Eureka! On the clustering of geniuses

Examining the social networks essential to innovation

I

Why do so many discoveries happen at roughly the same time to different people? There are cool sounding names like multiple discovery and simultaneous invention here which only serves to make the amazing seem rather mundane.

  • Newton created calculus and definitely was an all round well credentialed genius. But at the exact same time Leibniz cooked it up too, somewhere far enough away that it seems like a spooky coincidence.

  • Evolution, probably one of the greatest discoveries known to man, so controversial that it's still about as counterintuitive as quantum physics, was simultaneously theorised by Darwin and Alfred Russel Wallace.

  • Did you know that neutrinos have mass? Well Arthur McDonald and Takaaki Kajita figured it out independently at the same time and got to share a Nobel for it

  • In the 1770s a few interesting things happened. There was that British colony that got uppity. But before that two people figured out what Oxygen was only a couple of years apart

  • Jet engines were built by the Allies and Germans independent of each other trying to one up the other side during World War II

  • Every chemistry student's nightmare of the periodic table of elements, with all its legendary internal coherence and annoying organisation, was discovered by Mendeleev and Lothar Meyer a year apart

What gives?

One option of course is that out of the total sum of all inventions and discoveries that we've made as humankind, this is a tiny fraction. And though it's a sampling, it would actually be even weirder if there were zero coincidences like this. That's where Wikipedia comes to the rescue and says actually there are a bucketload of them.

But still, it could be coincidence.

Though when you read through the list it seems like it contains an awful lot of what people would refer to as the greatest scientific discoveries ever made. So when I took a look at a sample (The Atlantic, Kendall Haven's book) there's actually a reasonable amount of overlap.

But a lot of the overlap is in the application of certain technologies rather than the discovery itself. Once Leibniz invented calculus, he didn't immediately set upon to use it and discover gravity. That remained Newton's to figure out.

So perhaps the curious fact isn't just that people managed to figure out things at the same time. But that they were looking at the same place at the same time.

That opens up the question of whether there were groups of smart people who all thought to explore similar research areas at the same time, and turns out there are! Clusters of people who looked at the same fields at the same time are everywhere! In fact when you look at innovation spreading in any era it usually follows a cluster of some sort.

I started writing about the weird fact that turn of the century Hungary was a weird place, seemingly filled with geniuses you could (but probably shouldn't) hit with a rock in any direction.

There's a mythology about Hungarian super geniuses who were born at the turn of the 20th century, all in a suspiciously small neighbourhood, often going to similar or the same schools, and growing up with each other. John von Neumann, Paul Erdos, Leo Szilard, Edward Teller, Nicholas Kurti, Peter Lax, Nicholas Kaldor, George Soros, Andy Grove, Dennis Gabor, George Polya, the list goes on and on. They towered over fields of mathematics, physics, chemistry, biology, economics, computer science, even business. There's barely a human endeavour where they did not make a large and towering difference.

This seemed to also be true in Silicon Valley. Paypal mafia, Stanford mafia, the old school Intel mafia, never forget Bell Labs mafia. There seemed to be quite a few groups of people who somehow seemed to have all found each other and created billions of dollars through happy happenstance.

I wondered whether there was a super radiation spike, or something in the water, or whether there was a Professor Xavier hidden somewhere in there, but none of those seemed particularly credible.

So there are three possibilities. One is that Hungary at the turn of the 20th century was hit with some sort of genius radiation causing a genetic mutation in a large fraction of the population, and hey presto, genius alert. It's unlikely, of course, but what else could explain the extraordinary co-location of such a large number of geniuses who were all so good, that they got to the top of their fields.

Another option is that they were all pretty bright, but something in the cultural zeitgeist combined with some special educational process or insight created their special abilities. Like Professor Xavier's school for gifted children, they created superheroes. The puzzle here is what made it stop. Surely if you make superheroes for a couple of decades, you don't stop. You keep going! But it doesn't feel like Hungary is particularly adept today at creating more von Neumanns.

My conclusion then was that somehow genius clusters make each other succeed more often than they would have otherwise. Which isn't a particularly controversial thing to say, since that's what seems to happen anyway. Put smart people together, they do even smarter things. Standing on the shoulders of giants and all.

II

And within this group of geniuses who have all found each other and exchange ideas to cross-pollinate and create even better ones, another constraint is needed. They need to have the right toolkit and ideas in the first place. In Ancient Greece, in the absence of much scientific progress, the advances were mostly made amongst subjects that can be done solely in the mind - philosophy and mathematics.

Though we don't know the names, I imagine there would have been similar cliques in architecture in Ancient Rome, or Egypt for that matter. In the Renaissance, they could build on the knowledge that had existed on arts and architecture to sculpt Pieta or build the Pantheon. In Central Europe in the late 19th century there were pressing problems in physics and chemistry which were opening up entire vistas of knowledge.

So the requirements become larger numbers of smart people, who have the time to cogitate and work on things that go beyond their daily bread, connected in networks that enable them all to rise together, rather than just succeed individually. That still seems like something that could exist, and did exist, at multiple points in human history.

In almost any society there is an upper class which has enough smart people with enough free time and who know each other from educational and social circles, that this criterion gets filled. So there has to be something more, something that defines not just their ability to work on something, but the ability to actually make progress on something.

I could work half a century on anti-gravity with other smart people, but I wouldn't place a bet on my ability to succeed in floating serenely above the clouds. Some things are harder than just to be a function (talent, time to work, network). It also needs the variable that is feasibility of making progress.

Which means that it's not just that there was a vast sea of ignorance that the geniuses could try to work on together, each guided by their own idiosyncratic likes and viewpoints. It's that oftentimes there existed translucent and tenuous maps that showed potential routes to actually conquer these fields.

It's not that Heraclitus had a brainwave that atoms could potentially exist. It's that Bohr created an actual model of what that might look like, and devised experiments to get to the bottom of the matter.

Sometimes fields get to a pregnant stage where everyone gets really sure that if they just work hard enough they will uncover all sorts of things. Sometimes they're right, and other times they are not. But the feeling is still there, and it's rather unmistakeable.

Whatever the right metaphor, the scene has to be set right. Nobody could invent Instagram, as much as they might have wanted to, before people had smartphones. Nobody could invent Microsoft before people had PCs. Nobody could invent CRISPR before we understood DNA and a whole lot else besides which I don't understand well enough to explain.

Standing on the shoulders of giants is a known cliché but it also happens to be true. More importantly it happens to mean that when there are giants standing around that you can climb up on, as long as there are enough of you, many of you will be looking at the same vista at the same time.

We're not waiting for the tutelary deity to inspire us, or for genius insights to strike like never before without prior warning, but often the problems we look at are the same, and the tools we use are often similar, and the solutions we come up with naturally rhyme. As the inimitable Pratchett wrote: about Leonardo da Quirm:

Particles of raw inspiration sleet through the universe all the time. Every once in a while one of them hits a receptive mind, which then invents DNA or the flute sonata form or a way of making light bulbs wear out in half the time. But most of them miss. Most people go through their lives without being hit by even one. Some people are even more unfortunate. They get them all.

III

One highly suggestive outcome of this analysis is also that ideas are everywhere and much less exclusive than one might imagine.

When looking into this I came across a list of 7 "greatest ideas" in history. Even forgetting the biases that might have been involved and the myriad ones that didn't make it into the list, it's a good starting point. The ones identified were:

  • Aristotle's Logic

  • Gödel's Incompleteness Theorems

  • Newton's Physics

  • Darwin's Natural Selection

  • Einstein's Miracle Year

  • Shannon's Information Theory

  • Tim Berners-Lee's World Wide Web

Even within this rarified list let's see what we find:

  • Godel's theorem was clearly a massive accomplishment. However in the theorems of limitations of formal systems there's Tarski's undefinability theorem and Church's proof and Turing's halting theorem, all of which revolve around the same few issues

  • Darwin's contemporary Alfred Wallace also developed the theory quite a bit, and did so independently

  • Newton's physics had Leibniz as a contemporary, if only in the mathematical analysis part and discovery of calculus

  • Lorentz and Poincare provided bases for the special relativity theory that Einstein developed - Lorentz showed his work on the time transformation theory and Poincare published Sur la dynamique de l’électron in 1905, which he expanded upon later

  • And re Shannon's theories - "Much of the mathematics behind information theory with events of different probabilities were developed for the field of thermodynamics by Ludwig Boltzmann and J. Willard Gibbs"

  • Tim Berners Lee wrote the first web browser in 1990, the World Wide Web, but Paul Baran and Donald Davies conceptualised packet switching, who together with Kleinrock created networks, and many more who contributed to bring this to reality

What this shows are two things - 1) that all genius insights are explorations of plateaus of knowledge that exist atop various mountains, each explored to a different extent, and 2) since this exploration of areas unknown is performed by people who "stand on the shoulder of giants", there's quite often the phenomenon of simultaneous discoveries.

This is also a constant argument in entrepreneurial circles. Examples one, two, three, ... infinity. Was pets.com a horrible idea, or a precursor to Chewy? Was Webvan destined to fail, or waiting to come back to life as Instacart? The answer is yes. It's not that the idea itself is somehow good or bad, it's that the collective hallucination of a new world that ought to be is not something to be imagined alone.

So are the ideas themselves precious, or is it about execution by the right person at the right time? Is progress a function of one lone genius after another, or is it an "idea whose time has come"?

To be successful clearly indicates that you need not just a killer idea, which apparently a bunch others would also be having an epiphany about at that same moment, but also execution. And that execution is made easier when you're surrounded by people with whom you have a trusted relationship.

An interesting aspect therefore is also that even if simultaneous invention seems to happen an awful lot for a bunch of reasons mentioned above, quite often true breakthroughs in all sorts of fields seems to happen with folks who seem to be part of a formal or informal cliques.

If you want to think one step beyond specific innovations, like Facebook, to groups of innovations like transforming an entire industry, then the presence of these clusters is inevitable.

For the semiconductor industry that happened in the eponymous valley. Simultaneous progress brought about by several visionaries pushing the industry forward. Same perhaps with Baghdad in the Golden Age or Edinburgh in the 19th century or Paris in the 1920s or the coffee houses of Vienna in 1900s.

But why are those folks in these clusters more likely to succeed in the first place?

It's Kevin Bacon all over again. If you're 2 nodes away from the dream customer, acquirer, partner, etc for your company, then life's pretty good. Chances of your success just went up. If you're 1 node away, the chances don't double. They go up much much more.

The feedback loops are much tighter within one. Success in one member seems to engender success in others if at all there is potential. The halo effect of being a part of something successful makes the next idea or venture that much more likely to succeed. And the feedback that can be gotten from others increases the chances of success. Plus you have a ready group of strong minds who are predisposed to listen to you out of their own volition.

And that's also the benefit of actually being in a cluster. Doing anything is hard. So it gets much easier when people around you, who are also trying to do stuff, help you make like an internal market for you, and just by that, lift you way up and above the rest.

Why do startups go and join Y Combinator, arguably the most prestigious accelerator in the world? It's not just for the money, or even the access to the valley. It's because chances are that the cohort you'll be a part of will have a crazy success in it, and being part of that orbit can be helpful.

IV

One other weird outcrop of this is that the null hypothesis is that there is a tremendous amount of groupthink happening everywhere in the world all the time.

People who are bona fide geniuses, at the cutting edge of their respective fields, seem to have an unerring instinct towards the same few ideas and exploration targets.

Tired of science clichés? Fine, let's do some business clichés.

  • Reid Hoffman had been apparently obsessed with social networks for a good long while before he built LinkedIn. He had first built socialnet.com, one of the first social networks, but along the way also invested in Friendster (may it rest in peace) and Facebook (may its soul find peace).

  • Uber and Lyft started within a year of each other, with both taking some hefty inspiration from each other

  • Microsoft and Apple started pretty much at the same time, and even roughly within the same area code!

  • Paypal started at about the same time as X.com, which it later merged with

That list is endless too. Any VC or entrepreneur in the valley would tell you that there are themes that sweep the industry regularly. Social networks were all the craze at one time. Then the app craze. Then the mobile gaming.

In the enterprise world there was MapReduce giving birth to Cloudera and Hadoop around the corner from each other. Want performance management? New Relic, Datadog, Elastic if you're into that, AppDynamics and more. Robotic Process Automation is your jam? Here comes Blue Prism, Automation Anywhere and UIPath.

It's not just a question of companies simply copying a trend and a few winners emerging as if through lucky rolls of the dice. Each company has founders and executives who have pretty deep knowhow in the space. In fact if you read any of the narratives about them it often reads like an adoring profile. You can't even conceive of a world where they didn't end up starting the company.

Now if it were true that successful companies are the result of incredible insights, this should happen much lesser than it does. But if they're the result of the envelope of what's actually possible slowly being pushed beyond where it is currently, then each point that pushes it forward seems to have a big wave supporting it.

So part of the idea might also be that the zeitgeist is a real thing that exists. The collection of actions people have taken, the ideas that are in the ether, and the Overton window that exists for potential in the future, they all meld together into a few new paths. That's also what results in the same few ideas, or at least research directions, being popular amongst those doing the searching.

V

So there's usually a relatively well understood list of problems that several of the discoveries (of the non-serendipitous penicillin types anyway) seem to have followed. And there's the enthusiasm that comes from the fact that the ideas rely upon strong execution to succeed, rather than a pure spark of unalloyed and irreplaceable genius.

How do the people who get the idea and want to pursue them get together anyway?

If you look at the Hungarian cluster, they all seemed to be within a few blocks of each other.

If you look at the Silicon Valley cluster, they all seemed to be within a few miles of the Stanford campus.

If you look at the exceptional outpouring of art in the Renaissance, it came from the few city states that had amenable, even modern, rulers.

While there are examples of clusters forming semi-virtually, leading to quasi-independent discoveries (hello again Poincare and Wallace), for the true breakout across multiple domains we have to move from individual discoveries to full fledged expansion. From a lone explorer trying to claim new land to a band together who settle a city.

One commonality that we find when looking across these clusters is that they are all geographically close by. This by itself shouldn't surprise us. Geoffrey West's fantastic work shows us that cities, for instance, help create exactly this dynamic. In bringing people together it increases the rate of reproduction of ideas, speeding up the emergence of new ones.

Another is that they often have highly similar social circles. This again shouldn't be too surprising. We can barely be said to be in an era where previous social standing and upbringing are of limited predictive interest, so it stands to reason that this would be one way in which people would find similar other people and band together.

There are degrees of exceptions for a couple of categories - if you've been educated in the same institution, so the social circle from upbringing can be substituted, or if you're all close to the top of emerging fields so that you could be said to be pushing whole fields forward together. And that stands true whether it's for fun parties in Bloomsbury or for serious work in Los Alamos.

The natural next question is about the types of people that need to be brought together in that particular geographical cluster for ideas to start pinging across each other's brains.

It could be that the smartest people, with the highest potential, automagically gather in the same place through a process of mystical osmosis. Let's call that Hypothesis A.

Or possibly it could be that the formation of the cluster is what transforms people into geniuses. Being a part of a group makes B players look like A players, and A players look like A+ players.

Speaking in clichés, the first is a "build it and they will come" sort of belief, and the second is "it is what you make of it". And that remains true regardless of the fact that you can only really recognise these things in hindsight.

Fairchild Semiconductors didn't know that they were sitting on a goldmine of talent necessarily, except looking back a few decades later and being flabbergasted. Did everyone in Paypal have unerring insights into how awesome they all were, or did that only become apparent once Elon Musk started catching rockets falling from the sky? Did the people in the Renaissance period clap each other on the back as new theories and artforms kept emerging, exactly as they had predicted?

Since going to school with Von Neumann and co would necessitate the invention of a time machine, and he's sadly not around to invent one for us, it's difficult to identify where there might be a cluster to become a part of. There seem to be two clear examples though of places and times when it was actually recognised in time, and others could take advantage of it were they able to look hard enough. Not clusters per se that you could become a part of, but cluster creation entities where you could move to increase your odds of being in one.

One is Silicon Valley. Every year its hype seems to be growing faster than reality, but that exponential curve dance is exactly what causes smart and ambitious people to move there to try things. There's a reason the Collison brothers and Zuckerberg and Larry and Sergey were all tempted to go there. And even once it was fully recognised that this was a cluster-creation-center, you could still move and try to randomly bump into or join one. It might not be as stratospheric as joining the Paypal mafia, but it's also better than trying to go it alone in West Virginia or Hamburg or Abuja.

It is also worth marvelling at the fact that its been a known center of innovation for more than half a century now. It's had a few booms and bubbles, and there still are people present who remember them. It is incredibly weird when you think about it, but the guy who helped invent the web browser to, you know, browse the web, runs a VC. You can pitch him an idea and get funding. How much closer to the center of a cluster can you get?

The second, in exactly the same way, is New York for finance. In an odd way it's a much narrower cluster, since unlike technology or innovation, finance is a fairly well defined discipline. But if you're smart and ambitious and want to go into financial services for whatever reason, you know where to go. You know where you'll have the best chance of being a part of a group of few select smart, ambitious people who want to do something together and potentially create a cluster.

VI

It's unclear that there is one clear takeaway from this. We've learnt that specific inventions can often happen simultaneously in different places, that innovations at a larger scale often happen in clusters, and that the creation of clusters can happen within a meta-group arrangement like cluster-creation-centers like the Valley.

To get there there needs to be enough giants to step on. Whether that's in the form of simultaneous progress in a bunch of areas so new things can be put together, or bringing new ideas that can be tried and applied in new sciences.

The innovation landscape is more like a city of spires standing tall together rather than a solitary tower. The challenge is to find and encourage the formation of more clusters well in time where further expansion and attraction of talent can happen.