The first time that I heard about the Coronavirus was sometime in mid January. It was of course mostly from friends and acquittances within the tech industry. There are lot of theories about why the tech industry was the first to react to the virus outside of China:
They always kept a close eye on developments in China and never underestimated Chinese infrastructure
They have been regularly exposed to how network effects can help something take off - usually a product, this time a virus
More people within the tech industry have characteristics that make them suitable to notice this - hyperaware, always online nerds
While I share some of these characteristics, I don’t think I’m on the extreme end of the spectrum on any of these - I’m a little more aware of China than the average population, a little too anxious and have observed second hand how network effects work.
Most of my thinking around this was driven by proxies that I trust - people like Balaji Srinivasan who were really early on this. And I think this may have been the case for a lot of people similar to me - somebody we respected or looked upto was early to recognize what’s happening and we imitated their decision.
Now I did have a lot of other narratives/decisions to choose from at this time (mid to late January). People in my immediate circles who are doctors ridiculed my idea to stock up on N95 masks. Popular magazines were derisive of the tech industry’s “excessive panic”. The president of the country I live in called the virus a hoax designed to undermine him.
I think I went with the tech industry narrative because of Preferential Copying
Preferential copying
Copying is ubiquitous in human culture. As children, we learn by copying and this is does not happen at random. They have a model bias in who they choose to learn from as written in this paper:
Children are biased towards those who intend to teach
Children are biased toward copying the most proficient models
Children are biased toward copying models belonging to a group which has a reputation for being proficient
Children are biased toward copying models that resemble themselves
Children are biased towards copying models with high status
While this explains why we copy, it does not explain why we choose the narrative of one group over the other or herd behavior.
Herd Behavior
Herd behavior is a heuristic for making decisions when the information and the quality of the information is uncertain as explained in this paper. In the absence of quality signals, we rely on the decisions made by decision makers before us to make our choice along with the information we have at hand. As succinctly captured by Sarah Perry:
The “herd behavior” demonstrated by adults can be modeled as a rational process: we copy others precisely in situations where their behavior offers cheap information about proper behavior, in situations in which we are uncertain, or acquiring information through other means would be too costly. Here again we have rational copying based on a complex underlying cognitive framework, rather than mere mimesis
Decision Queue - Prepping vs Keep calm and carry on
We can visualize the decisions of prepping vs keep calm and carry on as a long queue of people who have to make decisions one after the other based on:
Information they have about the virus
Information they have about the decisions made by their predecessors.
Within the tech herd, the earliest decision makers happened to be tech industry veterans such as Balaji Srinivasan and Bill Gates who have deep first hand experience in network effects along with expertise in healthcare. Hence, the quality of the information available to them was very high. I relied on their decisions to make mine and my guess here is that this is what transpired with a large segment of the tech herd. It also helped to a great extend that the tech herd shares information more freely and quite fast.
In the purely rational world of the scientific paper, the only variables for people in the decision queue would be information available to them and the decision made by predecessors. But real world decision making is not rational. Which brings me to herds that should have reacted faster but did not - journalists and doctors. I’m not sure what went wrong here but two simple theories I have are:
The journalist herd does not trust decisions made by the tech herd. To go back to the preferential learning biases - journalists have for a long time not considered the tech industry to be proficient in social decisions and does not fit their model of a high status signal in this context. So it is plausible that default tendency of journalists was to make the opposite choice made by “tech bros” before them in the decision queue.
Doctors don’t share information enough and even if they do, the existing mechanisms - scientific papers mediated by third parties are too slow. So the information available for early decision makers like the CDC is not high quality and hence the quality of decisions downstream is also poor.
One last thing I want to touch on is the lukewarm response of a large segment of the population to social distancing and shelter at home orders. The quality of information available to make decisions decreases downstream. This means options people choose to imitate tend to be based on decisions made by early decision makers who made their choice based on low quality information. This is indicated in this piece by Ben Thompson that describes how the quality of information available in social networks on a particular topic decrease with time:
“Now let me ask you another question: Which is the art of painting designed to be—an imitation of things as they are, or as they appear—of appearance or of reality?
Of appearance.
Then the imitator, I said, is a long way off the truth, and can do all things because he lightly touches on a small part of them, and that part an image. For example: A painter will paint a cobbler, carpenter, or any other artist, though he knows nothing of their arts; and, if he is a good artist, he may deceive children or simple persons, when he shows them his picture of a carpenter from a distance, and they will fancy that they are looking at a real carpenter.” - The Republic by Plato translated by B. Jowett
I guess I have to end with some declarative statement so I’ll end with - Plato is wrong, imitation/copying is a useful mechanism for learning but just because you chose the right imitation once does not mean you should use that same heuristic for all decisions
If you enjoyed reading this, sign up and share with others