Hands, in their ability to touch — fingertips serving as input data points to the brain — feed a constant stream of data to the brain that is enriched per the homunculus’s illustration. Not only are they sensitive, but they are the most adaptable limb, quick to contort or reposition themselves as reflexes or deliberate actions call for. They are particularly useful in novel situations. In art-making, they study tangible landscapes or tools that can be manipulated, reshaped after feedback from the brain. They are part of a feedback loop, cracial (arguably necessary) to the creative process. Without hands, it is very difficult to transform the mind’s abstract thought to a concrete reality. Regardless of the medium. This point is also why a computer is inherently a tool for the mind, as it deals with abstract concepts and reshaped interactivity that doesn’t necessarily require hands to explore the unknown.
I think a great example of McCullough’s argument is TiltBrush, a 3D drawing environment within virtual reality. The computer, in this situation, is the software that in a literal, warped reality, allows you to execute otherwise traditional tasks (sculpting, drawing) with your hands via the controllers. Feedback from how the “ink” dispenses from the controller to paint/sculpt your work of art is rendered before you in this reality, allowing you to reprocess and alter your actions to either explore things differently or repeat.