> They’re essentially very tiny, very specialized atomic robots engineered by nature.<p>This is a tantalizing interpretation, but overall not helpful.<p>Terms like "machine" and "robot" bring with them a concept of design and order, which doesn't reflect the sheer randomness that dominates that level of physics. Compare this caption from the essay:<p>> Kinesin, gracefully strolling across a microtubule<p>which this molecular dynamics simulation of part of the same step - <a href="https://www.youtube.com/watch?v=JckOUrl3aes" rel="nofollow">https://www.youtube.com/watch?v=JckOUrl3aes</a><p>The title of that MD video is "Molecular Motor Kinesin Walks Like a Drunk Man" and in the description "Although many previous cartoons depict kinesin step as unidirectional and deterministic, hundreds of month-long computer simulations based on the law of physics show that kinesin undergoes bidirectional and stochastic thermal motion while stepping on the polar track microtubule."<p>That gives a very different feel about what's going on!<p>There are at least two other problems with using the 'atomic robot' metaphor. We know how to design robots, and for at least 25 years there's been the optimism that we could reuse that knowledge to design molecular robots. For example, Autodesk, the CAD company, looked into the idea in the early 1990s (<a href="http://www.fourmilab.ch/autofile/www/chapter2_82.html" rel="nofollow">http://www.fourmilab.ch/autofile/www/chapter2_82.html</a> ), and in the late 1990s there was the nanotech hype of designing all sorts of things in rigid-like diamond, rather than soft, flexible proteins. The robot metaphor really doesn't transfer over that well.<p>We also see creationists use the "robot" analogy to argue that since robots are designed by an intelligence, and proteins are robots, then proteins are designed by intelligence.