I first started thinking about the meta-coordination 4 years ago, in the context of rationalists arguing about community norms. It seemed to me that people were getting into fights that involved a lot of wasted motion, and failing to accomplish what seemed like obvious shared goals.
For a few years, the bulk of my thought process was a vague, dissatisfied "surely we can do better than this, right?". Many of the people arguing eventually went off to focus on their individual orgs and didn't interact as much with each other. Maybe that was the right solution, and all this worrying about meta-coordination and norm arguments was just a distraction.
Then a pandemic hit. Coordination became much more practical and important to me, and the concept of coordination pioneering became more directly relevant.
Here were some issues that felt coordination-shaped to me. In this post, I’m speaking largely from my experiences with the Bay Area rationality community, but I think many of the issues generalize.
Negotiating policies and norms within a single household. Do you lock down? If so, how do you go about it? What do you do if people disagree on how dangerous covid is, what practices are effective, or what’s worth trading off for safety?
Community contract tracing. If someone at a party later gets covid, are people entitled to share that information? How do we negotiate with each other about sharing that information? This includes concerns about privacy, public safety, and how to socially navigate trading those off against each other during a crisis.
Maintaining social connection. This might involve negotiation with your housemates over covid policy, or the housemates of your friends. Even if you and a friend each live alone, figuring out what kind of contact to have during a pandemic is at least a two-player game.
Housemate swapping/matchmaking. Housemates hadn't generally been selected for "having similar preferences of how to handle pandemics". There were several reasons people might have wanted to relocate. But people also had reason to not necessarily want to advertise that they were looking for new housemates – they might risk antagonizing their current roommates, or airing drama that was still unfolding. Switching houses is also an effortful, high cost decision that was difficult during an already stressful time.
Allocation of labor (intellectual and otherwise). There was a lot of stuff to figure out, and to do. There was an initial flurry of activity as everyone scrambled to orient. I think there was a fair amount of duplicate labor, and a fair amount of labor allocated to "figure out wtf is up with the pandemic?" that could have been spent on people's day job or other non-pandemic personal projects.
Maintaining organizational sync. Most organizations went remote. I think some organizations can do a decent job working remote, but I think it comes with costs. Some forms of communication translate easily to zoom, and some are much harder when you can’t bring things up briefly without scheduling a call being A Whole Deal. This prompts two questions of “What were the best ways to shift to remote?” as well as “Was it actually necessary to shift to fully remote? Could better coordinated orgs have found ways to stay in person without undue risk?”, or “Were there third options?”
From my perspective, these all feed into two primary goals:
The physical and mental health of my social network.
The capacity of the rationality and EA communities to continue doing important work. (In particular, this could have been a year where AI safety research made differential progress relative to AI capabilities research. But my sense is that this didn’t happen)
I think all the previous bullet points are meaty topics, that each warrant at least one blogpost worth of retrospective. I’m not sure which topics I’ll end up deep diving into. In this post, I wanted to give a broad overview of why coordination innovation feels so important to me.
“Coordination” is a somewhat vague word to cluster all those topics together with. I think, ultimately, it’s helpful if you can taboo “coordination”, and focus on individual problems and processes. But as I write this, I’m still in the process of thinking through exactly what went wrong, or what could have been improved, and how to cluster those problems/solutions/concepts. In some cases I think the issue was more like "actually making use of existing good practices for coordination (at the object level)", and in some cases I think metacoordination, and the coordination frontier, are more relevant.
What all of those items share is that they are multiplayer games. In each case, individuals made choices, but some good outcomes required multiple people to agree, or to make synergistic choices in tandem.
This blogpost is the first of a few posts for helping me organize my own thoughts.
There are a few frames that stand out to me to look at the situation:
Skills that could have helped.
Outlooks and orientation that could have helped.
Systems that could have helped.
Organizational structures or leadership that could have helped.
And then maybe a fairly different frameset around "Who's 'we', exactly?". I think there's multiple scales that it's worth looking at through a coordination lens – a couple individual people, a loose network of friends and colleagues, particular organizations, the vaguely defined "rationality community", and the broader structure of different cities, states, and countries.
Analogies to future crises
I expect to learn many things from a Pandemic Coordination Case Study, that I'd wish I'd known in 2020. But the most important question is "whether/how will this be relevant to future crises?"
It's possible there will literally be another pandemic in our lifetimes, and that many lessons will directly transfer.
My biggest current worry is "accelerating AI technology either disrupt the economy, and create situations of high-stakes negotiations, where some of the lessons from the pandemic transfer." There are different ways that this could play out (a few individuals within an organization, negotiations between leaders of organizations, government regulation, industry self-regulation, intergovernmental treaties).
And then, of course, there could be entirely novel crises that aren't currently on my radar.
I first started thinking about the meta-coordination 4 years ago, in the context of rationalists arguing about community norms. It seemed to me that people were getting into fights that involved a lot of wasted motion, and failing to accomplish what seemed like obvious shared goals.
For a few years, the bulk of my thought process was a vague, dissatisfied "surely we can do better than this, right?". Many of the people arguing eventually went off to focus on their individual orgs and didn't interact as much with each other. Maybe that was the right solution, and all this worrying about meta-coordination and norm arguments was just a distraction.
Then a pandemic hit. Coordination became much more practical and important to me, and the concept of coordination pioneering became more directly relevant.
Here were some issues that felt coordination-shaped to me. In this post, I’m speaking largely from my experiences with the Bay Area rationality community, but I think many of the issues generalize.
From my perspective, these all feed into two primary goals:
I think all the previous bullet points are meaty topics, that each warrant at least one blogpost worth of retrospective. I’m not sure which topics I’ll end up deep diving into. In this post, I wanted to give a broad overview of why coordination innovation feels so important to me.
“Coordination” is a somewhat vague word to cluster all those topics together with. I think, ultimately, it’s helpful if you can taboo “coordination”, and focus on individual problems and processes. But as I write this, I’m still in the process of thinking through exactly what went wrong, or what could have been improved, and how to cluster those problems/solutions/concepts. In some cases I think the issue was more like "actually making use of existing good practices for coordination (at the object level)", and in some cases I think metacoordination, and the coordination frontier, are more relevant.
What all of those items share is that they are multiplayer games. In each case, individuals made choices, but some good outcomes required multiple people to agree, or to make synergistic choices in tandem.
This blogpost is the first of a few posts for helping me organize my own thoughts.
There are a few frames that stand out to me to look at the situation:
And then maybe a fairly different frameset around "Who's 'we', exactly?". I think there's multiple scales that it's worth looking at through a coordination lens – a couple individual people, a loose network of friends and colleagues, particular organizations, the vaguely defined "rationality community", and the broader structure of different cities, states, and countries.
Analogies to future crises
I expect to learn many things from a Pandemic Coordination Case Study, that I'd wish I'd known in 2020. But the most important question is "whether/how will this be relevant to future crises?"
It's possible there will literally be another pandemic in our lifetimes, and that many lessons will directly transfer.
My biggest current worry is "accelerating AI technology either disrupt the economy, and create situations of high-stakes negotiations, where some of the lessons from the pandemic transfer." There are different ways that this could play out (a few individuals within an organization, negotiations between leaders of organizations, government regulation, industry self-regulation, intergovernmental treaties).
And then, of course, there could be entirely novel crises that aren't currently on my radar.