We simulated if you can really reach anyone in 6 steps

Veritasium

5,534,930 views • 10 months ago Save 23 min 10 min read

Video Summary

The concept of "six degrees of separation" suggests that any two people on Earth are connected through a chain of acquaintances in six steps or less. While a random network of 8 billion people would indeed have this property, real-world networks are far from random; they are highly clustered geographically. This clustering, paradoxically, also facilitates "small-world" connections. Through simulations, it was found that introducing even a small percentage of "shortcuts" dramatically reduces the degrees of separation, making the world feel much smaller.

This phenomenon of small-world networks, where dense local clusters coexist with efficient global connections, has profound implications. It explains how information, diseases, and even disruptions can spread rapidly. The presence of "hubs," like major airports or influential websites, further facilitates these rapid connections, acting as crucial nodes that make the entire network more accessible. However, these hubs also represent an "Achilles' heel," where a disruption at a central hub can cascade through the network, causing widespread effects.

Furthermore, network structure can influence behavior. While random interactions might foster defection in games like the Prisoner's Dilemma, the ability to choose connections and the presence of familiar "clumps" can promote cooperation. Ultimately, our networks shape us, but our choices and actions also shape the networks themselves, highlighting the power of individual agency within these interconnected systems.

Short Highlights

The "six degrees of separation" principle suggests any two people are connected within six steps, a concept proven mathematically in random networks but complicated by real-world clustering.
Small-world networks combine dense local clusters with efficient long-distance "shortcuts," drastically reducing separation degrees.
"Hubs," like major airports or popular websites, act as central connectors, making networks highly traversable but also vulnerable.
Network structure significantly impacts behavior, with local clusters fostering cooperation while random connections can encourage defection.
Individuals have agency within networks; choosing connections and proactive behavior can lead to cooperation and influence the network's overall nature.

Key Details

The Six Degrees of Separation Paradox [00:00]

An experiment in 1999 showed a falafel salesman could connect to actor Marlon Brando in six steps.
The concept posits that any two people on Earth can be connected in six steps or less.
This is mathematically true for random networks where each person knows 100 individuals.
In such a random network, connections expand exponentially: 2 steps connect 10^4 people, and 5 steps connect 10^10 people, more than the global population.
The core idea is that the world is surprisingly small due to interconnectedness.

The paradox arises because real-world networks are not random. They exhibit high clustering, meaning people naturally connect within geographical or social groups. This dense local structure, while seemingly limiting connections, paradoxically facilitates the "small world" phenomenon, making global connections surprisingly efficient.

"The question is not why is the world small, it's really how could it be otherwise."

The Impact of Small Worlds [01:16]

The increasing interconnectedness of the world is making it smaller, but this also exposes individuals to negativity.
Diseases and malevolent information can spread more easily through these conduits.
The "small world" effect, though mathematically predictable in random networks, has tangible negative consequences when applied to real-world scenarios.

The shrinking of the world due to increased connectivity is a double-edged sword. While it brings people closer, it also opens pathways for harm and negativity, impacting how diseases and undesirable information propagate.

"It's not only dangerous in terms of disease propagation, but anything malevolent now has conduits that it didn't used to have."

The "Small World" Problem and Network Clustering [02:41]

Real-world social networks are not random; people naturally cluster geographically.
Most acquaintances live nearby, and they are more likely to know each other.
Network clustering measures the fraction of people known who also know each other.
A highly clustered model, like people arranged in a circle knowing only their neighbors, would require millions of steps to connect distant individuals.
The paradox is how we perceive both local clusters and global connectivity simultaneously.

The reality of social networks is a stark contrast to random models. Geographic proximity and shared social circles create dense clusters, making direct connections between distant individuals seem improbable. Yet, the experience of the world suggests these connections are readily available.

"The truth is that people naturally cluster geographically."

Watts and Strogatz's Small World Model [04:23]

Mathematicians Duncan Watts and Steve Strogatz investigated the "small world problem" using computer simulations.
They studied networks that were neither perfectly ordered nor completely random, but somewhere in between.
Their model started with a regular network (like people in a circle) and gradually introduced random "shortcuts."
A shortcut is a connection to someone outside one's immediate social circle, like knowing someone in another country.
Introducing just a few shortcuts dramatically reduced the average number of steps needed to connect any two nodes.

The groundbreaking work of Watts and Strogatz demonstrated that even a small number of random connections can drastically shrink a network, merging the properties of ordered, clustered networks with the interconnectedness of random ones.

"As soon as he introduced a few shortcuts, the world immediately gets as small as a random graph."

The Power of Shortcuts and Clustering [06:07]

Rewiring just 1% of links to shortcuts in a simulation dropped the average separation from 50 to 10.
Crucially, clustering remained high even as the network became more connected due to shortcuts.
This means a network can simultaneously possess high local clustering and be a "small world."
For 8 billion people, only about 3 out of 10,000 friendships need to be shortcuts to achieve an average separation of six degrees.

The remarkable finding is that a small number of random connections can make a vast network highly connected while preserving its inherent local structure, explaining how both dense communities and global reach coexist.

"You could simultaneously have the clustering that we know is real and the small world that we know is real."

Real-World Small Worlds and "Weak Ties" [08:22]

Watts and Strogats tested their model on the neural network of the worm C. elegans, finding an average separation of just 2.65.
The model also proved true for Hollywood actors (average separation less than 4) and power grids.
This concept is related to the sociological phenomenon of "strength of weak ties."
Weak ties (acquaintances) are often more valuable for finding jobs or new information than strong ties (close friends).
These acquaintances act as shortcuts, bridging different social circles and making the world feel smaller.

The small-world phenomenon isn't confined to abstract models; it's observable in biological and social systems. The surprising utility of distant acquaintances, or "weak ties," highlights how these connections function as critical shortcuts in our social networks.

"The strength of weak ties. Cuz you ask people how they got their job and people would say, 'Oh yeah, I heard about it from, you know, Randy.'"

The Impact of Small Worlds on Dynamics [10:46]

Understanding small-world networks has implications for how things spread, such as diseases or information.
Simulations showed that introducing a few shortcuts dramatically accelerated disease spread compared to a purely clustered network.
A disease spread through a clustered network took 73 days to infect the whole population.
With just 10% shortcuts (a small-world network), the same disease took only 26 days to spread.
A fully random network spread the disease in 25 days, highlighting the efficiency of even minor shortcuts.

The structure of a network significantly impacts the speed at which things spread. Small-world networks, with their blend of clustering and shortcuts, facilitate rapid propagation, whether it's infectious diseases or other forms of information.

"Boom. Wow. That's really dramatic, right?"

The Virality of Network Science [12:50]

Watts and Strogats' 1998 paper on small-world networks became highly influential, garnering tens of thousands of citations.
The paper's impact spanned diverse fields, from neuroscience to English literature.
The concept's success led to unusual applications, including inquiries from the FBI regarding hair and fiber transfers in criminal investigations.

The findings on small-world networks had a profound and far-reaching impact, sparking research across numerous disciplines and even leading to unexpected practical applications in forensic science.

"It's probably worth making that distinction that citations are one measure of impact."

Preferential Attachment and Hubs [16:29]

Albert-László Barabási studied the internet, finding it also exhibited small-world properties despite its immense size.
Unlike Watts and Strogats' model, the web's structure was dominated by "hubs" – highly connected websites.
These hubs, like Yahoo, linked to thousands of other pages, making the web a small world.
Networks typically grow over time, and new nodes tend to connect to already well-connected nodes (preferential attachment).
This "rich get richer" mechanism naturally leads to the emergence of hubs in growing networks.

The internet's interconnectedness is not primarily driven by random shortcuts but by the emergence of dominant "hubs" through a process of preferential attachment, where new connections favor existing popular nodes.

"The more connected nodes simply because you are more likely to know a more connected node than a less connected n."

The Achilles' Heel of Networks [22:01]

Hubs fundamentally change network behavior, enabling efficient global connectivity but also creating vulnerabilities.
A disruption at a hub, like a thunderstorm grounding flights at Chicago O'Hare, can have cascading effects across the entire network.
This vulnerability of hubs is referred to as the "Achilles' heel" of networks.
Understanding this can be beneficial, for example, in network medicine for targeting crucial parts of disease networks.
Targeting hubs was also effective in curbing the HIV epidemic in Thailand by focusing interventions on high-contact individuals.

The concentration of connectivity in hubs, while facilitating rapid spread and access, creates critical points of failure. Disruptions at these hubs can destabilize entire systems, but this understanding also offers powerful strategies for intervention and control.

"So this is what we call the Achilles' heel of networks. And this could be good news or it could be bad news, right?"

Network Structure and Behavior: The Prisoner's Dilemma [24:24]

Watts and Strogats explored how network structure influences behavior using a simulation of the Prisoner's Dilemma game.
In this game, defection is the rational choice for individuals, but mutual cooperation yields a better collective outcome.
In a clustered network, cooperation could thrive, as small groups of cooperators could foster mutual benefit.
However, introducing shortcuts disrupted these cooperative clusters, leading to a breakdown of cooperation and widespread defection.

The way individuals are connected can profoundly alter their behavior, even with the same inherent strategies. While dense, familiar networks may encourage cooperation, the introduction of random connections can undermine these systems, leading to a rise in self-serving actions.

"All you're changing is the way they're connected. And you go from a world where everyone's completely nice and working together to one where it's filled with nastiness and people betraying each other only by changing how they're connected."

Agency and Choice in Networks [31:11]

Experiments on the public goods game showed that network structure initially had no effect on cooperation.
This was because people were not choosing their connections; they were assigned.
When players were allowed to choose their connections, cooperation significantly increased.
This highlights the power of individual agency: choosing to interact with positive influences and disengaging from negative ones.
Individuals have the power to initiate movements and drive change by acting decisively and choosing their associations.

The ability to choose our connections is a powerful factor in fostering cooperation and positive outcomes. By actively selecting who we engage with, we can shape our environment and contribute to beneficial collective behavior, demonstrating individual power within interconnected systems.