By now, the front-end writers for Science know just how to calibrate the correct balance of honesty and sarcasm in discussing the idea of assembler-based nanotechnology. Here Robert Service has enlisted Richard Smalley to do the dirty work:

Richard Smalley, a Nobel Prize-winning chemist at Rice University in Houston, Texas, says that there are several good reasons to believe that nanomachines of the sort imagined by Drexler and company can never be made. "To put it bluntly, I think it's impossible," Smalley says.

Well, Smalley sat on a panel with Ralph Merkle in a Congressional hearing in June of last year, and said nothing about MNT being "impossible".

As he sees it, the idea of little machines that grab atoms and assemble them into desired arrangements suffers from three faults. First, he says, it's wrong to think you can just manipulate an individual atom without handling the ones around it as well.

Never mind that no serious MNT theorist pretends that you could.

"The essence of chemistry is missing here. Chemistry is not just sticking one atom in one place and then going and grabbing another. Chemistry is the concerted motion of at least 10 atoms."

I suppose this explains why no diatomic molecules exist.

That means to move that one atom where you want it, you'll need 10 nanosized appendages to handle it along with all of its neighbors.

Of course any design for an assembler system will have to take account of the motions of atoms in the vicinity of the reacting unit (which need not, and in many cases probably would not, be a single atom, but a group, possibly even a rather large molecular unit). But this does not imply that the reaction dynamics will be impossibly complicated to predict and control. Most of the motion will be confined to one or a few degrees of freedom involving only a few atoms. The motions of surrounding atoms need not always be completely controlled, but only constrained within certain limits.

Which raises the second problem–what Smalley calls the "fat fingers" problem. A nanometer is just the width of eight oxygen atoms. So even if you're trying to build something hundreds of nanometers in size, "there's just not enough room" in that space to fit those 10 fingers along with everything they are trying to manipulate.

First of all, "fingers" is a naive and incorrect image. Of course you don't have little fingers just picking up atoms and placing them one at a time. What you have is an assembler arm which determines position, angle, and applied forces, torques, voltages, etc. The end of the arm is a binding site which binds reversibly to a variety of "tool molecules" which carry the units to be added to the workpiece-in-progress. When the reaction is completed, and the unit has been transferred from the tool to the workpiece, the remainder of the tool is docked in its proper recycling station and unbound from the arm.

Of course the girth and extent of the arm will constrain the types of structures that can be assembled. Many chemically stable structures may prove inaccessible by direct assembly, but other tricks involving "robotic" actuators as well as self-assembly and traditional chemical synthesis pathways may make some such structures accessible as subunits to be assembled into larger systems. The fact that not all structures may be directly accessible does not prove that assemblers cannot "close the loop" of self-replication and serve as the basis for a versatile manufacturing system.

Finally, there's the "sticky fingers" problem: Even if you could wedge all those little claspers in there with their atomic cargo, you'd have to get them to release those atoms on command.

Again, it's the wrong image. Anyway, we do know how to reversibly bind and unbind. STM experiments have already demonstrated that this can be done repeatedly in reactions with single atoms. Again, this is a cheap shot, lazy and dishonest. But I wonder if Richard Smalley still makes these arguments, or if Service was only quoting some statements he may have made several years ago.

"My advice is, don't worry about self-replicating nanobots," says Smalley. "It's not real now and will never be in the future."

What, me worry?