[0]Known, boring technology with 99.9% of the engineering warts and technical debt already solved.
The petrol in the ground is free, sunlight is free, biodiversity is free. The cost is only a human factor.
Higher efficiency on the other hand usually means lower resource consumption.
A weak proxy for cost ? Or a strong one for survival ?
A cheaper steel bike is less efficient than an expensive carbon fiber bike when your goal is to get to the mountain top before the flood submerges the valley.
Arguing that, in the judgement of process for the conversion of one form of energy (sunlight) to another (hydrogen), thermodynamic efficiency, or the measure of how well a process converts one form of energy to another isn't relevant is somewhere between obnoxiously pedantic and wishful thinking.
Funnily enough I read the article after your comment, and it is even less useful, interesting, than my initial assessment (not even direct solar to H2). It has approximately the novelty and utility of porting DOOM to <insert processor here> and running it in a refrigerator.
If you need the end product at a particular location, the transport costs can overwhelm any differences in efficiency.
The same also applies if you have excess energy in a particular location that you want to use.
In either case, the transport costs can make using the less efficient process, the more energy efficient choice overall.