I think that the implications of Thaler's and others' work in behavioral economics goes deeper than what he and others admit. Experiments show that humans systematically deviate from rationality by being present-biased, prone to temptation, unfocused, and inordinately susceptible to social influence. Thaler points out that these are all things that can systematically stop us from achieving our goals.
They also raise serious questions, though, about whether humans can be said to have coherent goals at all. That's an important question because many of the judgments economists do or do not make on policies depend on the mid-20th century idea that one can help individuals satisfy their preferences but not make paternalistic judgments about those preferences or comparing different individuals. If humans don't have coherent goals (or don't necessarily have coherent goals), then any policy evaluation suddenly involves some degree of paternalism, and the line between policies that require moral judgments and those that do not evaporates.
Economist and blogger Tyler Cowen wrote an interesting piece on the use of nudges by conservatives, countering the predominant view that Thaler's ideas are roughly center-left. When it comes to the theoretical ideas undergirding economics, though, Thaler's findings suggest some quite radical changes.