OK I’m conflicted.

On one hand: hawt.

On the other hand, calling it a perpetual motion machine has triggered the engineer side of my brain and it’s kinda hard to concentrate when I’m considering feasibility.

Let’s start with the obvious: This is not a perpetual motion machine.

It is extracting energy from the nutrients contained in the material stream and converting it to mechanical force via intestinal smooth muscle contraction. If left running long enough it would deplete the energy contained in the material stream and then deplete the energy stored as glucose and fats in the processing nodes.

Now given the usually high energy extraction percentages for omnivorous mammals, I would expect a Linear Human Centipede (LHC) to be able to extract nearly all the available energy in the material stream with minimal number of nodes. Contrast a Cyclic Human Centipede (CHC), the LHC does not require a separate process for extracting depleted material. Simply put if you injected fresh nutrients inbetween the nodes you will eventually exceed the system material handling capability and cause a general system failure.

Because of energy extraction efficiency, even with only a handleful of nodes in a LHC we would likely see net energy loss in the tail node, so it would likely require that new node be added to the head, while the tail node is removed after it is inactive. Alternatively, you can cycle the rear node to the front after certain energy store metrics have been depleted.
One possibility for a CHC is that if the number of available nodes is small enough that multiple passthroughs are needed for the desired energy extraction. In which case material stream through put monitoring or even material stream energy measuring can be used to determine when to break the cycle and on a temporary basis, act as a LHC on already depleted material.

You see what this thing did! I came here to fap, not write an engineering paper!