Advantages do accrue to early adopters of new growth modes, but these gains seem to have gotten smaller with each new [transition]. … 1. The number of generations per growth doubling time has decreased. … 2. … As we get better at sharing info in other ways, the first insight-holders displace others less. 3. Independent competitors can more easily displace each another than interdependent ones.

Earlier today I wrote about the inequality at each point in time, in the eras between transitions:

The number of species per genera and individuals per families has long declined with size as a tail power of two. After the farming revolution, cities and nations could have correlated internal successes and larger feasible sizes, giving a thicker tail of big items. In the industry era, firms could also get very large. Today, nations, cities, and firms are all distributed with a tail power of one, above threshold scales of (three) million, thousand, and one, thresholds that have been rising with time.

So, the unequal success that comes from some moving sooner in a big transition between growth eras has declined in more recent transitions. Yet the within-era inequality at a moment in time between groups like nations, cities, and firms has increased over time. As larger groups have become feasible, with more internal correlation in their success, the high tails of very large groups has gotten thicker, until they are now Zipf distributed evenly across many size scales. And in such Zipf distributions, typical group size increases with the both minimum efficient scale and total population, both of which have been increasing.

“But that is not all, no that is not all!” (Said the Cat in the Hat.) In the future, the feasibility of much longer lifespans will allow more lifespan variation, which I expect to induce a Zipf power-one tail of individual wealth, up from the current tail power of ~1.4. So individual wealth inequality should increase with this change, and also as population increases.

While a Zipf tail power of one produces more inequality than a tail power of 1.4 or two, yet more inequality comes from a tail power of less than one. And in fact, that sort of inequality is common in our world, among the firms within an industry. For example, this paper estimated power laws for:

The market shares for 506 brands in 48 product categories of foods … [in] a large urban market in the Southwestern US, … [and] for 665 brands in 43 product categories of sporting goods … for the entire United States. … The data are restricted to brands with market shares no smaller than 1%.

Out of the 48 categories with a 5%-significant power estimate, only 17% had powers over one — the other 83% had powers less than one. That is, for most product categories, most market share is held by a few big firms. This isn’t a total winner take all situation, however, as the maximum market share for any firm in this dataset of 1171 brands was 75%. But it is a heavy skewing toward the largest feasible shares.

So if firm size within each product category is distributed with a power of one half, how can the distribution of all firms have a power near one? It must be that the size of product categories is distributed with a power of near one, and firms suffer when they try to handle too many categories at once.

The example of firms within a product category shows that is possible for environments to favor a few large groups, near the maximum feasible size. If some new management trick let firms better manage diverse products, perhaps firms worldwide would merge to a few big firms. And if travel congestion costs that now limit city sizes becomes less relevant, perhaps most everyone would live in a few enormous cities. And that would force everyone to live in a few vast nations, unless a single city could be divided up among several nations.

Even families might become more unequal in the future. Today people with the same surname break up into many small units that mostly succeed or fail on their own. But if those who shared a surname became organized enough to promote a common style and reputation for their members, they might tend more to succeed or fail as a unit.

The future whole brain emulation (em) scenario I’ve been exploring plausibly includes most of these sources of increasing inequality. Em cities should have larger minimum efficient scales, and greatly reduced congestion limits to achieving the many gains of close interaction. And the “clan” of copies of the same original human would make sense as a unit of reputation and governance. For example, I’d expect each occupation to be dominated by a few clans in the same way each product category is now dominated by a few firms.

There might also be substantial advantages to being one of the few hundred “types” that most folks know well. We might return to a forager situation where we prefer to deal with the types we have come to know well over a lifetime of interactions, relative to less known “strangers.”

In sum, I envision an em future with Zipf distributed individual wealth and firm size, where most of trillions or far more of ems come from a few hundred well-known clans, and most live in a handful of cities in a handful of nations. Even so, probably no one individual, clan, firm, city, or nation would hold more than a third of their respective total. In a vast world, even great concentration is consistent with robust competition.

Now imagine that you work in a virtual office. That is, while you are actually lying at home in your VR pod (or being an em brain in a data center), you experience yourself as sharing a virtual office complex with your work colleagues. Sitting at your desk working at your computer, talking in a meeting, chatting with a neighbor in his doorway, or perhaps walking the cubicles to feel the buzz.

OK, now ask yourself: how could we design more effective virtual offices, for the purpose of making an efficient workplace not needlessly taxing its workers? For example, what features of office spaces today would we jettison if we could, since they mainly deal with physical constraints that need not apply in virtual reality?

Maybe each person would feel the temperature and humidity they like best. Maybe walls would glow, instead of all light coming from glaring overhead lights. Maybe you’d always feel like you were walking barefoot on soft grass. Maybe all surfaces could be of the most luxurious textures and styles. Your computer “screen” might fill up a wall, or be 3D in a vast warehouse-sized space. But what else?

People might just appear in each other’s offices, instead of having to walk there, but that might feel disruptive. Perhaps hallways could be lots shorter, with each person having a huge personal corner office looking out on a spectacular view. But would it be ok if the shapes and views of offices and halls made no sense relative to each other?

In meetings it might be possible to let each person see and hear others in great clear detail, even adding biometrics on if they felt scared, tired, etc. You might even be able hear their thoughts if you wished. Or at the other extreme, each person might instead be able to project a pleasant attentive appearance no matter how they actually felt. You might even appear to be in several meetings at once. Where along this spectrum would typically make for the most productive meetings?

If each person could make the walls etc. look however they want to, then how will other people know what they are seeing in order to interact smoothly with them? Would you like the ability to look out at any time and see dozens of people as they work, if the cost were that dozens of people could you look at you at any time?

I’ve read a lot about speculation about virtual reality over the years, but I’ve not seen much that took these sort of questions seriously.

I see these as Watkins main mistakes:

1. Overestimating coordination capacities. Watkins said we’d cut underused letters like C,X,Q from our alphabet, eliminate mosquitoes and house-flies by ending their breeding grounds, put all city traffic below or above ground, and accept many American republics into the USA union. All of these require far more coordination than we seem capable of.
2. Underestimating wealth indulgence and signaling. Watkins said we’d adopt an engineer’s efficiency attitude toward food preparation and personal fitness. People unable to walk ten miles at a stretch would be weaklings, and we’d use central cooking instead of personal kitchens. But rich folks don’t want to work that hard, and humans have long asserted wealth and autonomy via personalized vs. communal dining. Institutional communal food, such as in dorms, ships, military bases, boarding-house, etc., has long been avoided a sign of low status.

Added 10a: The institutional food that is cheapest, and lowest in status, makes you eat where they say, when they say, and what they say. Yes of course a restaurant is “institutional” in some ways, but it costs more because it offers customers more flexibility in time, location, and food.

A majority of citizens today perceive cryonics enthusiasts as kooky. … I share some of this skepticism. … Wouldn’t any reasonable person — one worthy of revival — dedicate a lifetime’s accumulated resources to helping their children and posterity, instead of splurging it all on a chancy, self-important gamble for personal immortality?

Consider:

“Median total [US Medicare] expenditures in the last 6 months of life [in '00 to '06] were $22,407.” (More)
“Out-of-pocket medical expenditures … for the years 1998-2006 … in the last year of life is estimated to be $11,618 on average.” (more)

Since US medical spending has more than doubled since then, we must now spend over $50K per person on the last six months of life. And this spending seems to, if anything, reduce lifespan. In contrast, a ~$40K (30 + 10) cryonics procedure gives a chance of a whole new life, and increases the chance of others gaining the same benefit at a lower cost. So why don’t Cowen or Brin first complain about selfish end-of-life care?

Brin continues:

Some people who sign up for storage believe their bank accounts alone — set up to earn dividends until some future era — will suffice to make them worthy of being thawed, repaired, and given full corporeal citizenship in a coming age of wonders. Somehow, I wouldn’t give that bet anything like sure odds, no matter how many technological barriers future people overcome.

Let me get this straight. People who suffer ridicule and fierce conformity pressures to pay to take a chance to avoid death and help others avoid death, who actually end up being right, and who in addition save money that gets invested in the world economy to help it to grow faster and larger, in order to generously pay future folks to revive them, do not deserve to be revived?! Even if they are quite willing to work to pay their way upon revival? Future folk should instead steal their money and refuse to revive them?! Why doesn’t Brin suggest that we today kill old folks a few weeks early to save thousands in medical costs? How exactly are they deserving yet cryonics patients not?

Btw, a second person has finally taken their cryonics hour. Any more takers?

This distribution of firms and cities seems to result from their being tolerably effective across a wide range of sizes, having long unequal lifetimes, having little net local growth, and holding a roughly fixed total number of people. In contrast, individuals have more equal lifespans, are psychologically inclined to spend more as they get richer, and have spending habits that correlate only weakly across generations. (“Rags to rags in three generations.”)

How might these change in the future? In the em era, I expect firm distributions to stay similar, but expect city and individual wealth distributions to change. I’ve talked before about how I suspect strong gains to em concentration, as they suffer less from travel congestion, leading perhaps to most being in a few dense cities. In this post, let me talk about em wealth.

Since em lifespans should be limited mainly by em wealth, em lifetimes can vary a lot more than human lifetimes, and ems can have more long-term spending consistency. While some ems will spend their wealth on more copies, others will hoard their wealth. Some may even manage to consistently reinvest most of their wealth via something like a Kelly criteria. This seems likely to make future em wealth evolution more akin to today’s firm and city evolution. I thus expect a near Zipf distribution for the high tail of em wealth.

This change in tail power should make em wealth distributions more unequal. Under a tail power of ~1.4, today’s richest person has about $75B, which is about 0.04% of the world’s $200T wealth. Under a power of ~1, the richest person might be about a hundred times richer, holding ~4% of the world’s wealth, or $7.5T.

Since a Zipf distribution has an unbounded expected value, its inequality also depends on the total population size (which follows it). The following table shows this dependence:

The “% of Richest” column says what fraction of the total wealth is held by the one richest person. The “MidW %” column shows the (smallest) fraction of the population that holds half of the total wealth. And the “MidW/ave” column shows how much richer is the mid-wealth person (for whom half of all wealth is held by richer folks) than the average person.

For a Zipf wealth distribution, as the population gets larger wealth gets more concentrated. Even so, the very richest person holds a smaller fraction of the total wealth. The same should apply to firms and cities if they retain a Zipf distribution — the firm and cities that hold most people will get larger, even though the largest firm or city would be a smaller fraction of the total.

In sum, as the population gets larger, I expect firms and cities to get larger.  And for “immortal” ems, I also expect a more unequal distribution of wealth. Even so, as population increases the very largest firms, cities, and rich folks should hold smaller fractions of their respective totals.

Added 11p 14Jan: This post has now been up for a whole day, with zero comments and one vote. Which has to be some sort of record for reader disinterest. This is especially noteworthy, given that I’m especially proud of this post, culminating several days work trying to understand something important about the future. Alas that I  sometimes bore readers, but I’m writing this blog mainly for me, so I’ll continue to write about what most interests me, even if past responses suggest readers won’t be as interested.

Is there a better option? Imagine a copy of my current mental state is saved, and then revived for brief periods on special occasions, like major ceremonies, consultations, and votes. These revivals might decline in frequency with time, but spread over hundreds or even billions of years. When the accumulated effects of these revivals threatened to cause too much divergence from the original me-now, that original could be revived instead, to start another cycle.

That seems to about as much life as is feasible for me-now to have. And this sort of me-now immortality seems cheaper that the usual sort. That is, the (likely future) cost to give this sort of immortality to a me-now seems substantially less that the cost to ensure that a mind continues to evolve at something like its current rate and capacity for trillions of years. Of course this cost is still high, too high to offer to all me-nows. But neither is it ridiculous.

Yes, there is an ambiguity in how big a mental difference would count to create a different me-now. But the ordinary concept of immortality is also ambiguous when minds can be copied and run in parallel — how many of parallel copies need to last forever (or a very long time) for “me” to be “immortal”?

I expect a few future ems to be “immortal” in the sense of a single copy that continues on for a very long time. But I expect far more me-now-immortality, archived minds brought back with declining frequencies for rare ceremonies and consultations. This approach is cheaper, better serves the needs of others, and may even offer more of a reward to ems who identify more with me-now than their distant changed descendants.

How far have we come, and how far might we go? Imagine this progress as climbing a ladder on the side of a tall building, where at each new floor we get ten times more capable. If floors are ten feet apart, and ladder steps are one foot apart, then each ladder step represents a decibel of growth, i.e., a factor of 1.26.

From a population of perhaps a thousand folks two million years ago, humanity has so far climbed about nine floors, or 87 steps, up the ladder. We climbed 36 steps as foragers, taking fifty millennia per step, 23 steps as farmers, taking three centuries per step, and so far 28 steps as industrialists, taking five years per step. (All estimates rough.) How much further might we climb?

An em civilization might climb even faster, at least for a while, taking a step per day! A few months ago I estimated population sizes for very advanced em civilizations:

[Using] android ems … a two kilometer deep Earth city could hold a … thousand billion billion ems. And a solar system civilization might fit a billion billion billion ems.

Reaching this last level would require us to climb about fifteen more stories, or 150 steps. And since the observable universe contains a thousand billion billion stars, to reach a universe-spanning civilization with this density we’d need to climb an additional twenty-one stories, or 210 steps. So, we’ve climbed nine floors as humans, but have perhaps another thirty six floors to go, to reach the “top” forty-fifth floor. That’s four times as far to go as we have come.

Yes, we might find ways to grow more after that, and our pre-human ancestors must have also done a lot of growing. But my main point is that we hope to climb much further than we have climbed so far.

What more can we say about this future climb? Our biggest fear should be falling off the ladder to the ground below, never to rise again. To avoid such a fall we must find the right tradeoff between climbing slower to carefully inspect each new step, and climbing faster to avoid the hailstones, wind gusts, etc. that we expect from time to time, and which we expect are less problematic further up.

Of course we should also wonder what we may become as we rise. We are no longer the foragers who began this climb, nor the farmers who climbed just a few floors below, and those ancestors would probably not be pleased with everything we have become. We’ll probably also have misgivings about what our descendants become. But hopefully we will on net be proud of them, just as our ancestors would probably be proud of us.

Attempts to see more clearly up the ladder tend to fall to two extremes. Many folks study trends over the last few steps, and tentatively conclude that such trends may continue for another few steps. And many such folks also declare it to be impossible, as a matter of principle, to see further. They often respond to concrete arguments about more distant forecasts with “that must be wrong, since we can’t see that far,” as if that were some sort of law of physics.

At the other extreme, some folks claim to see all the way to the ladder’s top. We’ve been getting less fertile and more green, rich, peaceful lately, so very advanced civilizations must be idle rich zero-population-growth leave-nature-be peaceniks. Or we’ve been getting more into movies and video games lately, and like fast net connections, so very advanced civilizations must be virtual reality navel-gazers crushed into tiny balls. Or because an AI might improve itself, one AI will soon take over the world and then live on forever as indisputable ruler of the universe.

Now I do suspect that a few of the basic physics limits we now see apply to our universe’s true physics, of which the physics we see is only an approximation. For example, if light speed and entropy limits persist, then growth and innovation must eventually slow way way down. And evolutionary theory seems robust enough to embolden me to estimate that in the absence of strong central coordination, competition must force our descendants to be well adapted to their environments, in an evolutionary sense.

But I don’t yet see how to say much more about the very long run. That isn’t impossible, but it does seem very hard. So I prefer to focus on trying to see into the next era, beyond our industry era. While that does take more than a simple projection of current industry-era trends, it takes far less than is needed to see life in the very distant future, way up on floor forty and beyond. To see into the next era one needs only good general theories about how societies work, and a good guess regarding the key changes by which the next era will differ from ours.

We can imagine progress as climbing a ladder on the side of a tall building, where at each new floor we get ten times more capable, and each one-foot-apart ladder step represents a decibel of growth. We’ve climbed through three or so floors of foraging, two and a bit of farming, and may climb through three to six floors of industry, before coming to a new era, which may last through two to seven more floors. It seems to me that at least some of us should try to use our best social science and guesses on the biggest upcoming changes to try to understand this next post-industry era, even if we don’t know how to envision the five to ten or more eras which may follow after.

I see seven relevant considerations:

1. Some think subsistence-wage ems an abomination, and so prefer to delay or prevent them. Conversely, others think that vast em numbers times lives worth living makes the em world a good well worth hurrying.
2. Some want to delay the em transition, to give more time for its serious consideration. Others want visible em efforts to start sooner, fearing that serious consideration won’t start before then, and expect an earlier start to give a better total discussion. Still others think that, as with nanotech, early public anticipation of such events tends to make them go worse.
3. The richer and more capable our civilization gets, the lower seem its chance of being extinguished by most disasters. Ems would make us richer faster, and ems survive biological disaster especially well.
4. During the em transition our civilization is especially vulnerable to collapse, or to a central power grab. This transition is less disruptive when the last tech to mature is computing power, and most disruptive when that last tech is cell-modeling. This argues for hurrying scan and cell-model tech, relative to computing tech.
5. Many fear that a single self-improving AI will suddenly grow vastly in power and take over the world. Some want to delay this event until they see how to pre-provably control such an AI. So such folks want to delay most other AI tech advances, including ems.
6. Assuming pre-provable control is infeasible, on-the-fly control seems better when the people controlling are many and fast relative to the controlled AI. Since ems can be much faster and numerous than humans, this argues for hurrying ems.
7. Great filter and anthropic selection considerations greatly raise our estimates of existential risks that could leave the universe empty. These do not much raise AI risk estimates, however.

On #1, I confidently estimate em lives to be numerous and worth living. On #2, I weakly estimate little benefit from delay or early publicity. Points #3,4 are the strongest I think, especially #4, and both argue for speedup. Since I think a single machine suddenly taking over the world is pretty unlikely, I give #5,6 less weight, especially when taking #7 into account. So on net I favor hurrying em cell-modeling tech most, em scan tech less, and weakly favor delaying em computing tech.

Added 11a: More considerations from the comments:

8. Future people may evolve to differ from us via competition and changed circumstances. Some hope Earth will soon collectively organize to regulate to prevent such change, and so want to minimize change and competition before then. Since ems give more faster change, they prefer to delay ems.
9. It seems humans can live on as ems, and non-poor ems need never die. Not dying is good, suggesting we hurry ems. Conversely, if uploading really kills humans, perhaps we should delay ems.
   ]]> http://www.overcomingbias.com/2011/12/hurry-or-delay-ems.html/feed 57 http://www.overcomingbias.com/2011/11/science-fiction-is-far.html http://www.overcomingbias.com/2011/11/science-fiction-is-far.html#comments Wed, 16 Nov 2011 13:30:49 +0000 Robin Hanson http://www.overcomingbias.com/?p=28287 Continue reading →]]> SF author Greg Benford posts a ’97 Peter Nicholls talk:

   I decided that I would write ALIEN ARTEFACTS but call it BIG DUMB OBJECTS. … But the joke was on me, because as I came to write the entry, I realized that the subject– which was vast alien enigmatic artefacts–was at the heart of what attracted people to science fiction. And even stranger, I realized that no matter what literary shortcomings you found in Big Dumb Object sf – and believe me, there are plenty – that Big Dumb Object stories were often successful, that even if badly written they were usually good to read. Why? …

   There is in science fiction, even or especially (as I will argue later) in so-called Hard science fiction, something which in other context we tend to think of as unscientific, be it called sense of wonder, or the sublime, or the transcendent as the Panshins have it, or the romantic. And one rather mechanical way of creating this effect is for the storyteller to imagine something very very big and mysterious, like the spaceship Rama, or like Larry Niven’s Ringworld. That is, the mysterious something in science fiction often has its locus classicus in the Big Dumb Object. …

   Of the BDO novels I’ve cited, [big] voyages in space become [big] voyages in time in the majority of them: … It is, as the celebrated cliché has it, the last frontier, and this ties in with what one does in frontiers of all kinds, one meets the “other”. I think the meeting of humanity with the other is now generally accepted as one of the great themes of science fiction. … The sublime … is dehumanising. It makes us feel small and unimportant and indeed hardly there at all. I think this feeling of our vulnerability and littleness in the context of cosmic vastness and indifference, is one of the root feelings of space fiction. … Sf writers capable of perfectly good straightforward, journeyman prose, tend to fall into florid poetics of the most excruciatingly embarrassing kind when trying to imagine what transcendence might feel like. …

   BDO fiction … is about being dwarfed by space and hugeness, about attempting to maintain our own humanity, warts and all, in the light of this vastness, while at the same time yearning to be better or other than what we are. And this is not a theme that is intrinsically scientific at all, which makes it all the odder that it is in the hardest and most scientific sf that we tend to find the purest examples. …

   I began by saying that I had recently re-read a dozen or so classics of hard science fiction, and I listed them. What I didn’t say then is that it was a rather disappointing experience. … The main problem is the sense of wonder, that feeling you get when confronted by the truly awe-inspiring in sf. It doesn’t tend to occur so poignantly the second time round. (more)

   Immersing ourselves in images of things large in space, time, and social distance puts us into a “transcendant” far mode where positive feelings are strong, our basic ideals are more visible than practical constraints, and where analysis takes a back seat to metaphor. Many “hard science” folks who won’t allow themselves ordinary religious feelings do allow themselves these transcendant feelings. Which seems ok, but for the risk that it might overly infect their practical beliefs.

   ]]> http://www.overcomingbias.com/2011/11/science-fiction-is-far.html/feed 11