Smaller innovations are probably worth protecting, but I would still give the edge to their execution rather than the pure idea.
I think Andrex is confusing the following:
"Damn it would be sweet if I could build a solar panel that extracted more energy from the sun"
with
"Damn, I have an idea how to build a solar panel that extracts more energy from the sun. We just need to put ____ with _____ and connect it to _____" and so on...
No one is trying to protect the first line with an NDA because everyone already has that idea. But why the hell wouldn't you want to keep people's mouth shut about the second idea?
He also using this argument to downplay the importance of good ideas:
"I'm a regularly guy and I had a great idea for a solar panel but I don't have the money or the technical experience to execute it."
vs.
"I'm an engineer at a major company, so I will build an awesome solar panel."
Yes, execution matters in this stilted comparison, largely because one person is actually in a situation where they can execute and the other isn't. Sure, execution in that situation matters a lot. But that discounts this situation:
"I'm an engineer at a major company and I have just been handed two ideas for a solar panel: one is really good and the other is terrible."
But, ah, I guess the idea doesn't matter very much as long as the execution is good, he can just pick whichever he wants and random and it'll turn out okay in the end.