Hinds holds a PhD in Organizational Science and Management from Carnegie Mellon and is the Rodney H. Adams Professor at Stanford, where she co-directs the Center on Work, Technology, and Organization and is a senior editor of Organization Science. Most of her research is on distributed and cross-boundary teams (culture, language, and conflict across national borders), and more recently on human-robot teams.
The work cited here is earlier and narrower, and it is the evidence the page’s least intuitive claim rests on. In “The Curse of Expertise” (1999) she ran two studies, one comparing experts, intermediates and novices, one manipulating expertise experimentally, and found that among those with task experience the more expert the predictor, the worse their estimate of how long a novice would take, and that standard debiasing techniques did not fix it. The relationship is not a straight line: intermediates were the most accurate of all, more accurate than novices as well as experts. That is the evidence under the instruction to expect an expert to be bad at explaining their craft to you, and it is why the page tells you to plan the questions rather than trust the conversation to find its own way.