To the first point, I think the issue is that in any setting, it is impossible to avoid constructing some choice architecture. I actually think the recognition of this should reduce special interest control, since the government will be able to employ people who think about the implications of the choice architecture, rather than letting special interests set it.
Further, a nudge is not the same thing as an intervention. It is important that Libertarian Paternalism does not imply an expansion of interventions–just smarter construction of existing interventions.
To the second, a defining characteristic of Libertarian Paternalism is that the decision space is not changed in any way. Consumers/citizens can choose the exact same set of goods under a different choice architecture. (Ideally there are no costs to doing this, but I think this probably plays out differently depending on the specific example.)
In the policy areas they make recommendations on, the broad finding is that a majority of people don’t care or are not expert enough to figure out what is best for them. I see no problem with attempting to help those people make smarter decisions, as long as they are nudges and not shoves.
Given this, it seems that your problem is not with Nudge, but with the government. In my reading, the authors are agnostic about the appropriate level of intervention. If your complaint is against the government manipulation of markets, your fight should not be against efforts to make that manipulation smarter, but against the manipulation itself.
More to come.