I was a bit surprised in early biochemistry classes to learn of the low efficiency of human energy metabolism. I don’t remember the precise figure quoted, but the 33% cited in this website [link] was probably what I was taught. It needn’t be too concerning coming from a design perspective, as energy transitions, including in human-engineered systems, are seldom particularly efficient (my previous post makes reference to one example, where a particular transition was 50% efficient, not counting further losses that would occur downstream), but the details may be worth exploring in light of the argument of the link above. It turns out (as addressed in the site) that the 33% figure seems to be an underestimate, based on a misunderstanding of Gibb’s free energy, albeit an oft-quoted one. Something to look into in future anyway.
With some quick searching, I found this paper [link], which notes difficulties in specifying what is “optimal” in a thermodynamic system – it depends on the desired engineering outcomes. Intriguing!