Great post. Clearly energy available constrains the production of nearly everything, including metals.

A general observation - if a "universal mining machine" is ever developed, it will probably be used first to mine the Moon (or asteroids). Free access to unlimited solar power (for up to two weeks out of every month at least), and free access to a vacuum (needed to separate atoms efficiently). Lunar produced helium-3 for fusion reactors has been suggested as an alternative energy source. Agree that there are probably no ore deposits on the Moon to speak of, unless you count metallic broken pieces of meteorites, or metallic ores possibly concentrated at the base of thick lava flows. Impacts have homogenized everything else (free comminution). John Lewis (University of Arizona) has argued in a couple of books in favor of mining asteroids instead, mainly owing to the value of platinum-group elements alloyed with the native iron-nickel, the presence of water and other volatiles, and freedom from gravity constraints on transport. Very high (perhaps far too high) initial investments will be required for any type space mining operations, however, and it's unclear that anyone could make a profit on them.

Minor quibble with your statement about the sea floor: Diamonds and even nuggets of tin oxide have long been recovered from fairly shallow water at very low energy cost (many from illegal dredging operations - basically a pump at the end of a very long hose - a type of placer mining). The main constraint on sea floor mining is the huge amount of turbidity pollution produced, and the legal problem that most of the sea floor currently belongs to no single country. Metallic mineral sulfide deposits are continually being created by sea floor hot springs ("black smokers") at spreading centers (mid-ocean ridges) and are being destroyed as sea floor is subducted at trenches. Other metals are present in manganese oxide nodules at the surface. I suspect that, if the need becomes acute and industrial civilization continues its present course, despite PO, the 70% of the planet that is sea floor will be the next metal prospecting frontier - way ahead of outer space. By then little marine life may survive anyway. More likely even less would survive such a mining episode.

Another quibble. As Deffeyes (2005) discusses in his book, the 1970's assertion by Brian Skinner of a concentration valley between high-grade ores and geochemical background levels of metals has yet to be demonstrated scientifically, AFAIK, and Deffeyes makes a pretty good case that it doesn't hold for uranium in particular. Whether or not such a "valley" exists has no effect on your basic argument - that the energy costs of extraction eventually get prohibitive as ore grade goes down, implying that eventually metal production must peak unless energy continues to get cheaper OR if the metal is needed "at any price". Uranium is probably such a metal - the cost is at present a tiny fraction of the cost of building a new power plant, and could therefore increase 10X or even 100X without affecting the basic economics much, especially if fuel reprocessing was used. Indium for solar cells might be another "at almost any price" metal if no better technology is developed.

The basic problem of economists is that they think of petroleum as metal ore, with a very long "tail" on concentrations in nature (with or without bumps). As Campbell and Deffeyes discuss at length in their books, unlike metals, petroleum is destroyed both at depth and at the surface ("petroleum window"), and there was only one way to produce and trap it, during very limited periods of Earth history, under very, very special conditions. When it's gone, it's gone. Coal likewise. In contrast, metals like copper, for example, are never destroyed, only redistributed, and there are many, many different geological ways to concentrate and preserve them as ores in nature. As you and others have noted, in the event of a collapse, the metal miners of the future may be digging mainly in old landfills.