Photo by Christine Roy on Unsplash

Crypto and Network Effects

Ideas that propagate, and the staying power of dominant networks

Crypto and religion have a bit in common.

• They are both ideas.
• They both live in the collective imagination.
• They are both made by people.
• They both band people together.
• They are both network effects.

Network effects

In the book Sapiens by Yuval Noah Harari, Harari outlines a natural limitation of direct relationships and gossip to form bands.

In the wake of the Cognitive Revolution, gossip helped Homo sapiens to form larger and more stable bands. But even gossip has its limits. Sociological research has shown that the maximum ‘natural’ size of a group bonded by gossip is about 150 individuals. Most people can neither intimately know, nor gossip effectively about, more than 150 human beings.

This approximate 150 threshold is also known as Dunbar’s number. In Sapiens, Harari contends that it is collective myths that band groups of humans far past the 150 threshold.

Any large-scale human cooperation — whether a modern state, a medieval church, an ancient city, or an archaic tribe — is rooted in common myths that exist only in people’s collective imagination.

Collective myths (or ideas) are present in nearly all human interactions and are often valuable to humans as they enable new types of games for humans to cooperate and /or compete.

The internet has drastically impacted the speed at which ideas can spread, and has introduced new levels of compounding to new and existing collective ideas.

With network effects, dominant collective ideas can quickly emerge, settle, and defend their position in the collective imagination.

Dominant collective ideas

Technology

• The internet (25–30 years old, 4.6 billion users)

Source: Internet World Stats

Social media

• Facebook (10 years old, 2.6 billion users)

Source: Statista

• Twitter (14 years old, 330 million users)

Language (spoken)

• English (1,400 years old, 1.1 billion users)
• Mandarin Chinese (800 years old, 1.1 billion users)

Source: Visual Capital

Language (code)

• Javascript (25 years old, 10.7 million users)
• Python (31 years old, 7.0 million users)

Source: Daxx.com

Money (fiat)

• US dollar (106 years old, 370+ million users)
• Euro (21 years old, 340+ million users)

Source: Visual Capital

Money (code)

• Bitcoin (10 years old, approx. 50–100 million users)
• Ethereum (5 years old, approx. 25–50 million users)

Religion

• Christianity (2,000 years old, 2+ billion participants)
• Buddhism (2,500 years old, 300 million participants)

Political

• Nation states (500 years old, 7+ billion participants)
• Roman empire (1,000 years old, 60–70 million participants at peak)

Law

• US constitution (233 years old, ~330 million participants)

Economic

• Capitalism (300+ years old, 7+ billion participants)
• Communism (170+ years old, ~1.5 billion participants)

The most dominant collective ideas continue to grow:

1. They have the most individuals to spread the idea
2. A rational individual would choose to play an existing game with the most value

But 1) growth does slow, 2) network effects can unwind, and 3) networks can evolve.

1) Growth does slow

Here’s a example that looks at the growth factor of COVID-19.

Since network value increases with each additional participant, existing users are incentivized to find new participants, but eventually growth does slow.

The above are purely illustrative growth patterns for networks starting with one million participants.

E.g. in the 2.00x GF scenario (yellow line):

1. In Years 1–5, participants are sticky, once they join they stay, and each network participant is able to find one new participant, on average, every year (GF = 2x). The network doubles each year. Growth is exponential.
2. In Years 6–10, participants are less sticky and only 90% of network participants are able to find a new participant, on average, every year. Exponential growth comes to an end following an inflection point in Year 9 as more participants exit the system than participants that join.

Diminishing growth is expected, eventually

• Maximums: There are 7.6 billion people in the world. The possible growth in number of internet users this year is bound between 1 and 65 percent.

Caveat: no consideration is given here to machine users or “bots”

• Switching costs of alternatives: While the English language has about 1.1 billion participants, growth to 7.6 billion is unlikely given the switching costs associated with learning a new language.

If we are able to put in 10 hours a day to learn a language, then basic fluency in the easy languages should take 48 days, and for difficult languages 72 days. Accounting for days off, this equates to two months or three months time. If you only put in five hours a day, it will take twice as long. — The Linquist

• Geography-specific value: The value of the US constitution is concentrated, with its participant, in the US.

For things like the internet and Facebook, with few alternatives and few geographical restrictions, slowing growth can be attributed in part to population maximums. For the more fragmented collective ideas like nation states and language, switching costs and geography seem to explain the slowing of growth more than population maximums.

Historically, the value of geography-specific networks, like nation states, has been protected through physically fought wars. Wars in which the sheer strength of the network could be directed towards a competing networks through the assembly and application of militia.

But all networks defend their value. Even if defense mechanisms are less tangible (twitter bot armies, paid google search results, strategic social media trends, etc). And innovation and network iteration also seems to be a viable network defense (see ‘forks’ below).

Source 1, 2, 3, 4, and 5

2) Network effects can unwind

Empires

• The rise and fall of empires can be attributable to a variety of reason, but Khan Academy synthesizes these reason to provide high level overviews of fall factors of historical empires.

Money

• Resemblant to the succession of empires, money systems have evolved over time, with new iterations replacing the old.

Source: Thismatter

• And specific currency networks eventually unwind. The creation of the Euro in 1999 replaced 15 existing money networks that were systematically unwound and replaced with the Euro.

Source: OANDA

Social media

• Myspace experienced nearly vertical exponential growth 2004–2006, flatter growth 2006–2009, and declining growth 2009–2011. The unraveling of Myspace’s network has been very slow. Myspace has approximately 8 million monthly active users today.

Source: Vator

Gaming

• While WoW’s network effect is slowly unwinding, the gaming protocol’s ability to continually iterate its software has extended its lifespan.

Source: Techspot

The resilience and staying power of network effects is exemplified by both WoW and Myspace. After nearly a decade of deteriorating users, both networks still exist today.

3) Networks can evolve

While it is possible that there are OG WoW users from 2005 still playing today, for the most part users today were likely acquired over time given Blizzard’s ability to continually morph and evolve.

Similarly, Myspace’s network has had staying power as it has pivoted to an online community for music listeners.

Forks are common in software and enable the creation of new, but similar, code to respond to changing environments.

Money has undergone a number of forks, from commodity money, to representative money, to fiat, to crypto.

Within crypto, ethereum and bitcoin have both undergone a number of protocol updates.

With effective iterations of previous games, social networks, or money systems, new iterations are able to retain and even amplify existing network effects.

Competing networks

Generally, the flaws of existing systems are more likely to get the benefit of doubt, and new systems are more likely to be scrutinized or diminished as networks voraciously compete for users.

As network size increases so too does its value.

The same way network effects can work for emergent networks, the network effects of other, competing networks, work against these new, emergent networks.

Small networks that prove legitimacy in the short term, will likely either be acquired or incumbent networks are forced to work with and cooperate with the new network.

First movers

First mover networks often move towards a dominance that is resilient to competing networks. E.g.

• The internet (first open communication network)
• Twitter (first open short communication network)
• Bitcoin (first open value network)
• Ethereum (first open programmable value network)

Imitators of the above have had a tough time gaining traction. Quickly becoming games of the few vs. the many.

Unless significant enhancements are made over existing first mover networks that it is an entirely new type of network. Then the network is no longer competing, but a first mover in its own right.

Types of networks

Generally, networks can be:

• Opt in or opt out
• Unbiased or biased
• Open or closed

Opt in vs. Opt out networks

As David Hoffman talks about in his piece, “A Bankless Nation”, there are protocols that people opt-in to and protocols that people opt-out of.

Earlier in a protocol’s life the majority of growth may require individual discovery and opting in (e.g. new nations, crypto).

Later in a protocol’s life the majority of growth may be seamlessly passed on from parents to their kids (e.g. nationality, fiat money).

Trust takes time to build. As networks grow they are compounding trust.

Play iterated games. All returns in life, whether in wealth, relationships, or knowledge come from compound interest.

— Naval

For existing, opt-out protocols you start from a position of inherited trust. For new, opt-in protocols you may start from a position of distrust and skepticism.

This doesn’t mean one is more deserving of trust than the other. Both should still be evaluated based on their respective merits.

Unbiased and neutral networks

It’s not always clear why network effects grow, but what does seem to be clear is that once networks do grow, they have tremendous staying power.

The more open and neutral a protocol, the easier time it seems to have growing and sticking around, e.g. the internet.

It seems that once bias is placed at or near the foundation of a protocol, the less viable the protocol becomes, the more participants search for an alternative.

As Hoffman puts it, neutral protocols have no concern for

• Gender
• Race
• Religion
• Age
• Language
• Ability
• Nation-State citizenship
• Personal connections

Network privilege

There seems to be an interesting relationship between geography and network privilege.

Unimpeded access to dominant networks is a privilege. Depending on where you live,

• access to the internet (the dominant communication network) may be limited
• access to the dollar (the dominant value network) may be limited
• access to english (the dominant language network) may be limited

Optional, unbiased, and open protocols

The internet is one of these protocols that is optional, neutral, and open to all.

Similarly, we could praise the internet for what it’s not. The internet is not mandatory, it’s not biased, and it’s not closed to only select, privileged participants. Caveat being if layers atop the core internet protocols block access, e.g. ISPs introduces accessibility limitations to users.

The internet is a tool and a public good. Enabling the free, open transmission and consumption of information and new ways for humans to communicate and interact.

There is a need for a similar protocol to the internet, but for value. A protocol for value that is optional, neutral, and freely accessible. Where individuals are on equal footing.

Social contracts

On the latest Bankless podcast episode with Vitalik Buterin, social contracts are discussed (timestamp, 1:23:30). Social contracts are also a central theme in Hasu’s “Framework for sceptics”.

The idea of social contracts resonated with me and is good place to close this article.

Social contracts are present in each of the above collective ideas and are the glue that bands together these vast networks. Social contracts can be defined and enforced by the below:

• Action
• Speech
• Text
• Code

Over time, and with the advent of the internet, more and more social contracts are being defined in open source code. Open source code can move with the movement of its underlying social contracts.

It’s worthwhile to take a moment to reflect on the social contracts you’re currently signed.

Acknowledging how these contracts are:

• Defined (code/speech/text)
• Enforced (optional or mandatory)
• Changed (dynamic or unvarying)

And if these contracts are subject to:

• Privilege (access is closed, limited, or free)
• Bias (participants are on equal or unequal footing)

But perhaps most importantly, what types of social contracts do you want to be a part of?

References:

• Yuval Noah Harrari, Sapiens
• David Hoffman, A Bankless Nation
• Naval, on Compounding Relationships
• Bankless, Ep. 21 with Vitalik Buterin
• Hasu, A framework for sceptics, pt. 2

Get Best Software Deals Directly In Your Inbox