I'd agree with much of this - in particular you don't know what question is being answered if you say 'do you like this', and people do love to vent on what they think other people would find useful or not. However there are stages in product development where it can be useful to ask if a respondent would find a new feature useful if other methods of insight aren't available. Sometimes a person will say, in response to that question, "no, because....". So whilst you can't necessarily take it at face value if someone says yes, the reasons they give why it won't (or will) be useful can be immensely useful to the researcher.