Rara Avis
How can we know if we're ready for black swans, and change course if we're not? Great question.
I like to reason about collective human systems (societies, organizations, etc) using the same set of tools I would for studying a machine learning systems. Partly, this is because when you've got a nice new hammer, everything takes on a distinctively nailish aspect. However, it also seems that despite humans being agents exercising individual autonomy, collections of humans are awfully good at climbing uphill relative to metrics you can get them to agree to care about. We are all living longer, children are dying less frequently, and we are on average more wealthy. Also like a machine learning model, we often improve these metrics to the detriment of things outside of our collectively agreed upon reward function (externalities) and we don't distribute the average across all people, but that is a discussion for another time.
So, I guess my first answer to your question about what we can do to improve our resilience to black swans is this: learn to shape the reward functions of collections of people. Shape it towards what? That's where things get a bit tricky. Black swans by their very nature are hard to predict. And machine learning systems, like most systems, are extremely bad at dealing with "out of distribution data" - situations that are absent or extremely rare in the set on which they were trained. So, what can societies do? To take a page from machine learning, I think we must learn to both diversify and augment the dataset.
I think the strongest non-values based argument for building as diverse a society as we can is that it fundamentally expands our resilience to external shocks. By building a society (or an organization for that matter) that represents and simultaneously values the contributions of people of various backgrounds, genetic makeups, personalities, etc, our hill climbing system must satisfy many objectives at the same time. It is known in machine learning that such systems take a much more robust approach to solving the problem and are less disturbed by perturbations in the input. We see this in human systems as well. Take, as an arbitrary example, airlines which must service a wide range of customers both rural and urban. Different planes and technologies are developed for short haul flights to remote regions of Canada as long-haul flights from New York to Hong Kong. And so, when there is a shock to the system (a volcanic eruption, for example) we have a set of tools available to us to work around it: using short-haul routes to reroute around the volcanic clouds. Diversity is the strongest asset we have in improving our resilience to these black swan events.
What about augmenting the dataset? Humans have an amazing capacity that we are just learning how to replicate in machines. We can come up with wild fictions and learn from them! The communicator on Star Trek (1966) inspired Motorola's first flip phone (1973). Elements of the moon landing (1969) were predicted in Jule's Vernes From Earth to the Moon (1865). Holograms, radio, smart watches, antidepressants, the list goes on and on. Humanity has speculated about the future of science, of government, of war, of art, of exploration. These speculations aren't perfect, they're not meant to be. But they serve the useful purpose of equipping us with a framework, with vocabulary and imagery, to explore situations we have yet to find ourselves in and think at least a few steps ahead. So, fundamentally, to build our resilience to black swan events we must invest deeply in our creatives and listen to the signals they send us about the future, even if they are incorrect.
I'm not naive enough to think that these steps are sufficient. But I'm fundamentally an optimist and I like to think about what we do have the power to control. And, even the scenario of the coronavirus is one with a message of hope. At a time when the world seemed to have built up a great deal of political tension, governments found ways to cooperate, to control the spread of the virus, and help each other's citizens. It is a tragedy, but it demonstrates that for a big enough challenge there are ways of overcoming that which divides us to tackle it together.
I like to reason about collective human systems (societies, organizations, etc) using the same set of tools I would for studying a machine learning systems. Partly, this is because when you've got a nice new hammer, everything takes on a distinctively nailish aspect. However, it also seems that despite humans being agents exercising individual autonomy, collections of humans are awfully good at climbing uphill relative to metrics you can get them to agree to care about. We are all living longer, children are dying less frequently, and we are on average more wealthy. Also like a machine learning model, we often improve these metrics to the detriment of things outside of our collectively agreed upon reward function (externalities) and we don't distribute the average across all people, but that is a discussion for another time.
So, I guess my first answer to your question about what we can do to improve our resilience to black swans is this: learn to shape the reward functions of collections of people. Shape it towards what? That's where things get a bit tricky. Black swans by their very nature are hard to predict. And machine learning systems, like most systems, are extremely bad at dealing with "out of distribution data" - situations that are absent or extremely rare in the set on which they were trained. So, what can societies do? To take a page from machine learning, I think we must learn to both diversify and augment the dataset.
I think the strongest non-values based argument for building as diverse a society as we can is that it fundamentally expands our resilience to external shocks. By building a society (or an organization for that matter) that represents and simultaneously values the contributions of people of various backgrounds, genetic makeups, personalities, etc, our hill climbing system must satisfy many objectives at the same time. It is known in machine learning that such systems take a much more robust approach to solving the problem and are less disturbed by perturbations in the input. We see this in human systems as well. Take, as an arbitrary example, airlines which must service a wide range of customers both rural and urban. Different planes and technologies are developed for short haul flights to remote regions of Canada as long-haul flights from New York to Hong Kong. And so, when there is a shock to the system (a volcanic eruption, for example) we have a set of tools available to us to work around it: using short-haul routes to reroute around the volcanic clouds. Diversity is the strongest asset we have in improving our resilience to these black swan events.
What about augmenting the dataset? Humans have an amazing capacity that we are just learning how to replicate in machines. We can come up with wild fictions and learn from them! The communicator on Star Trek (1966) inspired Motorola's first flip phone (1973). Elements of the moon landing (1969) were predicted in Jule's Vernes From Earth to the Moon (1865). Holograms, radio, smart watches, antidepressants, the list goes on and on. Humanity has speculated about the future of science, of government, of war, of art, of exploration. These speculations aren't perfect, they're not meant to be. But they serve the useful purpose of equipping us with a framework, with vocabulary and imagery, to explore situations we have yet to find ourselves in and think at least a few steps ahead. So, fundamentally, to build our resilience to black swan events we must invest deeply in our creatives and listen to the signals they send us about the future, even if they are incorrect.
I'm not naive enough to think that these steps are sufficient. But I'm fundamentally an optimist and I like to think about what we do have the power to control. And, even the scenario of the coronavirus is one with a message of hope. At a time when the world seemed to have built up a great deal of political tension, governments found ways to cooperate, to control the spread of the virus, and help each other's citizens. It is a tragedy, but it demonstrates that for a big enough challenge there are ways of overcoming that which divides us to tackle it together.
Comments
Post a Comment