Can welfare economics avoid paternalism after the behavioural turn?

In economics, choices, policies, and institutions are evaluated by how well they serve individual preferences. If everybody in the office prefers full fat milk, the welfare economist would recommend we get full fat rather than skim milk to stock the office fridge. This approach is given an anti-paternalist rationale. However, orthodox welfare economics faces a challenge from behavioural economics. My milk choices depend a lot on the default setting in the coffee machine, or on whether I have just been reminded by my surroundings of the need to live more healthily. Johanna Thoma asks, if we can’t derive a consistent preference from people’s choice behaviours, can we still live up to the anti-paternalist ideal?