Rough working notes on scenius.
The artist Brian Eno has introduced a notion of collective intelligence that he calls "scenius":
"Scenius stands for the intelligence and the intuition of a whole cultural scene. It is the communal form of the concept of the genius."
As an example, conside the town of Florence in the 15th century. At the time it had perhaps 100,000 inhabitants, whose number included Michaelangelo, Botticelli, Masaccio, Fra Angelico, Donatello, and a host of lesser but still great painters.
How remarkable a clustering of painters was this?
Numerical attempts to value creative work are always arguable, and I don't want to get caught up in arguing over details. However, let us conservatively grant that several of these painters could be listed among the 100 greatest painters of all time. For example, four of them appear on this list. A modern country of, say, 100 million inhabitants, might, if it were lucky, produce more than four painters of such quality during a century. But it is almost inconceivable that it would produce (say) 40 such painters.
In other words, if you were an artist who aspired to be considered a great painter, you upped your chances by a factor of perhaps a thousand by the good fortune of being born in Florence in the 15th century, rather than (say) in North America in the 20th century.
We often think in terms of people such as Michaelangelo, Botticelli et al being born with extraordinary, one in a million (or more) levels of talent. But the lesson of Florence is that this is wrong. These people were more like one in 10,000 types of talent. Indeed, they were probably far less: after all, painting was closed as a career to the great majority of Florentines, most of whom were of the wrong gender, or too poor, or faced some other near-insurmountable barrier.
What truly set Michaelangelo et al apart wasn't talent. It was their great good fortune in being in the right place at the right time. Their genius derived from scenius.
Can we really claim that scenius caused Florence's 15th century outbreak of artistic genius? Maybe some other causal factor was responsible?
To see what I mean, consider a case where it seems as though scenius was occurring, but upon closer examination that explanation seems unlikely. The example is the project to build the atomic bomb, carried out in the small town of Los Alamos during World War II. Living in town at the time were scientific luminaries such Feynman, Fermi, Bethe, and (on occasion) Bohr and von Neumann, as well many other scientists and engineers of great distinction.
This is a galaxy of talent not unlike Florence in the 15th century. But, of course, the environment in Los Alamos did not create such talent. Most of the luminaries in town were already exceptionally accomplished before going to Los Alamos. Perhaps the environment did make some positive difference to their future accomplishments, but it seems unlikely that it was primarily the environment which caused their future accomplishments.
By contrast, the great painters of Florence named above were all born in Florence, or in the countryside nearby. And so it really was living in Florence during the 15th century which caused these people to become better painters.
Another way of countering the scenius hypothesis is to wonder whether ordinary statistical variation might explain the existence of a town like Florence. After all, the world has contained many communities of 100,000 people through history. And some of those communities will contain an unusually large number of excellent artists, just by the luck of the draw. Maybe the reason for Florence's success wasn't scenius, it was luck. Some community had to have the most great artists, and maybe it just happened to be Florence.
In the appendix I use a statistical model to test the idea that Florence was just lucky. In particular, I test the genius hypothesis, the idea that what makes a great artist such as Michaelangelo is some intrinsic property of the artist, not the environment in which he found himself. According to this hypothesis, Michaelangelo could have been brought up in another small town of 100,000 and, provided the town had some sort of artistic community, he would still have matured into a great artist.
What I show in the appendix is that it's not possible to account for the creativity we find in Florence, even by the luck of the draw. Instead, we must take seriously the scenius hypothesis, i.e., the notion that while intrinsic properties are important, the environment in which a creative type works may be far more important.
Imagine it is the 18th century. You are living in New York and you hear the first news of a major gold strike in California. Many people rush out to California, dig for gold, and an unusually large fraction become wealthy. Should we conclude from this that there was some special scenius in California that imbued those people with a special talent for generating wealth? Were they picking up special wealth-generating skills from their friends and colleagues?
The answer, obviously, is no, it would be wrong to attribute the wealth to anything particularly important in the social environment. Rather, they chose to put themselves in the right place at the right time.
You can make a similar case for Florence. There is a sense in which the 15th century Florence was a rich frontier waiting to be mined for great art, not unlike the mining of gold in the California gold fields. At least three factors contributed to Florence being such a frontier: (1) it had been a leading European center for art for hundreds of years; (2) the rise of the Medici family created a business model to support a large number of artistic endeavours; and (3) the ideas, techniques and technologies of art had advanced to a point where they opened up a whole new world of artistic possibilities.
Any one of these three facts would not have been enough to establish Florence as pre-eminent not just in its time, but arguably for all time. But with all three factors occurring simultaneously, Florence became the place to be for ambitious artists.
This explanation for Florence's pre-eminence is subtly different from Eno's explanation of scenius. It's not that the great painters of Florence were learning and benefiting from each other. Instead, in this explanation they were more like gold miners who just happened to be first on the scene, just as a new vein of gold was found.
This is what might be called the rich frontier hypothesis of genius: there's a huge advantage to getting in very early as a rich frontier is opening up. There are relatively few other people present initially, and so not only pioneers stake a major claim, they're often able to remain ahead of the curve for many years or decades, even as the field matures.
So now we have an interesting question. How much were the great Florentine artists benefiting from being at a rich frontier, and how much were they benefiting from a rich social environment? I don't think it's at all clear what he answer to this question is.
How does this affect our notion of scenius?
There are two paths we could take at this point. One is to adopt a narrow notion of scenius as the benefit a creative worker gets from their social scene. I'm prety sure this is how Eno intended it. If we adopt this notion of scenius then it explicitly excludes the rich frontier. Any attempt to explain Florence and similar events must engage the question of the relative contribution of both scenius and the rich frontier.
The other approach is to embrace a broader notion of scenius which includes both the benefits from working at a rich frontier and the benefit from a rich social environment. In this essay, I'm going to adopt this broader notion of scenius. I'll still occasionally find it helpful to break out both components of scenius --- the rich frontier and the rich social scene.
At this point these really do turn into disconnected notes. Still, I hope they each make sense individually.
What am I trying to do in the essay? I'm not sure. Partially I'm just exploring a fascinating phenomenon for anyone interested in human achievement or human institutions: the clustering of talent. Why does it happen? What does it mean? Are there interesting hypotheses about the clustering that we can rule in or out?
Partially I'm interested for much more personal reasons. I'm trying to figure out what I should be doing. There are two problems with this: (1) it's very personal, and it's hard to write about personal stuff; and (2) it's incomplete: am I writing about scenius, or about how to choose work? Because of this there is a natural pull to make this merely part of a much larger and more ambitious essay on how to choose work. But that feels like more than I can or want to bite off at present.
Partially I'm interested in the question as an institutional design problem.
I'm left in an uncomfortable position: I have what seem like many interesting observations. But I don't have a natural frame to put them in.
One possibility would be just to stop, to make the first part of the essay, above, the entirety of the essay, with some re-organization. But I feel like I can do better.
Another possibility: just go with something cheap: "Five signs of scenius" or something like that. Or maybe "Five signs you haven't found scenius". I'm perhaps trying too hard to find a narrative thread, and should instead just tie things up, and let it go.
Why do I feel discomfort with the whole essay? I've talked a bit about this above. But there are more reasons.
There's a best-selling self-help book by Steven Covey entitled "The Seven Habits of Highly Effective People". It's about what you'd expect --- advice on how to live a better and more effective life. Linus Torvalds once wrote of Covey's book: "XXX [random insulting remarks here]".
Torvalds' comment sums up the attitude many technical people have about books on "soft" human-centred subjects, like how to live your life. We don't have a science which answers questions such as how to spend one's time, how to deal with your spouse, and so on. Of course, such questions still matter, and so self-help books such as Covey's rack up millions of sales.
Intellecturals sneer at such books as witch-doctory, but they can't point to obviously superior solutions. There have been attempts by psychologists to colonize this field --- many write books using the fact that they're REAL SCIENCE as a marketing credential --- but the results are often oversold, the evidence and reasoning flimst. In one or two hundred years I expect we'll genuinely know a lot about these sorts of questions. But for now, these "scientific" self-help books don't seem so much more trustworthy to me than any other.
The problem is that just because we don't have a science of how to live, doesn't mean we get to opt out of questions like how to live a happy and interesting life. The right question isn't whether Covey and his ilk are witch doctors. Rather, when you only have witch doctors to choose from, you want to make darn sure you find the best.
(Parenthetically, I've occasionally talked to the intellectual sneerers. I've noticed a few odd things. When discussing questions like how to choose work, they'll often make extremely black-and-white statements supporting some obvious conventionalism. When pushed on why they believe those things, they often respond with sneers, or with flimsy or non-existent evidence. Their commitment to evidence is often thin, disguising a discomfort with how to deal with not knowing.)
In any case this essay deals with such soft issues, and because of that is making me feel some discomfort in talking publicly about such issues. I feel the need to be bulletproof in my discussion, when of course that is not feasible. This kind of thing is necessarily speculative.
Where is scenius nascent today? I don't know where scenius is nascent today. I have speculations: maybe it's in robotics, in energy, in space, in 3d-printing, in synthetic biology, in gaming, in virtual worlds, in self-driving cars, in augmented reality, or in the maker movement. Probably it's in areas edgier than any of these, areas I've never heard of.
But identifying promising subject areas is not the same as identifying high-concentration communities already doing great work, or which (even better) will soon start to do great work.
Part of the problem is public relations. It's easy to look at the pubicity coming from an organization like SpaceX --- XXX, wow! Any maybe it really is a great place, a place that is enabling its engineers and scientists to grow by leaps and bounds. Equally well, maybe it's all a few brilliant engineers, and everyone else is a drone. I don't know.
High-prestige, fashionable, or rich fields and organizations: You see many magazine articles, books written about fashionable fields: the web, energy, space, and so on. And you see lots about organizations such as Google, SpaceX, the MIT Media Lab, and others. My empirical observation is that the density of stimulating people to learn from in these fields and organizations is often not as high as publicity suggests. Sometimes, it's quite low. A possible reason is that these are fields in which it is relatively easy to do well, and so attractive to lots of people who want to coast. When money is pouring into a field, so will the people who just want to take the easy way. When prestige or fashionability is pouring into a field, so too will the poseurs.
The counter-argument is that lots of very smart, very determined people will also be in these fields. My gut feeling is that while this is true, the density becomes relatively low. Not as low as in many other fields, of course, but it's not amazing: it's not Florence in the 1400s, or Los Alamos in the 1940s.
Is the internet still a frontier? People have been saying the web is over almost since the web was begun. My instinct is that this is wrong. Think about te huge problems that remain to be solved: we're a long way from every object and every event in the world being networked; we are drowning in co-ordination problems; we're drowning in data analysis problems; we're a long way from artificial intelligence. Maybe there's insurmountable bottlenecks that will stop progress toward all these goals. But I'd be surprised. One thing that shocks me especially: there's so little work on new internet protocols and new clients. Danny Hillis made the observation that the web --- i.e., the http protocol --- is the "slime mold of the internet", and that seems exactly right.
A corollary to the above is to look for communities that are held together by something other than fashion, money, or prestige: Often, they'll be held together by a shared ideal, or something else that is shared. It may just be shared awareness of a frontier.
An example is the case of Camp 4 at Yosemite. Climbers at Camp 4 will often make a living by collecting and recycling aluminium cans. They'll live off boiled rice and peanut butter. They're not doing it for the money. And they're not doing it because it's fashionable or prestigious in the wider society. Rather, they have a shared ideal of what matters, and they believe so strongly in it that they're willing to live this way, pushing the frontier of climbing.
Is there a rich frontier to mine? Think of painting at the end of the 15th century. Painting at the start of the 20th. Quantum computing in the early 1990s. Physics in the 1920s. The internet (especially the web) in the 1990s. Possibly also today. In each case there were major frontiers on which it had become possible to make rapid progress.
Suppose we assume the genius hypothesis as true, and believe that someone will achieve greatness as a painter with a fixed probability, say p = 1 in ten million, independent of their environment. Then in a community of n = 100,000 people, the probability of having k or more great painters (k a small positive integer) is approximately:
(np)^k / k! = (1/100)^k / k!
When k = 4, this probability is less than one in a billion. And so if the genius hypothesis is true, then Florence is far outside the bounds of what can be expected statistically.
Of course, the outcome here depends a lot on exactly what we mean for a painter to be "great". If we adopt more modest standards, say p = 1 in a million, then the probability above increases quite a bit, to about 1 in a quarter million. This is still too small -- there have not been anything like a quarter million such communities in history -- but is at least getting closer to statistical plausibility. The problem is that if we adopt this value of p, then we're saying that we expect a great painter to emerge from about 1 city in every 10 of size 100,000. While some notable painters will likely come out of such a collection of cities, I frankly think that's far too modest a criterion. So we can confidently reject the genius hypothesis.
Thanks to Jen Dodd, Ilya Grigorik and Hassan Masum for many conversations about these issues.
Feedback: Michaelangelo at some point must have realized that he had an opportunity to be one of the great artists of all time. That would have motivated him to work far harder. The best way to sail confidently is with a wind at your back.
Phil Agre: The information scientist Phil Agre has proposed an interesting way to identify important emerging fields. He suggested asking the very smartest undergrads what they're interested in, and making a list of their replies. Cross the already fashionable answers (cosmology, nanotech etc) off the list. Look instead for fields that you've never heard of, that sound strange, and slightly off the wall. Some of those fields will fail, of course. But chances are that some of biggest fields a decade or two hence will be on that list.
How to recognize scenius?
What's going to be important in 10-15 years' time? Are there high-concentration communities doing interesting work in the area today? Space, robotics, energy.
The adjacent possible:
Teleportation was the adjacent possible for a long, long time: It's tempting to think this is just about low-haning fruit. But it's not. Quantum teleportation is a simple enough protocol that it could have been discovered in the 1920s. But it wasn't. It had to wait until 1993.
What isn't scenius? If something is fashionable or there's lots of money to do it (or both). This causes an inrush of poseurs and 9-to-5ers.
However, there may still be identifiable sub-veins within those locations:
Is Google an example of scenius? Maybe there are places inside Google where that's true.
It doesn't come out of companies: I wonder. Many of Michaelangelo's greatest works were, of course, produced on commission. I think a problem with companies is that people in them often don't get the credit for what they do, and they often have to fit into someone else's tighty controlled narrative. Neither is good for doing excellent work.
Brainstorming questions: How to find scenius? Where is scenius today? How to create scenius? I have a practical question: what impact should it have?
There are two natural questions to ask at this point. If you are still not yet decided on a career, and ambitious to do great creative work, then you should ask how to find scenius. After all, what else increases your chance of doing great things by a factor of 1,000 or more? The second natural question is how to create scenius? Although I have many thoughts about the second question, I am going to focus on the first.
The question then becomes: how to find scenius? And to answer that, the we can ask Where is scenius today? Let's look first at the related question: what are some examples of scenius in recent times? Here's a few that seem like likely candidates:
-
In the 1970s Xerox PARC made crucial contributions to the development of laser printers, the mouse, windowing systems (and graphics more generally), the ethernet, and many more.
-
Pixar: 11 of Pixar's 12 feature films have been critical and financial successes. For the most part they're not to my taste, but clearly Pixar has figured out something which virtually none of the other studios have.
Difficulty in writing about these issues: It's difficult to write about these issues. Questions like "how to choose where you work" or "what should you work on" are challenging questions.
What question do I want to ask in the second half?
Paypal as an example?
Might it only be founders-only? Maybe early-employees only? Is there much evidence for that?
Do you need some selection process for quality? Yes, a ruthless one. Ideally, one that will concentrate the best people in the same locations, but at the same time not dissuade them from being friendly with one another, and supporting one another.
OPen, but with ruthless standards: Or maybe the right way of thinking . It's an old observation that there is no democracy of talent. More important, to my mind.
Problem: Many of the notions at the start seem bland: They are only bland when not set up in opposition to their opposites. E.g., it may seem trite to acknowledge the role of others. But it's really an assertion that innate talent just doesn't matter that much.
The talent myth: This idea of talent just seems utterly destructive. If Michaelangelo had grown up in Waterloo, he'd be a talented local artist, admired within a small sphere, but not regarded as remarkable, and forgotten by history.