@akhilrao i don't think it's a great take. we (in aggregate) want productivity improvements, rather than a high price to keep labor bid into lower productivity work. people in high productivity gigs (according to cost disease assumptions, if not experienced reality) get paid well — that's the reason you have to bid high to keep people in low productivity gigs. it'd be better if everyone was just paid well for high productivity rather than for low, ceteris paribus. 1/

@akhilrao all that said, there are lots of circumstances where "cost disease" shouldn't be taken as a disease at all. for example, a teacher of a small class is, superficially, a cost disease example: holding class sizes and learning pace roughly constant, productivity seems not to have increased. but we would not want productivity in that sense to increase, because the actual value comes from a not-scalable human connection, not anything a MOOC can provide. 2/

@akhilrao the better way of thinking about this kind of case is that there's no cost disease because productivity HAS increased. the value of the kind of education that can only come from direct human relationships has increased as the economy has grown more specialized, complex, and prosperous. so even though the goods delivered are precisely the same, it's not really a cost disease, just maintaining a similar share of value provided. /fin

@akhilrao (epilogue: i think a key question is, would we WANT "productivity" in the tangible-outputs-to-inputs sense to increase in a sector. in the infamous string quartet example, the answer is "no" — by assumption, people don't want recorded music, but an actual live quartet. then it's not really a cost disease, is it? we are perfectly capable playing a streamed recording, but instead we pay up for the quartet. we are valuing the quartet more highly, then, not suffering from cost disease. 1/

@akhilrao contrast that with a construction worker framing a new home with a nail gun. if five workers could be replaced by one person and five robots for half the cost, no intangible value would persuade us to keep the five workers, as long as those workers could find other remunerative things to do. so the construction worker's rising wages for the same framing represents cost disease, while the iconic string quartet example actually does not.) /fin