My own view is that the ultimate scarcity in today's system comes from what the political economy of our societies and polities can bearArbitrary production constraints on our time, are increasingly at variance with the growth and output levels needed for general equilibrium stability. So long as human capital potential remains left on the sidelines, non tradable sectors will continue to make excessive demands on general equilibrium, which only further displace economic participation at the margins.
If citizens were free to explore their full economic potential, less exponential tradable sector growth would be necessary, to maintain production and consumption levels in today's dominant services sectors. Even though Georgescu-Roegen was concerned about entropy and limits to growth, time value is mostly squandered when citizens have too few meaningful means of economic engagement with others. Greater freedom of economic participation, would allow more inputs to flow into the system as a whole. This matters, given the fact entropy becomes more likely when systems lack sufficient inputs to balance extensive demands on their output.
What of the concerns Henry George held, regarding land as fixed scarcity? Fortunately, land doesn't hold the same relevance as a fixed factor, as it did in the past. However, one reason land appears excessively valued or scarce, is due to the centralized nature of productive agglomeration, whereby important forms of (land correlated) knowledge use are artificially scarce. Productive agglomeration could be greatly expanded, if new patterns for knowledge and mutual coordination were adapted for use in decentralized settings. After all: Via greater possibilities for productive agglomeration, the land which qualifies as meaningful economic access, would become less scarce as well.
The constraints of daily life in the 19th century, likely made it difficult for individuals such as Edward Gibbon Wakefield to imagine the possibility of knowledge dispersal in smaller communities and via the coordination of multiple skill levels. Even though he envisioned humans as the ultimate scarcity, he seemed to dismiss a wide range of human capital potential if it were not also well reimbursed on monetary terms.
Alas, others lack this appreciation as well. Consequently, it is becoming an increasing struggle for groups with widely divergent skills sets to live in close proximity to one another, given today's inflexible divisions of labour which are physically separating the most basic of human functions. Given the understandable desire of individuals to live among others who hold similar values and skill capacity, more groups might eventually resolve this coordination problem by sharing both their higher workplace aspirations and lower valued skill sets. By internally coordinating a full range of skill sets, groups with a similar income range would also share similar levels of responsibility for the local infrastructure and other local amenities they find most desirable.
For that matter, today's digital realm makes it all the more compelling, for societies to begin sharing better integrated skill sets wherever people work and live. In particular, automation poses its own challenges to societal expectations of divisions of labour and specialization, especially since the IQ "losers" often have to settle for work which makes it difficult to meet their personal responsibilities. One reason computer deep learning may be unsettling for professionals, is that the same dismissal these groups tended to assign to "deplorables", is the same dismissal that automation could impose on them, should the arguments of efficiency and rationality be carried even further to "cold" logical conclusions.
Perhaps the main problem for Robert Solow's framework of complete interchangeability, is that we increasingly struggle to substitute the fixed scarcity of time value, for other resource capacity in services dominant economies. Granted, it was once a simpler matter to substitute time value with other factors. Given that tradable sector dominance assigned extensive production roles for time value for so long, perhaps it is understandable, how Solow and many others were convinced this approach was valid.