Last December, we ran a workshop on exploring civilizational sanity. Our core team consisted of the lead organizer and two co-organizers, one dedicated mostly to operations (ops). Other staff included a cook and two part-time volunteers. Ten people participated in the event. Overall, it was a success! Some things went really well. Some things we messed up [1]. If you want to run similar events, this post might be relevant for you.
Content
What we Explored
Our intention was to explore how the structure of our social systems and institutions influences individual behaviour, how you can lean into or protect yourself from that influence — as well as how to shape those systems as an individual [2]. To do that, we introduced ideas like incentive design, group rationality and inadequate equilibria. We drew inspiration from books likeSeeing Systems (Barry Oshry), Inadequate Equilibria: Where and How Civilizations Get Stuck (Eliezer Yudkowsky),Fair Play (Eve Rodsky), The Gulag Archipelago (Aleksandr Solzhenitsyn) and the Simple Sabotage Field Manual(CIA).
Content Sessions
We held a number of content sessions to introduce the aforementioned concepts. Participants were encouraged to share their experiences within their own social groups and tried to connect these experiences with the presented concepts. For example, when talking about a model of social roles within institutions [3], participants related the model to their own workplaces, academic institutions and volunteering groups.
In addition, we hosted discussion rounds where participants brought up topics of their own choice on the theme of civilizational sanity. Here, participants tried to apply the frames we provided to the questions they cared about. We discussed topics like the characteristics and origins of cult-ish group dynamics, the advantages and problems of groups without (explicit) hierarchies [4], as well as a concrete presentation and discussion about a governance issues in a philanthropic organisation one of the participants was working at.
Guest Talk
One highlight of the weekend was a talk by our guest speaker Jan Kulveit. Jan has previously advised the Czech government on COVID-19 measures during the pandemic. In his talk, he shared his experiences as a case study for understanding and navigating dysfunctional, high-stakes social environments. He highlighted examples of civilizational insanities he observed in government reactions to COVID-19, relating them to the incentive landscapes inside and between the institutions involved in decision-making. The small event size allowed for an extended Q&A session; this way the participants could hear about how the dynamics we addressed during content sessions play out in real life. Jan's talk was well reviewed by participants and we are very happy that he made it to the event!
Equinox
On the final evening of the event we played a negotiation role-playing game designed specifically for the weekend: Equinox. The purpose of the game was for the participants to embody the dynamics we had discussed theoretically in the content sessions. The participants really enjoyed this part of the event — it was the most well-received part of the weekend. One inspiration for applying roleplay as part of the workshop was the EXP camp run in 2025. teaching rationality-related topics via experiential education seems very promising. We plan to further experiment with this in potential future projects.
Equinox is set in a fictional bronze age world where two neighboring tribes have been at war for years. The two tribe councils have to come together to negotiate a peace agreement. The participants took on various council roles (King, Minister of War, Minister of Coin, etc.) with different and diverging interests, allegiances, and powers. The incentive design of the game made it difficult for the participants to reach a peace agreement as various dynamics lead characters to have different priorities that were hard to reconcile. Among these dynamics were in-group out-group dynamics, information asymmetries, conflicts of interest, the credible commitment problem or dynamics from selectorate theory.
After the game had ended with a failure to find peace [5] and a violent coup in one of the tribes, we discussed what participants had experienced and what they had learned from the game. The participants reflected on many of the aspects we wanted them to take away. One commonly shared experience was that time pressure and information overload was one of the prime factors that made it difficult to come to an agreement. Most of the players felt very immersed in their roles and emphasized more with the difficulties of making hard decisions under uncertainty in a political leadership role. One dynamic that was also discussed was that roles that had incentives to sabotage the negotiations had a much easier job than the roles that tried to actively achieve peace.
While we do not currently plan to release Equinox as an independent game we are interested in further developing the game and potentially running it with interested groups of 6-12 participants. If this sounds interesting, feel free to reach out to us!
Culture Setting
Staff
We had some blind spots related to culture setting. Mainly, we neglected to explicitly align some central expectations within the organizing team. The lead organizer wanted the organizer-participant relationships to have a friendly-professional tone, while one of the co-organizers wanted them to be informal and intimate. This led to conflict in the organising team during the event. Culture setting is something the lead should have tracked explicitly, with the support of all co-organizers!
With regards to other staff, we were happy for them to participate and contribute to the program in their time off (if they wanted to!). However, that was not clearly communicated to participants — some of whom didn't know how to relate to those "part-time participants". Related to neglected expectation setting, we should have better briefed the staff on what vibe we wanted to cultivate at the event and how they might fit in.
Participants
The first session we ran was on event culture. In addition, some culture setting was done implicitly, e.g. by the choice of venue design and participants. One of the central ideas for participant selection was to bring people from different communities together — including rationalist communities, EA, math academia, cognitive science, and Go. We wanted to emphasise how these communities complement each other. This was mostly a success!
However, we still didn’t account for some foreseeable culture clashes. In hindsight, some expectations should have been spelled out more clearly. Notably, some people imported a casual cuddling culture from rationalist and EA community events in Germany; this was uncomfortable for some participants who didn’t come from that background. The lead organizer tried to discourage public cuddling but did not communicate it clearly enough. Afterwards, we received some complaints about strong displays of affection in the public space.
One piece of culture that was readily taken up was giving participants a lot of autonomy. No part of the schedule was mandatory [6]. People could run their own sessions and activities in the afternoon. Most participants brought in what was of interest to them and actively pursued what they wanted to get out of the weekend. Another win for event culture was assigning tasks to participants — refilling the tea tank, taking pictures of the flip chart notes, etc. The tasks were small enough for participants to be able to focus on the event while being actually useful and making the space more pleasant!
Main Takeaways
If you plan to run a similar event, the following questions might be valuable: What kind of vibe do you want to cultivate? What exactly does the role of participant, volunteer or organiser entail? How do you want participants to relate to each other? How do you want for participants to contribute? What should the relationship between staff and participants look like? How will people learn about these expectations?
Logistics
Outreach and Admissions
We started the outreach process later than ideal — our announcement was posted a little more than two months before the beginning of the event. Admissions functionally ended two weeks before the weekend. This type of short notice and uncertainty likely skewed our applicants young [7].
Moreover, we did not check the EA/rat event calendar, which meant losing some strong applicants to EAGx Amsterdam. Relatedly, the comparatively low contribution to career capital of the event turned away at least a couple of people who had “better” things to do. On the other hand, this meant people who did apply were generally a good fit. Interviews were time-consuming to organise, but we endorse them as participant fit is extremely valuable to such a small event.
Making people confirm their participation by sending a fee [8] served the primary purpose of making sure people either come or cancel on time. We got only one short-notice cancellation out of eleven participants, which we consider a success.
Venue and Food
We wanted the event to have a personal, cozy, and introspective vibe and made design choices to instantiate this. Firstly, we chose a venue with a rustic ambiance, which was well-received and helped elicit the event atmosphere we hoped for. Secondly, we employed a cook that we personally knew and thought would fit the tone of the event. Based on the anonymous feedback form, people generally enjoyed the food, which helped add to the cozy, homely aesthetic.
The main error we made regarding the venue was not insisting on checking it out in advance. We asked, but did not press when our question was ignored. The venue had a couple of “quirks” that made it logistically annoying to handle. Knowing this in advance would have helped us prepare adequately. One thing we did well was booking the venue for one day before the start of the event. This was indispensable because it gave us time to set up the venue and let staff settle in.
We offered a shuttle service from a nearby transport hub to the venue, as it was hard to reach by public transport. In the original acceptance form, nearly everyone claimed that they would take it both ways. In practice, some of them changed plans about how they wanted to arrive or leave. We should have given a monetary incentive for participants to report their intentions around the shuttle accurately [9].
Ops Staff
Logistics for the event turned out to be too much for the dedicated ops person to handle. We made the correct call to ask for volunteer help from friends, ensuring things went smoothly. In retrospect, it would have been ideal to have two people dedicated to ops before and during the event. For rural venues, ops people should all have a driver's license.
Central Insight
Nothing goes quite according to plan from the ops perspective. We experienced a wide range of unexpected difficulties that we didn't mention in detail [10]. While you should definitely plan ahead, run pre-mortems and improve your plans, you'll benefit from having the slack to adapt to contingencies that weren't accounted for. Generally, slack comes in the form of redundancy; for instance, it's wise to have extra ops/content people on hand, even if fewer are likely enough to get everything done. Part of our event's success was due to us having adequate back-ups (e.g. volunteer help).
Examples of hiccups or difficulties include our car needing unexpected maintenance, coordinating with our guest on short notice, and adjusting ad-hoc to issues with the venue.
Last December, we ran a workshop on exploring civilizational sanity. Our core team consisted of the lead organizer and two co-organizers, one dedicated mostly to operations (ops). Other staff included a cook and two part-time volunteers. Ten people participated in the event. Overall, it was a success! Some things went really well. Some things we messed up [1]. If you want to run similar events, this post might be relevant for you.
Content
What we Explored
Our intention was to explore how the structure of our social systems and institutions influences individual behaviour, how you can lean into or protect yourself from that influence — as well as how to shape those systems as an individual [2]. To do that, we introduced ideas like incentive design, group rationality and inadequate equilibria. We drew inspiration from books like Seeing Systems (Barry Oshry), Inadequate Equilibria: Where and How Civilizations Get Stuck (Eliezer Yudkowsky), Fair Play (Eve Rodsky), The Gulag Archipelago (Aleksandr Solzhenitsyn) and the Simple Sabotage Field Manual (CIA).
Content Sessions
We held a number of content sessions to introduce the aforementioned concepts. Participants were encouraged to share their experiences within their own social groups and tried to connect these experiences with the presented concepts. For example, when talking about a model of social roles within institutions [3], participants related the model to their own workplaces, academic institutions and volunteering groups.
In addition, we hosted discussion rounds where participants brought up topics of their own choice on the theme of civilizational sanity. Here, participants tried to apply the frames we provided to the questions they cared about. We discussed topics like the characteristics and origins of cult-ish group dynamics, the advantages and problems of groups without (explicit) hierarchies [4], as well as a concrete presentation and discussion about a governance issues in a philanthropic organisation one of the participants was working at.
Guest Talk
One highlight of the weekend was a talk by our guest speaker Jan Kulveit. Jan has previously advised the Czech government on COVID-19 measures during the pandemic. In his talk, he shared his experiences as a case study for understanding and navigating dysfunctional, high-stakes social environments. He highlighted examples of civilizational insanities he observed in government reactions to COVID-19, relating them to the incentive landscapes inside and between the institutions involved in decision-making. The small event size allowed for an extended Q&A session; this way the participants could hear about how the dynamics we addressed during content sessions play out in real life. Jan's talk was well reviewed by participants and we are very happy that he made it to the event!
Equinox
On the final evening of the event we played a negotiation role-playing game designed specifically for the weekend: Equinox. The purpose of the game was for the participants to embody the dynamics we had discussed theoretically in the content sessions. The participants really enjoyed this part of the event — it was the most well-received part of the weekend. One inspiration for applying roleplay as part of the workshop was the EXP camp run in 2025. teaching rationality-related topics via experiential education seems very promising. We plan to further experiment with this in potential future projects.
Equinox is set in a fictional bronze age world where two neighboring tribes have been at war for years. The two tribe councils have to come together to negotiate a peace agreement. The participants took on various council roles (King, Minister of War, Minister of Coin, etc.) with different and diverging interests, allegiances, and powers. The incentive design of the game made it difficult for the participants to reach a peace agreement as various dynamics lead characters to have different priorities that were hard to reconcile. Among these dynamics were in-group out-group dynamics, information asymmetries, conflicts of interest, the credible commitment problem or dynamics from selectorate theory.
After the game had ended with a failure to find peace [5] and a violent coup in one of the tribes, we discussed what participants had experienced and what they had learned from the game. The participants reflected on many of the aspects we wanted them to take away. One commonly shared experience was that time pressure and information overload was one of the prime factors that made it difficult to come to an agreement. Most of the players felt very immersed in their roles and emphasized more with the difficulties of making hard decisions under uncertainty in a political leadership role. One dynamic that was also discussed was that roles that had incentives to sabotage the negotiations had a much easier job than the roles that tried to actively achieve peace.
While we do not currently plan to release Equinox as an independent game we are interested in further developing the game and potentially running it with interested groups of 6-12 participants. If this sounds interesting, feel free to reach out to us!
Culture Setting
Staff
We had some blind spots related to culture setting. Mainly, we neglected to explicitly align some central expectations within the organizing team. The lead organizer wanted the organizer-participant relationships to have a friendly-professional tone, while one of the co-organizers wanted them to be informal and intimate. This led to conflict in the organising team during the event. Culture setting is something the lead should have tracked explicitly, with the support of all co-organizers!
With regards to other staff, we were happy for them to participate and contribute to the program in their time off (if they wanted to!). However, that was not clearly communicated to participants — some of whom didn't know how to relate to those "part-time participants". Related to neglected expectation setting, we should have better briefed the staff on what vibe we wanted to cultivate at the event and how they might fit in.
Participants
The first session we ran was on event culture. In addition, some culture setting was done implicitly, e.g. by the choice of venue design and participants. One of the central ideas for participant selection was to bring people from different communities together — including rationalist communities, EA, math academia, cognitive science, and Go. We wanted to emphasise how these communities complement each other. This was mostly a success!
However, we still didn’t account for some foreseeable culture clashes. In hindsight, some expectations should have been spelled out more clearly. Notably, some people imported a casual cuddling culture from rationalist and EA community events in Germany; this was uncomfortable for some participants who didn’t come from that background. The lead organizer tried to discourage public cuddling but did not communicate it clearly enough. Afterwards, we received some complaints about strong displays of affection in the public space.
One piece of culture that was readily taken up was giving participants a lot of autonomy. No part of the schedule was mandatory [6]. People could run their own sessions and activities in the afternoon. Most participants brought in what was of interest to them and actively pursued what they wanted to get out of the weekend. Another win for event culture was assigning tasks to participants — refilling the tea tank, taking pictures of the flip chart notes, etc. The tasks were small enough for participants to be able to focus on the event while being actually useful and making the space more pleasant!
Main Takeaways
If you plan to run a similar event, the following questions might be valuable: What kind of vibe do you want to cultivate? What exactly does the role of participant, volunteer or organiser entail? How do you want participants to relate to each other? How do you want for participants to contribute? What should the relationship between staff and participants look like? How will people learn about these expectations?
Logistics
Outreach and Admissions
We started the outreach process later than ideal — our announcement was posted a little more than two months before the beginning of the event. Admissions functionally ended two weeks before the weekend. This type of short notice and uncertainty likely skewed our applicants young [7].
Moreover, we did not check the EA/rat event calendar, which meant losing some strong applicants to EAGx Amsterdam. Relatedly, the comparatively low contribution to career capital of the event turned away at least a couple of people who had “better” things to do. On the other hand, this meant people who did apply were generally a good fit. Interviews were time-consuming to organise, but we endorse them as participant fit is extremely valuable to such a small event.
Making people confirm their participation by sending a fee [8] served the primary purpose of making sure people either come or cancel on time. We got only one short-notice cancellation out of eleven participants, which we consider a success.
Venue and Food
We wanted the event to have a personal, cozy, and introspective vibe and made design choices to instantiate this. Firstly, we chose a venue with a rustic ambiance, which was well-received and helped elicit the event atmosphere we hoped for. Secondly, we employed a cook that we personally knew and thought would fit the tone of the event. Based on the anonymous feedback form, people generally enjoyed the food, which helped add to the cozy, homely aesthetic.
The main error we made regarding the venue was not insisting on checking it out in advance. We asked, but did not press when our question was ignored. The venue had a couple of “quirks” that made it logistically annoying to handle. Knowing this in advance would have helped us prepare adequately. One thing we did well was booking the venue for one day before the start of the event. This was indispensable because it gave us time to set up the venue and let staff settle in.
We offered a shuttle service from a nearby transport hub to the venue, as it was hard to reach by public transport. In the original acceptance form, nearly everyone claimed that they would take it both ways. In practice, some of them changed plans about how they wanted to arrive or leave. We should have given a monetary incentive for participants to report their intentions around the shuttle accurately [9].
Ops Staff
Logistics for the event turned out to be too much for the dedicated ops person to handle. We made the correct call to ask for volunteer help from friends, ensuring things went smoothly. In retrospect, it would have been ideal to have two people dedicated to ops before and during the event. For rural venues, ops people should all have a driver's license.
Central Insight
Nothing goes quite according to plan from the ops perspective. We experienced a wide range of unexpected difficulties that we didn't mention in detail [10]. While you should definitely plan ahead, run pre-mortems and improve your plans, you'll benefit from having the slack to adapt to contingencies that weren't accounted for. Generally, slack comes in the form of redundancy; for instance, it's wise to have extra ops/content people on hand, even if fewer are likely enough to get everything done. Part of our event's success was due to us having adequate back-ups (e.g. volunteer help).
I appreciate the irony of an event on civilisational sanity still running into a bunch of civilisational insanities!
More on our motivation in the original announcement
See "Seeing Systems" by Barry Oshry
A discussion that self-ironically started with a twenty minute power struggle about the purpose and leadership of the discussion group.
We were not that surprised. The game is set up in a way that makes success pretty hard to achieve.
Participants were encouraged to attend the roleplay game Equinox since it was designed to be the climax of the event.
Only two participants were over 30
100 euros
We did this successfully to make participants report their intent to come to the event, and should have just done the same thing for the shuttle.
Examples of hiccups or difficulties include our car needing unexpected maintenance, coordinating with our guest on short notice, and adjusting ad-hoc to issues with the venue.