Aa realisation I'd had a few years back was that networks -- and here I'm talking about any organisation of components with flows, not just data networks, but people, organisations, cities, companies, markets, transport, conversations or discussion boards, etc., -- have several characteristics which govern behaviour. These define a large class of hence related dynamics.

There's topology: (peer, star, chain, ring, mesh, tree, , web, complex...). There's scale: null, unary, pair, triple (largest size where links equal nodes), 4 (first where links exceed nodes), etc. And the scale effects discussed here resemble those I've noted.

At large scale, scale tends to domininate topology in large part as there's only one viable topology, the dendritic tree (this is West's bugbear).

There's also network depth -- the number of nodes between pairs, particularly on average. Even relatively shaallow depths can have huge significance, think Six Degrees of Separation (or Kevin Bacon): this gives you everyone within a substantial industry (cinema) or even the world (Facebook, national security agency threat matrices).

Related is the characteristic of specific individuals -- superstars and regression toward the mean. A Gresham's Law type effect means that the effective functional level of a group is set by its least capable rather than most capable members (absent some means of effective management or moderation), another tendency which favours smaller groups initially.

It's also worth noting that certain transition points, such as HR being required as headcount climbs above 50, are determined by regulatory requiremment -- beware outside influences in anacdotal observation.

A particularly pervasive similarity comes to mind between agriculture and media, and how scale, topology, node characteristics, and complexity result in conceptually similar activity models. This is, of course, broadcasting. Named for the farmer casting seed on fields, the analogy continues when the harvest is considered: the mass-gathering of commodity high-utility carbohydrates through a uniform and undifferentiated processing. In media, this is advertising. And the results are similar: focus is on nondifferentiation, uniform processing, and maximising harvest yield, not the field's (or audience's) quality experience.

It's also interexting to note that even before the term was adopted there was a trend to more targeted cultivation in ag, including Jethro Tull's seed drills (and those of China long preceeding him), pressaging moves to media segmentation and targeting.

It's also possible that there are other ag analogies which might prove useful or insightful: herd-tending (prevalent in religious contexts), orchards, forestry, foraging, hunting, fishing, vegetable vs. staple crops, reserving methods such as fermentation, etc. Food for thought, as it were....

https://www.nytimes.com/2017/05/26/business/dealbook/geoffre...

This was my immediate thought as well -- but by mentioning Scale you must also link the excellent review of it by Freeman Dyson(!):

http://www.nybooks.com/articles/2018/05/10/the-key-to-everyt...

Another point is that it's easy to mistake a power law for a sigmoid and vice versa.

There's remarkably little study of technology, in the general sense, as its own thing. There are a few good book,s, W. Brian Arthur's especially, and numerous ones that strike me as oversold (Kevin Kelley's strikes me this way).

Arthur, like West, is affiliated with the Santa Fe Institute, and it strikes me that much of the most interesting work I've seem in recent decades comes from there, see alse Farmer, Holland, van der Leeuw, Krakauer, and others.

There's a philosophy of technology (good overview at http://plato.stanford.edu) but it focuses almost exclusively on socialy and ethical interactions, and not what technology is or its mechanisms. Much of the literature is quite woolley (Heidigger, Elul), though not without merits. Others (Schumpeter) seem to me so far to miss a critical essence. Often its the study of failure which seems most instructive (Perrow, etc.).

I'm not quite sure what to make of your comment on companies: yes, large-scale networks do somewhat tend to resemble one a nother, and no, a given network maw not be exclusive to a domail (think: water, electricity, gas, comms, data, sewage, transport, social, financial, all or most of which likely connect your residence). At scale, the interesting questions tend to relate to interplay between various networks.

What's your glass metaphor indicating? Brittleness? Transparency? Amorphous state? Other?

Siqmoid and power laws strongly resemble one another, until they don't. They're also not the universe of growth and decay functions.

Using a Presidential analogy, it advocates a style much like Reagan. Founders should focus on vision and being the flagbearer of that standard. Then have the cleverness to delegate much of the responsibility for keeping the ship afloat to talented individuals without regard for ego. As a general process, seems reasonable. Probably less stress, more focus on important tasks. To an extreme, you become detached and unaware of the undercurrents within your firm. The vision can wander because it is decoupled from reality on the ground.

On the counter side, I would argue there is value in being more akin to a Clinton style figurehead. Fairly well known for keeping abreast of everything, to the point of annoyance. Leaving people people free to do what they do, ie, not micro-managing, but being aware of what is occurring - remaining engaged. To an extreme, you are obsessive and obstructing rather than staying informed. High stress, and not enough time maintaining vision and outlook.

Needs both, and fundamentally disagree that we as entities are limited to a "linear" growth. It may just seem linear as you approach the event horizon. To those far away, your change is remarkable.

In Australia, P.M. John Howard (1996–2007) slowly centralised a great deal of policy making power by moving it from departments (each headed up by a minister) to his own office/department, PM&C.

Clearly he was capable of the demands this put on him, and was able to construct a Cabinet (inner-circle of top ministers) to help in sufficiently.

Since 2007, this change seems to have destabilised several governments by allowing a micro-managing figure to bite off more than they could chew, infuriating their own party, who then threw them out, partially because of internal stresses on the party strucutre. Particularly the first Rudd gov. (07-10) and the Abbott gov (13-15).

This is a very specific interpretation however, other Australians may well disagree.

They can be highly beneficial and efficient, rapidly responsive to change, like a sword through red-tape, or oppressive and controlling. Largely dependent on the character of the captain of the ship. And viewpoint. Caesar was loved by many, killed by a few. Lincoln, J.F.K, Gandhi, similar.

And they can often be destabilizing when they end. Even if not by such drastic means as above. A classic problem of monarchies that if you had no children or incompetent children, your gov't was in trouble.

Not that pure republics and democracies with strong decentralization are a whole lot better. Just different issues. Why they oscillate. And even those often function best with a strong figurehead.

Possible: Yes, I guess. Probably: No. @3-10 people when growth is at or above expectation then the staff are almost always loyal to their invested interest unless there is a very big scandal.

I see a lot of Military leadership or Military strategy ideas referenced around the startup world as valuable lessons. Yet actual Military members or Veterans, outside of some "hiring veterans" efforts, aren't held in particularly high regard in the valley.

Why the Delta?

A) Biology is not inherently linear, at all. If I had to pick a single scaling function to identify with biology, I would pick the exponential.

B) "Biologically, we are sprinting beings, not marathon beings." is utterly untrue. We evolved as persistence hunters, we're woefully ineffective sprinters. Basically all large animals are dramatically faster than us.

Good article, but man do I hate woo bad analogies like this. It's almost as bad as the guys that talk about "Quantum" crap when I'm just looking for decent meditation help.

If you haven’t read about cursorial hunting, the Wikipedia article [1] is worth the time. Humans are one of like five known persistance hunters. Our ancestors (and certain Kalahari contemporaries) would literally chase their prey until they died of exhaustion.

From the perspective of the prey, it is uniquely horrifying. You outrun your predator. But every time you stop to catch your breath, before you’ve had a chance to cool down, they reappear, in the distance, on the horizon. Each time nearer. Hours or even days on end. Repeatedly until, finally, your legs or heart give out.

[1] https://en.m.wikipedia.org/wiki/Persistence_hunting

Reminds me of the film "No Country for Old Men". One of the most thrilling movies I've ever seen.

Basically after a few days the poor horse gets totally exhausted and just stops.

Might make a claim that a human walking on two legs is especially efficient use of food energy.

To increase the amazement, supposedly the US military has discovered that on really long exercise, say, all-out exercise over several days, say, walking 100 miles through deep snow, women have more endurance than men. The explanation is that the women start out with more fat reserves and are able to use them. So, too soon the men run out of fat and, thus, out of energy.

So, maybe in running down a horse, African zebra, or antelope, female humans might be better than male humans!

Some horses do have tremendous endurance capabilities, but not many. The average companion-horse probably would not comfortably go 56km / 35 miles in one day. The average rider would find 35 miles too far, as well!

The Wikipedia article referred to up-thread talks about the Human vs Horse race being 35 km / 22 miles, with the humans on foot having a 15 minute head start.

I always assumed those movie sequences where a bunch of folk on horseback go after a person on foot, who had a few minutes head start, were a dramatisation. Turns out it's probably more realistic than I thought, with the person on foot having the advantage of being able to traverse terrain a horse would have to go around.

1. https://www.quora.com/Horseback-Riding-How-long-can-a-horse-...

almost nothing grows exponentially unless there's a recurrence relation or difference equation involved.

http://firstround.com/review/hypergrowth-and-the-law-of-star...

see in part and relevant to the title of this thread

> “Human beings grow biologically and linearly. A year from now, you will be a year older — there's no growth hacking we can do to make that happen faster. Even if we were looking at the metrics of ‘you’—the age of your bones, your height, everything — you're not going to grow exponentially. You're going to grow linearly because all biological systems do that,” says Halim. “A company, which is a collection of biological systems called humans, can grow exponentially. Especially in tech, companies exist in a world in which you can be serving 100 customers one day, and a million a year later.”

Want to find the root of a function? Under mild assumptions, Newton iteration gives accuracy that improves as an exponential function of the number of iterations. This situation is especially well known and easy to see for finding square roots.

I will make an argument the other way around, that it's humans who can grow exponentially (for a while) and companies grow at best linearly!!

Okay, the first example is Leo Szilard walking across a street and then soaking in a bathtub: He thought, one neutron hits one uranium nucleus which emits two neutrons, ..., four neutrons ... and after positive integer n such generations we have 2^n neutrons emitted. So, that was exponential growth, all from just Leo walking and later soaking in a bath tub.

Then for an organization, at Oak Ridge the were getting about, what was it, 10 grams a day of separated uranium. So, for 100 grams a day, have 10 times more working units, that is, linear growth.

Then when it appeared that maybe one approach to uranium enrichment might not work, try three. Linear growth.

When it appeared that uranium might be too tough to get, try plutonium. Work with two elements instead of just one, double the number of elements, people, facilities, money, results, etc. and get two paths to success instead of one. All linear.

Then Edward Teller had another exponential idea: Let the X-rays from fission squeeze deuterium for another essentially exponential growth rate.

For another example, Alan Guth wanted to do something so dreamed up the big bang, where exponential growth seems even an understatement. So, Guth had exponential growth.

Since then various organizations have gotten data that confirm what Guth thought, and the growth of those organizations have been essentially linear, that is, twice as much money, people, time yield twice as many new results, e.g., the 3 K background radiation and accelerating growth of the universe.

For more, for the traveling salesman problem in Euclidean spaces, R. Karp had an idea -- on problems with positive integer n cities, for any probability p < 1, and for any epsilon > 0, a simple application of minimum spanning trees will as n grows, with probability greater than p get feasible solutions within epsilon percent of optimality. So, as n grows, the number of traveling salesman tours grows exponentially, really n! (use Sterling's approximation to get an exponential), and Karp's idea was exponentially powerful, that is, the minimum spanning tree effort grows only as a polynomial in n but beats the exponential challenge of the problem.

For another, linear programming commonly has the number of feasible region extreme points growing exponentially with problem size, but the polynomial algorithms can find the possibly unique extreme point in only polynomial time. If divide by the polynomial, then get linear effort, and the exponential number of extreme points divided by the polynomial is still exponential. So, the polynomial algorithm idea, one of them from the head of just one person, was exponentially powerful.

For another example, Moore's law, so far still exponential but from the brains of mostly just a few people and brought to market by essentially linear time, money, people, etc. Want twice as many chips? Okay, that linear; build two fabs instead of one and get twice as many chips -- linear.

For another, Alexander Fleming looked at some bread mold and how it fought off some bacteria. Presto, bingo, he had penicillin that saved exponentially many lives.

There are lots of other examples where people grew by having exponentially powerful ideas while organizations were limited by twice as much in people, time, and money gives at best, twice as much revenue, maybe less due to the quadratic growth of the overhead of internal communications as the number of people in the organization grows.

Net, the big stuff, the stuff with exponential power, is from people, commonly one or at most only a few people, where a company to exploit the exponentially powerful work grows only linearly, that is, twice as many people, etc. in the company result in twice as many customers served.

https://news.ycombinator.com/newsguidelines.html