Have you ever considered joining a community for the sole purpose of meeting like-minded people? Have you ever been encouraged to join such communities in order to find members who share your interests? Well, I know I have. Meet-ups for people interested in machine learning. Online groups for people interested in software development. Group chats for people interested in personal knowledge management. All those communities are built on the premise that if you group like-minded people together, you get a comfortable environment where healthy discussions can take place. That’s a reasonable argument, but is there anything missing? Is like-mindedness the best catalyst for healthy discussions, or is it actually limiting?

One possible drawback of such a community is the fact that members are all familiar with the topic. Why is this a drawback? Well, if people are familiar with it, they are less likely to question it. Their similar background knowledge becomes implicit. Everyone assumes that this is how things are naturally going. There’s no “Wait, I don’t get it, why do you actually do that?” or “This doesn’t make sense, how does this follow?”. There’s no incentive for critical thinking when people don’t bat an eye to underlying assumptions.

Another drawback of a like-minded community is the problem of limited perspectives. If members have a similar background, they’ll likely have similar views on the topic at hand. Even with a minority of members with diverging views, the culture which forms around a topic leaves its homogeneous mark on the community. At the same time, many of the world’s toughest challenges require the collaboration between different fields and schools of thought. The pieces of humanity’s greatest puzzles aren’t to be found in any one particular field, but in a rich network of radically different ones. A post from Farman Street puts this nicely:

Here’s another way to think about it. When a botanist looks at a forest they may focus on the ecosystem, an environmentalist sees the impact of climate change, a forestry engineer the state of the tree growth, a business person the value of the land. None are wrong, but neither of them are able to describe the full scope of the forest. Sharing knowledge, or learning the basics of the other disciplines, would lead to a more well-rounded understanding that would allow for better initial decisions about managing the forest.

What if, instead of creating communities based on similar interests, we created communities based on dissimilar ones? Despite the reluctance, disparate fields will often prove converging because of the highly interconnected nature of human knowledge. Get together an architect, a doctor, a developer, a lawyer, an entrepreneur, a researcher, a teacher, a social worker, an accountant, and others, and incentivize them to explain things to each other. Let’s term this resulting community a Feynman cluster.

This configuration solves both previously-mentioned problems. Each person will have to constantly explain themselves in front of the others, as they can’t rely on any implicit domain-specific knowledge anymore. Moreover, because the rest are not familiar with the content, they’ll be able to easily probe the other’s knowledge for gaps and inconsistencies in the process of understanding. What’s more, due to the fact that they have different interests, they’ll likely have different views on many topics, which enables them to gain a rich collective understanding of a wide variety of them.

One might think of a torrent platform as an analogy for a Feynman cluster. Each member brings their own knowledge to the table, and they can tap into the whole entirety of the others’ knowledge in return. However, this analogy does not capture a unique property of such a community. On a torrent platform, it’s only considered fair that users balance the amount of content they provide (seed) with the amount of content they receive (leech). If a user accesses way more content than they offer, then they’re cheating. In contrast, in a Feynman cluster, members benefit from both tapping into the collective knowledge and from sharing their own knowledge. Why? Because of the Feynman technique.

When you know a lot about a topic and you’re talking with someone who doesn’t, you’re forced to systematically break down concepts, clearly explain the relations between them, and provide representative examples. Not only is this useful for your interlocutor, but it is also very beneficial for you. Trying to explain a topic forces you to extract its essence from the noise, helping you realize what actually defines it. It forces you to seek the fundamental principles, and it brings to light the shaky parts of your knowledge in the process. This is extremely valuable, as it incentivizes you to further develop your own knowledge, despite the fact that you’re not interacting with new information per se, just looking at it differently.

Considering this, it can be said that a Feynman cluster is a positive-sum game. Everyone benefits from sharing knowledge. People routinely switch between the role of the teacher and the role of the student, and they gain value from both. Compare this to a torrent platform, where you wouldn’t directly benefit from sharing information.

Concretely, Feynman clusters might take on various shapes and sizes. For example, such a community could consist of a group chat of highschool friends who went on to study different things in university. It could also be based on an online group where users have flairs describing their background (e.g. economics, medicine, architecture) and moderators (or bots) incentivize discussion between users with different backgrounds. It could also take the form of a meet-up with dynamically changing topics, which might consist of small-group and plenary discussions. There’s still a need for a reason to bring everyone together, but once that is settled, thought-provoking discussions can start to take place.

Like-minded communities and Feynman clusters are not in conflict. You could be part of communities of both types. However, time spent in one means time not spent in the other. This, together with the enhanced comfort of hanging around like-minded people, may prevent many from experimenting with Feynman clusters. Therefore, there’s a need for actively trying things out beyond the comfort zone.

To sum up, we are social creatures with extensive collective knowledge, and we also individually benefit from explaining things to each other. Why not combine the two in order to build a community like the Feynman cluster?