Physics. You can't make light waves and photons smaller. You can make 'pancake' type lenses, but you give up maximum aperture and some of the optical quality gained by using larger front elements and more corrective elements that are needed with those larger front elements.

The only place getting the rear of the lens closer to the sensor helps in terms of lens size is with wider angle lenses that would need to use a retrofocus design on an SLR but can use a conventional design with the shorter registration distance (often called the flange focal distance) of the mirrorless camera.

A retrofocus design is almost like a telephoto lens turned around backwards. This is required if the lens needs to be a shorter focal length than the distance from the sensor to the optical center of the lens. The greater the difference in registration distance and focal length, the larger the retrofocus lens tends to be. For instance, a 35mm f/2 lens can be made smaller than a similarly designed 20mm f/2 lens for the same registration distance of 44mm.

With typical registration distances in the 40-46mm range, most SLRs must use retrofocus lenses for any focal length shorter than about 40-50mm.

With registration distances in the 16-22mm range, most mirrorless systems can use non-retrofocus designs down to about 20mm in focal length. Even for a lens with a focal length a bit shorter than a 20mm or so registration distance, the lens can still be more compact than a lens of the same focal length made for a 44mm or so registration distance.

Sensor size also plays a role. There are 50mm lenses available for mirrorless cameras with smaller sensors that are nowhere near the size of the one pictured on the FF mirrorless camera in the question. There are limits, though, to how much of an angle you want to allow light rays from the rear of the lens to strike the sensor.

Even with focal lengths at which lenses can be made smaller, ergonomics plays a role in the size of the camera body. We can only make a camera body so small before it becomes difficult to hold and control in our hands. The more direct control interfaces we wish to place on the camera, and the larger the LCD screen, the more this becomes an issue.