures shot at a different period to models for learning and calculation through deep machine learning. After four months, it finally produces the sky image above the lake by applying machine learning. With the application of those models which have been trained, it produces in-time images based on the lake sculpture. The wind speed and direction control the trigger at the lakeside of Vallée de Joux. The constant change of data makes the images change slowly, then subtle fluctuation is added by the fluid models, and finally, it presents the image by water projection. In the real and subtle shadow changes, the audience can sense the supernatural power of Vallée de Joux's field domain within the work. When you overlook the device that makes you seem at the top of Vallée de Joux, what you see is a real illusion and an illusionary reality; it is a simplified presentation and a progressive guide from the surface to depth. Through in-time connection, it continues to explore the unknown in the known.