This claim seems dubious, perhaps someone has more expertise to comment. The justification
> Operating this lock involves rotation at the joints by up to ∆θ≈1 rad. The model system analyzed in  is an excerpt of the links and joints shown in the closeup on the right of Figure 24. From [11, Eq. 2], this rotation dissipates bout 2.4×10−27J per rotary joint when operating at f = 100 MHz.
Seems akin to saying "We operate our microchip at 1 microvolt / 1 pico ampere at f=100MHz, giving 10-26J per operation." -- which seems like a silly aspiration without carefully analyzing noise and quantum mechanical constraints. (without which it would seem almost any computing device could operate at arbitrarily low power).
Things ignored to make that work:
Materials which are likely to ever exist.
A degree of mechanical perfection on the order of one of NASA’s fused quartz spheres.
An idealized means of transferring energy to the device.
A perfect vacuum.
Perfect shielding from external influence, especially vibration.
The faster you crank this up, the more energy is in the form of Ke in the device, and like a flywheel, it may come to resemble a bomb.
In fact, people probably use a lot of mechanical devices which can serve as logic gates without knowing it --- although everyone focuses on electronic digital computers, the mechanisms on which computers can be built are surprisingly vast and simple.
Also, the PDF is rather bloaty for the content, because all the figures are extremely high-resolution bitmap images instead of vector graphics, despite looking like they were created with a vector graphics program.
I wonder if the fact that the gates etc. are in a single plane could be used to make it function more effectively as well.
It certainly is using a different mechanical approach than the Differential Machine.
> In particular, a molecular version of this architecture could use stiff covalently-bonded nanotubes for the links and single bonds for the joints.
In that case joint friction would not be a consideration, or at least I would imagine things like electrostatic or molecular bonding forces would be larger.
This is discussed thoroughly in Nanosystems (Table of Contents - http://e-drexler.com/d/06/00/Nanosystems/toc.html)
It's turing complete. I'm wondering if the mechanism described in paper does something above and beyond...
Couldn't you simply connect the output of each holding lock to the respective input of the next cell and get the same results?
Merkle wrote a later paper on a different approach, buckling-spring logic: http://www.zyvex.com/nanotech/mechano.html I'm not sure how this latest design is supposed to be better still. I haven't really dug into either of them.
BTW this is the same Merkle who's always getting mentioned in blockchain articles.