Utility fog is a concept conceived by nanotechnologist Dr. J. Storrs Hall in his search for a replacement for the seatbelt. Rather than a belt, he envisioned a cloud of tiny machines with interlocking arms securing him in case of a collision. Of course, this concept can be radically extended -- if we had a swarm of reprogrammable airborne robots, we could direct them to simulate a wide range of objects and perform a wide range of functions. Hence the name utility fog.
Since its conception, utility fog has been the subject of a dozen or so technical papers. Most designs assume "foglets" about 100 micrometers in diameter, with about arms extending in every direction. Manufacturing these foglets would likely require molecular manufacturing, that is, a form of manufacturing capable of designing products to atomic precision, although a utility fog fabricated using advanced microtechnology techniques is also imaginable.
In "idle mode", the foglets would float in the air without connecting their arms. Their density would be low enough to avoid decreasing visibility or impeding human movement in any way. The foglets would be completely undetectable until they condensed into "active mode".
Because foglets would likely be made out of covalently bonded carbon, or diamond, they could be very strong, capable of manifesting as such durable objects as furniture. They could then disperse if they were no longer needed. Eventually, entire cities might be made out of utility fog. In such a city, persistent structures might be a thing of the past, with buildings morphing to conform to the immediate needs of the inhabitants.
A human being suspended in midair by utility fog would be able to fly without any other aerospace equipment, giving the illusion of independent human flight. A large wall of utility fog could simulate a wide range of backgrounds by changing surface color or optical properties accordingly. Although utility fog sounds like a fantastic, far-future technology, it could truly be feasible in the next couple of decades, but would likely require molecular manufacturing as a prerequisite.