A problem with that metaphor is that a machine learning paper refers to an objective reality. If a philosopher came along and said 'I'm not sure how this method

A problem with that metaphor is that a machine learning paper refers to an objective reality. If a philosopher came along and said 'I'm not sure how this methodology works, but it is clearly optimized for the wrong metrics as evidenced by your sales since you implemented it', the data scientist should know to listen to the philosopher no matter how he came to that conclusion. The equivalent in philosophy is, unfortunately, very different for someone deeply engaged with philosophy and an educated layperson. To a layperson, the measure of philosophy is something like the moral and mental well-being of the philosopher. That is why the laity cares and most would say should care about philosophy. Unfortunately, this is not the metric used by professional philosophers. Social validation may be a more rigorous heuristic than moral well-being, but that does not make it more relevant to what makes philosophy itself relevant. It seems likely to me that the reason laypeople so consistently criticize more educated philosophers is that philosophy optimized for social validation is not optimized for moral well-being. That is an alarm bell that should matter to the intelligent philosopher, much as a 'but sales are down' warning should matter to the intelligent data scientist.