The Decadent Society is a recent book by New York Times columnist Ross Douthat about “how our turbulent age is defined by dark forces seemingly beyond our control”. Amazon recommended it to me after I finished reading “Cult of the Dead Cow”, a similarly antiestablishmentarian book with technocratic overtones. I highly encourage anyone who finds this review interesting to pick up a copy of the actual book on Amazon or elsewhere, as it is much more dense than I could ever capture here.

By way of preface, here is an excerpt from Douthat’s op-ed advertising the book:

The truth of the first decades of the 21st century, a truth that helped give us the Trump presidency but will still be an important truth when he is gone, is that we probably aren’t entering a 1930-style crisis for Western liberalism or hurtling forward toward transhumanism or extinction. Instead, we are aging, comfortable and stuck, cut off from the past and no longer optimistic about the future, spurning both memory and ambition while we await some saving innovation or revelation, growing old unhappily together in the light of tiny screens.



Part 1: The Four Horsemen





Part 2: Sustainable Decadence

Comfortably Numb

A Kindly Despotism

Waiting for the Barbarians

Giving Decadence Its Due

Part 3: The Death of Decadence




My Thoughts

While looking for other reviews, I found this one, which suggested checking out Peter Zeihan’s Disunited Nations. I found the Youtube video to be entertaining and informative. According to the comments he delivers a similar presentation whenever he releases a new book, so at least he’s are consistent.

I am personally a big fan of Peter Thiel’s diagnosis of the problems facing innovation today, much of which is parroted in these sections. To paraphrase a section from Zero to One, Americans are “positive indefinite”; we have a general sense of positivity about the world, but no clear goals to accomplish, so our media ecosystem rewards the appearance of political action without anything fundamentally happening. Reading through this book was the first time that I had seen the innovation glut as a consequence of policy created by an aging society that is naturally more resistant to change.


I found the portions about algorithmically-generated recommendations hurting the diversity of content to be an interesting technical question. Recommender systems are a pretty interesting area of machine learning. The simplest recommender system would just figure out the most popular movie and suggest it to everyone, since without knowing anything about you that would be the safest bet. In machine learning terms, this is a bias. I think this particular critique basically boils down to saying that algorithmic recommender systems don’t work very well, but I suspect that as more modern machine learning approaches are applied to these problems and more data is collected about individual users and tastes, these recommender systems will exhibit less bias.

It’s important to note that there is a lag time before this takes effect, because the ecosystem takes time to adapt. For example, when I first started looking for video game streams on Facebook, all the top recommendations were for Fortnite and Call of Duty, which I’ve never really played, simply because the machine learning backend hadn’t been built yet. As it improves, it will attract more streamers for a more diverse set of games, and users will learn to look for content they like from those outlets, encouraging this symbiotic relationship and fostering more creative approaches.

Algorithmic classification problems can basically be distilled down to three components: some features to use as input, like a user’s age, gender, interests, time spent using a computer, past interactions; a signal that measures something that you’re trying to optimize, like the likelihood of someone clicking on an advertisement or the amount of time they spend watching a video; and a model that predicts the signal from the features, which can be used to make decisions, for example, about which ads or videos to show someone.

Advertising has historically optimized for clicks because it is the easiest thing to optimize for, but I think the world is moving towards something more substantive, partly thanks to the massive surveillance state we’ve constructed, and partly thanks to the big gains that the AI world has seen in the last couple years. Facebook and Google have basically saturated the advertising market for large companies and are starting to try to onboard small and medium sized businesses. I think this will be a positive direction for people - we have the potential to use the massively effective advertising tools at our disposal to generate something akin to community engagement. I’ve wondered in the past whether or not advertising platforms could be used by well-meaning individuals to broadcast socially helpful messages, similar to public service announcements, although I suspect this would be a pretty tall order for our current government.


The section about counterculturalism reminded me of David Foster Wallace’s critiques of irony as tearing apart social fabric. Instead of proposing new ideas, it is simpler and oftentimes viewed more favorably to critique existing ideas, masking the weaknesses of your own argument behind layers of sarcasm and irony. If my girlfriend is any indication, this is a very unhealthy behavior on both an individual and social level. I think this parallels (in different language) some of the discussion about the sclerotic political system.


By odd coincidence, this book was released right in the middle of the global Covid-19 pandemic. My girlfriend, who is currently a 3rd year medical student, has been talking extensively about the effects that the pandemic is having on our medical system. The US healthcare system fits many of Douthat’s definitions of “decadent”; it is bureaucratically bloated; expensive; monopolistic; risk-averse; patched together by kludge; resistent to any attempts at change.

There have been (unsubstantiated) reports on the medical school subreddits that my girlfriend frequents of nurses in New York (where the pandemic is hitting hardest) being paid $200 an hour, while medical students are being graduated early if they volunteer to work in an unpaid capacity as sub-interns. Other reports have suggested that medical students and residents, who don’t benefit from the strong unions that nurses have, are being used as meat shields for attendings, being sent into patient rooms without proper protective equiptment.

The American medical system is a perfect example of a system co-opted by (mostly) well-intentioned people that has resulted in an absolute mess. Salaries for specialists can be upwards of $1 million a year, while the number of residency positions for those specialists is kept artifically low by the American Medical Association. For a resident who does manage to make it into such a position, the threat of $500k in loans without any transferable skills deters many from unionizing or dropping out of even the most demeaning residency positions. Lack of public support for strikes when patient well-being is on the line also hurts organized labor efforts.

In some ways, the fact that the medical system is being hit first by this pandemic means that it’s a canary for the economic and governmental institutions that are likely to feel more delayed effects. While it is sustainable under ordinary circumstances, it seems that the highly unusual circumstances of a global pandemic may expose its many cracks.


Overall, this book does a good job of building a comprehensive narrative around Douthat’s vast cultural knowledge, although it is relatively sparse on details and some of the supporting evidence doesn’t hold up under scrutiny. Nonetheless, I think the message is urgently important, especially for those working in highly technical fields.