Amory B. Lovins, CEO Rocky Mountain Institute: Twenty Hydrogen Myths

[bananas]

2. He's talking about the complete statement "p implies q", not just the first half "p"

where p = "Making hydrogen uses more energy than it yields"
and q = "it’s prohibitively inefficient".

From the pdf:

Myth #3. Making hydrogen uses more energy than it yields, so it’s prohibitively inefficient.

Any conversion from one form of energy to another consumes more useful energy than it yields.
If it could do the opposite, creating energy out of nothing, you could create a perpetual-motion
machine violating the laws of physics. Conversion losses are unavoidable; the issue is whether
they’re worth incurring. If they were intolerable as a matter of principle, as Myth #3 implies,
then we’d have to stop making gasoline from crude oil (~73–91% efficient from wellhead to retail
pump) and electricity from fossil fuel (~29–35% efficient from coal at the power plant to
retail meter). Such conversion losses are thus not specific to producing hydrogen. Hydrogen production
is typically about 72 to 85 percent efficient in natural-gas reformers or ~70–75% efficient
in electrolyzers; the rest is heat that may also be reusable. (These efficiency figures are all

<snip>

edit: remove footnote numbers from the excerpt.