Incatena


In your blog you mentioned that “the CEO system for running corporations is a dangerous anachronism.” I was wondering what you would replace it with – in your perfect world, what system would you design for building and running a global business and ensuring its continuity from one generation to the next? What other cultural or economic changes would go along with the new system if it were implemented throughout society? (E.g. would there still be brand-name identification of consumer products?)

—Geoffrey

zorg

Not the ideal

Great question; I’m not sure I have a great answer. But that’s because we need a whole lot more experimentation. Anti-monarchists couldn’t necessarily describe parliamentary democracy in 1700, either.

First, let’s review the problem. In that post, I pointed out Trump as the epitome of the terrible CEO. He’s a lazy, incurious person who’s used to unquestioning obedience for his terrible ideas, and takes every reverse personally. But he’s not much of an outlier. I’ve met plenty of much smaller-level CEOs, who have the same arrogance and inability to understand their own business. The good CEOs I’ve met are generally the ones who know when to get out of the way of their own workers— the people who actually know the work and know what needs to get done.

Many people are in love with the idea of the strong leader of uncompromising vision. We’re not so enamored of the Louis XIV or the Napoléon these days, but Americans, at least, still admire the entrepreneur who builds up a company from nothing: Bill Gates, Steve Jobs, Henry Ford, J.P. Morgan, John D. Rockefeller.

I don’t necessarily want to touch that.  On the whole, society benefits when these people build something new.  They should obey labor laws and such, but we don’t need to throw out this part of the system. What we should throw out is the notion that their heirs, or some guy with an MBA, deserve the same deference. If you worship wealth and never address inequality, as Piketty demonstrates, what you get very soon is an aristocracy of rentiers, people who never created anything yet sit on all the accumulated wealth, and whose primary incentive is the maintenance of their own comfort.

So, how do we run a company instead? Two basic approaches:

  • Put someone in charge temporarily
  • Put a representative committee in charge

For now, this comes down to your view of human nature. If you think there needs to be a strong leader (even if his power is limited or we can get rid of him), then you pick Option A. If you distrust all single sources of power, you pick Option B.

Perhaps ironically, in theory option B is what we have now, once the founder is dead. Public corporations have a board of directors, and even rules that a good number of these must come from outside the company; and they pick the CEO. In the US, the board represents the stockholders (i.e. the owners), and it is very possible for them to throw out the CEO.

There are two problems with this.

  • One, boards are in practice not very independent-minded. They’re often friends of the executives and only meet for a brief period every few months.  They’re not deeply involved in the business nor inclined to second-guess the CEO.
  • And two, what the stockholders want is basically more money right now.  This may be completely opposed to the interests of employees, customers, the community, and even investors who are thinking beyond the next 12 months.

A minimal reform is to require that some of these other stakeholders get representation. In Germany, up to half of the board represents labor, though stockholders usually get the determining vote. The idea could be expanded by giving other stakeholders representation.

There are many interesting experiments in corporate governance today, such as:

  • Valve, the company behind Half-Life and Portal as well as Steam, is famously run on near-anarchic lines. Employees literally pick which product they want to work on— and if a product doesn’t attract workers it doesn’t get made. (One does wonder if this is why they never seem to get to release 3 with any of their products.)
  • The first company I worked for is now employee-owned.  It works out well for them, and they were able to weather the 2008 recession in part because they didn’t have the huge expense of a CEO’s salary.
  • NFL management is said to be a huge swamp… except for the Green Bay Packers,  owned by a huge mass of individual Wisconsinites. They’re competitive with other teams and the team stays in Wisconsin.
  • The Mondragon Corporation, of Spain, is a remarkable co-operative firm which employs over 75,000 people.
  • We actually have plenty of examples of non-monarchical institutions: churches, universities, co-ops, many arts organizations or activist groups.

We need more experimentation to see what works. The answer to a lot of objections is going to be “Maybe. We don’t know. We need to try things out.”

The obvious worry is that discussion and representation take time, or at worst tie the entity up in knots. Democratic politics is not exactly known for calm, civilized consensus. In response to that—

  • Again, monarchs suck. We put up with the inefficiencies of democracy because one-man rule really is worse. But the inefficiencies are definitely there.
  • People get better at democracy when cultural norms evolve to support it. I’ve been in endless, unfocused meetings— people flounder if they don’t know what they’re doing. That isn’t a condemnation of the system; it just means that the transition is tricky. People who are not used to power do not automatically know how to use it.
  • There are better and worse decision-making techniques. A huge, completely open-ended meeting is one of the worst. People are better at reacting to concrete proposals than they are at coming up with them. If a proposal is rejected, it’s better for a small group to take it offline than for a large meeting to attempt to redesign it. The group needs a way to table arguments, so it is not dominated by a few eristic individuals. And so on. Heuristics will develop to smooth the process considerably.

One big caveat: democracy is not a cure-all. I think we’d be happier if we could vote who runs the company, or at least vote the current bums out. But that’s not the same as saying we’d be happy.

On the other hand, looking at modern representative democracy, we have to remember that it’s optimized for the logistics of 1790. I’m sure we could do better. One big problem, for instance, is the bundling of policies. At the federal level, there’s no way to say that you want (say) more health care and less immigration.  You can only pick between the two major bundles that are offered. Maybe we need to vote on policies more than on leaders.

Your question on brands is just part of a much bigger question: the optimum size of corporations.  I’m sure a bunch of readers are eager to tell me that the problem is not how to fix corporations, but how to get rid of them. But leftists have, to my knowledge, only come up with a couple alternatives, and they’re contradictory:

  • Nationalize them. So the organizations become massive.
  • Have workers run enterprises directly.  So the organizations must be tiny.

In general, the first approach makes the problem worse. (Some public goods should be nationalized; but I do not want a government commissar, or even a People’s Soviet, deciding what books can be published or what crops can be grown.) And the second approach is largely untested, and of questionable utility for a planet of 7.5 billion people.

It might be nice if every firm was only the size of a village— 150 to 200 people.  But there is such a thing as economies of scale. Really big firms can grow corn, build computers or airplanes, and make action movies really efficiently.  A world of small firms is also, very likely, a world of high prices for consumers. There’s also the problem of competing standards. These should never be a monopoly, even a government monopoly.  And yet it’s kind of a nightmare if you have a hundred competing standards rather than two or three. And if you’re eager to break up Google, do you also want to break up Mondragon?  (They’re about the same size.)

Plus, there’s the inconvenient fact that large firms are far easier to regulate, and can be far more progressive.  A corporation with 75,000 employees can have a professional HR department, great worker amenities, and a commitment to diversity. They’d also be far easier to democratize.  Smaller firms are often run by the most conservative, cranky old despots.

Ideally we should be able to choose both options. Restaurants, for example: it’s nice to have a one-chef gem of a restaurant.  It can also be convenient to have a known brand where the type and quality of the food are predictable (and prices are cheaper).  In art, it’s great to have quirky one- or two-person projects; also to have behemoths that require hundreds of people working together.

Anyway, the one thing I’m certain about here is that future economics is going to surprise us. The modern corporation emerged only 150 years ago, with the invention of the telegraph.  (Adam Smith thought corporations were of limited utility.)  As late as the 1960s, the ideal of corporate governance was a class of professional, medium-income managers hired by the owners; the cult of rock-star CEOs paid in the megabillions was a (stupid) innovation of the 1980s. Right now things look kind of dystopian, but that doesn’t mean we’re stuck here.

Advertisements

One of the muses finally spoke– I’m not sure which Muse has the portfolio for science fiction.  Anyway, I suddenly have the plot for a new Incatena book.

Well, “plot” is too strong a word. “Predicament” maybe.

Areopolis-MapMorgan decides to quit the exciting but inconvenient life of an Agent, and on a whim decides to go back to Euko Teknik, in the α Centauri system, which is having a reunion for its alumni. The ex-Agent has changed sex (to whatever it wasn’t before) and had a half-mind-wipe in order to put diplomacy and espionage behind. But as we know, you can never really get away from the profession, especially in a spy novel. The boss activates some overrides in Morgan’s neurimplant. One more mission.

α Centauri, by the way, has the somewhat clunky traditional name Rigil Kentaurus. But it turns out to have the name Toliman too, and I’m considering using that. When you’re actually there, you probably don’t want to say “α Centauri A” every time you refer to the sun.

Anyway, the system contains two Incatena planets, Euko and Novorossiya, which have fought a few wars before. They have entirely different approaches and values. Euko is all about human or transhuman potential– they want to explore every possibility and rework humanity to match, and as  Euko has no native ecosphere they can rework the planet too. Novorossiya is into embracing our primate heritage, recreating the ancestral environment in the planet’s mixed alien/earthly ecosphere, and keeping a small technological footprint. (With Incatena technology, you can make your planet look like a jungle if you want– neurimplants are invisible and the high-tech infrastructure is only visible when you need a piece of it.)

But they each have their own planet, so it was not obvious how to get them into a major conflict. Finally I thought of something: The α Centauri is a double system (triple with Proxima, but it’s very far away), and orbits within it may not be stable indefinitely. So let’s say the system is getting unstable– perhaps a passing brown dwarf has destabilized it… only it turns out neither planet wants to take action to fix it. Thus Morgan’s interrupted retirement.

Now, I haven’t actually written a word yet, and it’s in line after a couple of other books anyway, so the two or three of you who’ve read APAF will have to wait a bit. But at least now I have a situation and not just a setting…

PZ Myers has a posting where he makes a short argument against transhumanist uploading.  This was relevant to my interests, because I think uploading is bonkers.

He has two arguments, really.  Unfortunately one (using entropy) is just wrong: entropy doesn’t prevent complex systems; it only requires that more entropy be generated to offset them. So long as you convert only a tiny fraction of the universe into computronium, entropy won’t stand in your way.

His other argument was better. but sketchy: uploaders prefer “what is good for the individual over what is good for the population”.  As he was arguing with Eliezer Yudkowsky among others, this is probably a misfire– judging from his Harry Potter fanfic, Yudkowsky does consider it an imperative that technology benefits everyone.

Still, there’s the germ of an actual good argument in there: that the uploaders think way too much about personally not dying, and way not enough about how to make what life we have worth living.  Morally, it’s hard to argue that our biggest problem is that people don’t live 1000 or 1,000,000 years.  If humans keep on with the sort of behavior and morality and economics they have right now, such lifetimes would be hellish.  Even if you have a wildly optimistic view of how well we’re doing, prolonging lifetimes even to a couple hundred years would be horrible for 90% of the population, and that’s assuming we can even keep our civilization going.  (If you want to live forever, climate change is not your grandchildren’s problem, it’s yours.)  So even if you want immortality, you’d better prioritize, well, almost everything else.

But that’s a discussion for another day.  I was caught up short by this comment, by one Gregory in Seattle:

There is a growing belief among memory researchers that the brain relies on “archetypes.” You actually have only one or two physical memories of the taste of bacon: all of the apparent memories of bacon link back to them. REM sleep is when the brain recompiles, tossing out actual memories from short-term storage and integrating the day’s experiences into long-term storage with heavy object reuse (pardon the computerese.)

According to this model, children learn faster because they have fewer archetypes: they are building a “library” and links into them are pretty straightforward. As we get older, though, the ability to store and link novel information becomes more difficult and memory begins to ossify. Someone who pursues life-long learning can stave this off, but not completely. To use another computer example, the problem does not appear to be one of storage so much as the storage becoming fragmented. The ability to link begins to suffer, and memories begin to get lost in the shuffle.

Without a major redesign of how the brain stores memories, very long lifespans will probably bring us to a point where novel experiences cannot be integrated at all. We see this sort of slow down in people who are 90 and 100; I cannot imagine what it would be like for someone who is 200, much less 500 or 1000.

I’d never heard about this theory, but then I don’t know anything really about memory research.  But it’s a fascinating idea, and one that makes a lot of sense as a way for a creature of limited brain to organize the reams of sensory data that swamp it daily.

Though it’s not so much an argument against long lives as an argument that if we want to have them, we’ll have to change some basic facts about ourselves.  That’s why, in the Incatena, I have people doing a kind of brain reboot every century or two: throw out a bunch of memories, loosen the connections, re-adolescentize the brain.

To put it another way, your basic personality, attitudes, ideology, politics, etc. are generally pretty well firmed up by the time you’re 30.  You can adapt to new things after that, but with increasing difficulty– by the time you’re 80, you’re a curmudgeon who hates the kids’ music and clothing and votes for reactionaries.  That’s acceptable when lifetimes are 90 years, but not if they’re 900.  If you refuse to die, then you have to do something to regain your adaptability, for your own benefit and for that of society.

I suppose every sf writer worries that something is going to come along to invalidate their vision, fictional though it is.  I’ve been having some worries lately about whether the sort of espionage pictured in Against Peace and Freedom will actually be possible in an info-saturated world.

We already live in a world where multiple entities have data about you that would be the wet dream of a totalitarian.  And it’s not even the government.  Facebook knows your whole social network, and peeps on what other sites you visit.  Stores know what you buy down to the last bag of candy.  Games give feedback to developers on how they’re played.  Cel phone conversations have been externally monitored.  Google has created a wildly detailed map with annotations– where you can’t turn left, what paths are suitable only for foot traffic, where the KFCs are.

Let this system develop for fifty years.  Would it then be possible for an Agent to come into the country, evade the secret police, and meet dissidents?  I’ve suggested that externalities can be reduced by low-level monitoring– bullets know who fired them, pollution is chemically marked to show its source.  In this kind of info-rich society, a human being is like a bull blundering through a dollhouse– there’s no way to miss their trail.

Now, the first thought is to minimize the clues.  This might be barely possible today: pay cash, never use a computer, use other people’s houses and cars.  But will cash even exist in 50 years, much less 2500?  And retail is already being transformed; it may not be long before it’s strictly impossible to buy a plane ticket without going online.  Besides, a suspicious government might well slap a nanobug on every traveler to add to the data trail.

Surveillance can also look for missing data.  Fifty travelers arrived; we have 49 hotel rooms rented.  Look up the missing guy in the airline’s database.  If he’s lost, maybe periodically scan everyone in random restaurants, see which ones can’t be linked to a valid identity.

The next solution is false identities.  To be honest, my account of Okura was based on the mechanics of visiting 20th century dictatorships.  Even today, my understanding is that it’s not easy to simply wander around China as you like.  Unless you look Chinese.  I have a Chinese-American friend who did just that; she could ignore all the restrictions on foreigners.  A totalitarian government can watch its citizens because it has key bits of leverage– they need to work and live somewhere, they have children who go to school, etc.  Watching everyone all the time is a hard problem and they take shortcuts that work for most cases; but with care, these can be avoided.

Creating false identities in an info-rich world would be possible, but tedious.  Imagine creating a fake Facebook account.  It’s highly suspicious without a bunch of commenting friends– we either have to invent them too, or co-opt real people to recognize the impostor.  The very idea of Facebook is based on shallow but wide-ranging connections… the person should have family, grade school friends, co-workers, neighbors, and all these interactions have to be plausible.

A foreigner might not be in the local Facebook, of course– but espionage frequently requires passing as a local, and again we run into the problem of missing data: the person with no online connections looks odd, and oddity invites scrutiny.

I refer to this a bit in the book– the idea is that the Incatena produces multiple, complex false identities everywhere, for Agents to step into when needed– if necessary, changing their features to match.  They’re probably mostly created by AIs, and it can be assumed that all spy agencies have been engaged for centuries in an arms race of fraudulence and counter-fraudulence.

If the systems are old enough, they might be riddled with hacks.  But I don’t buy the movie version of hacking– that any bright teenager can break into a system and make it do whatever they want.  Go and get some data, fine.  Add to a database– tricky.  Serious databases are not HTML pages you can hop in and add your anarchist message to; they’re carefully constructed to control and timestamp all access, and properly updating a web of records is actually a pretty complicated task that takes coders months, not minutes.  And adding hacked bits of code… again, a good system is housed in timestamped source control systems, and changes are looked at carefully.

(And yes, I know, systems do get hacked in grandiose ways– Stuxnet, for instance.  But Stuxnet wasn’t some kid breaking in from his mom’s basement.  I’m talking about things a single Agent can do.)

As for nanobots, I threw in a kludge– the arms race of nanobottery was statemated, and as a result no one really trusted their high-end measures and countermeasures.  In effect Okura doesn’t trust its own nanobots to stay where they’re placed; it relies on human agents to sequester travelers instead.

Another possibility is that an info-saturated society drowns in the density of the data.  It might be like the human genome: we have the data, but we don’t have the tools to understand how it all works.  It takes a long time to create new tools to dive deeper into the data, and by the time they’re written the data is denser yet, plus the databases aren’t compatible and the guy who really understood the schema quit to live offgrid in a cabin.

Anyway, APAF was set on peripheral worlds with backwards technology.  I’m tempted to set the next novel in a more central world– Earth, or Euko, or Sihor– some place that would showcase the more crazy-futuristic elements of the Incatena, and maybe a higher level of espionage.

You may remember my prediction that by 2100 corporations would be run as democracies rather than monarchies, an idea I also put into the Incatena.  This was partly based on conviction and observation, as well as the experience of a few collectively owned and/or run companies.

But there’s a new poster child for democratic governance of corporation– Valve, as explained and put in context here by its resident economist, Yanis Varoufakis.

Some companies famously allow employees to put a fraction of their time– 10% or 20%– into projects of their own choice.  At Valve, that percentage is 100%.  All employees choose which projects to work on.  And Valve is famously successful.

The immediate advantages are obvious: you’re not stuck in a job or project you hate, so motivation and retention are high. Plus dumb ideas, even if they come from the CEO, are likely to be suppressed.

Now, Valve makes creative stuff, so intuitively this model fits their business.  Still, it’s worth pointing out that most creative-stuff companies, from EA to book publishers to Hollywood, are as hierarchical as any tsardom.  If anything, creative types are more capricious and unresponsive than (say) manufacturers.  Physical things usually come with their own metrics, but who’s going to tell George Lucas that he’s doing storytelling wrong?

The obvious objection is that if your company performs a service, like banking or insurance or flying planes, there’s a lot of scutwork and it wouldn’t get done with the Valve model.  (This is my pet theory, in fact, on why Episode 3 and/or Half-Life 3 hasn’t come out.)  But this isn’t so much of a showstopper as a problem to be solved.  If it even exists: we won’t know if the model fails for banking till someone actually runs a bank this way.

As Varoufakis puts it, the genius of the market is that incentives take care of this problem society-wide.  If not enough people are making veeblefetzers, then there’s an incentive for entrepreneurs to get into that market.  In the Valve model the incentive internally is really employee interest, and fortunately people are interested in different things.  If that alone isn’t enough, there’s always more traditional incentives, like raising pay in the scutwork department.  Or maybe it turns out that you outsource the scutwork to a company that specializes in it (and which itself could be run democratically).

Why haven’t more companies tried this approach?  It can’t be because it doesn’t work or scale, because it hasn’t been tried enough for us to know that.  So I think it’s inertia.  People are just too used to the tycoon, despite a couple hundred years’ experience showing that most tycoons aren’t that smart after all.  (There were brilliant kings, too, but that doesn’t make monarchy a real success.)

Quick question about the Northern Cities Vowel Shift and future Englishes— it seems to me an awfully regional shift to be incorporated broadly into subsequent daughter languages or dialects of English. When I look back at other sound changes in English, like the Great Vowel Shift, their success seems to be predicated partly on the more regionally restricted nature of English before British colonialism really took off— not to mention the fact that nowadays English seems to have more than one prestige dialect (American English might be argued to enjoy a little more prestige that British English globally, but American English is hardly monolithic).

I could easily see in a civilization like the Incatena, where you have a new central focus of political and cultural influence being established (Mars) whose standard language is not necessarily the mother tongue of many people who speak it (cf. the number of people in India or Africa who speak English as a second language), various dialects contributing to the formation of a new standard. Obviously some will have outsize influence, based on population and cultural influence, but it doesn’t seem to follow that a regional shift like the Northern Cities Vowel Shift would come to dominate.

Just curious if you followed a different line of reasoning on this.

—James.

It’s pretty simple, really: designing a future English, rather than simply make up the first step, it makes more sense to use a sound change that’s actually in progress, though it’s currently below public perception. 

For the “regional” objection, two points.  First, what’s the alternative, given that I want to use a real sound change?  I don’t think there’s any sound change of similar phonetic importance that is more geographically widespread.

Second, if anything is going to affect near-future American English, the Northeast US is probably about the best place for it to start.  It’s already the core of General American, and other US dialects (New England, New York, the South) have historically been stigmatized and not spread outside their regions. 

A better objection might be that it’s highly urban, and doesn’t seem to have spread to the suburbs and thence to the media.  But Labov says that typically sound changes spread from the working class upward and downward, so again this is just what we’d expect for the origins of a major change.  It’s true that not all changes keep spreading; but this is sf, not physics.  I wouldn’t bet a lot on the NCVS not spreading; but I wouldn’t be surprised if it stopped, either.

Now, I said Stennud was “50C Earth English”, but this is taking a broad interstellar view.  To be more precise, I’m going to say that UK and American English will no longer be mutually intelligible, and Stennud is the descendant of the latter.  It’s likely to be more of a world standard simply because North America is a much much larger clot of English speakers.

This could change if India actually switched to English.  I don’t think this is likely, and it’s even less so in the Incatena timeline.  Currently there’s something like 86 million Indians who speak English as a first or second language; compare that to 520 million for Hindi.  In the Incatena future, Dravidian India splits off (which removes a significant barrier to the further spread of Hindi), and the Collapse reduces American influence worldwide. 

Recall that interstellar travel is expensive and time-consuming; the vast majority of people never leave their home system.  Thus planets tend to diverge linguistically; there is no “Incatenese”.  Interstellar travelers tend to use either Hanying, the major language of Mars, or Sihorian Franca.  Stennud is not that big a deal off Earth.

I’ve been reading William Labov’s book on social factors in sound change, and I’m finally getting to the good bit: where he starts explaining who are the leaders in change.

But more on that later.  For now, I realized that future Englishes should really incorporate the completed Northern Cities Sound Shift.  In particular, the “Stannud” of the Incatena should be Stennud or even Stiynnud.

Next Page »