AbstractConceptual structure of contemporary thermal physics is by no means simple, understandable and free of contradictions or ambiguities. As many students claim, it is a rather complicated doctrine which is far from to be easy comprehensible (Meltzer in Am J Phys 72:1432–1446, 2004). Therefore, in this polemic contribution, originally the summer school lecture, we will address some lesser-known points, which are likely responsible for such a situation. We start with Black’s path breaking discovery of two aspects of heat, namely the “matter of heat” and the “intensity of heat,” which were later identified with the quantities known as entropy (S) and temperature (T). It can be shown that the product of these two quantities can enter the energy balance equation, which is apt to interconnect various parts of physics. The genesis of these key concepts was, however, not straightforward. An original hypothetical model of heat, a subtle imponderable fluid, caloric (ς), was abandoned after the non-critical acceptance of the “milestone of thermodynamics,” Mayer–Joule’s postulate claiming the equivalence of heat and work. Heat then became a pure energy without any material carrier, but with seriously limited transformation abilities. In addition, the situation is also complicated by the very fact that the definition of Kelvin absolute thermodynamic temperature scale (T) is fully based on the caloric theory. Partial reconciliation brought about only introduction of Clausius’ entropy. This, however, was afterward recognized to be practically identical with the Carnot’s caloric. Furthermore, by analogy identification of phenomenological variables S and T with statistical parameters appearing in the kinetic theory has not withstand extension to the domain of special relativity. Thus, the full equivalence between thermodynamics and statistical physics is not feasible. Another realm of problems is generated by the recent tendency in metrology to define units by means of defining constants with fixed numerical values instead of materialized étalons. In case of thermal physics, one can be skeptic about such an approach, because the universal constancy of entropy and its unit, Boltzmann constant k = 1.380649 × .10−23 J K-1, are only unjustified assumptions, which can be a potential source of future difficulties.