The ultimate building blocks of matter are atoms and molecules. If we can control these we can truly build from the bottom up, for example, a computer made of atomic-scale components that would fit into the palm of your hand yet be more powerful than today's supercomputers. In 1986 the Nobel prize was won for the invention of a microscope that can image individual atoms and molecules, the scanning tunnelling microscope (STM). But more than that, in 1989 that microscope was used to assemble a man-made structure atom-by-atom with atomic precision; a 35 Xenon atom advert for IBM. Since then many great strides have been taken in the quest for ultimate atomic control, including breaking and making individual chemical bonds, constructing a single molecule transistor and even constructing logic gates out of carbon-monoxide molecules. So why, in the intervening 22 years, has such atomic scale engineering not become common place in modern technology? There are of course many technical challenges to working on this atomic level, for example, at the atomic scale everything sticks to everything and even at room temperature your IBM advert will boil off into space. But perhaps the most challenging limitation is not due to a fundamental problem with the quantum physics that governs atoms and molecules, but is instead the construction process itself: all these exquisite structures were built one atom at a time. That is quite a manufacturing bottleneck! This experimental proposal aims to explore a new way of controlling multiple atoms and molecules with the scanning tunnelling microscope - nonlocal atomic manipulation. Instead of manipulating only the atom that is directly in the microscope's sights, leading to the one-atom-at-a-time limit, here we'll spread the effect of the microscope (specifically its injected electric current) across a surface over distances of 10's of manometers. This nonlocal process allows thousands of individual molecules to be manipulated simultaneously. Many critical questions remain to be answered if this new mode of manipulation is to have any promise of constructing extended structures with atomic precision: what roles do the molecules and the surface play in the nonlocal process? Is there a difference to what happens for a molecule directly under the microscope (as in conventional atomic manipulation) and a molecule some distance remote? How is the electrical current transported from microscope to distant target molecules? And how general a process is this? By answering these questions we'll be closer to transforming atomic manipulation from an elegant laboratory technique to a manufacturing tool for creating (relatively!) large scale but atomically precise structures.