In Innovation, Meta is Max

Building on my intro to innovation, which summarized previous work, let me now offer a new insight: the max net-impact innovations, by far, have been meta-innovations, i.e., innovations that changed how fast other innovations accumulated. 

Concrete studies of specific innovations in biology, tech, and business suggest that most innovation value comes from bazillions of small innovations.  Yes once in a while a large innovation appears, like computers or a-bombs, but even these require scores of supporting innovations to fulfill their potential.  And if we consider innovations that improve not just one industry or region, but the entire world overall, then it is hard to identify any innovations responsible for identifiable deviations from the steady exponential growth (plus noise) that comes from the usual bazillions of small innovations.

Yet we know of perhaps four innovations responsible for deviations that were not only identifiable, but overwhelmingly so!  As I reviewed in my singularity econ article, these are the innovations that allowed animal brains, human brains, farming, and industry.  These innovations do not seem to have directly increased the level of the relevant economy much — instead within less than a previous doubling time each innovation increased the rate of innovation accumulation by a factor of sixty or more.  (If several innovations appeared together, they apparently had a single common cause.)

The first human minds, the first farmers, and the first industrialists were actually not much better hunters, gatherers, or builders than their immediate ancestors — they were mainly just better at getting better faster. 

So if you want to worry about big disruptive future innovations, you should definitely consider meta-innovations.  Admittedly, since the above data covers only innovations that improved the world overall, it does not include predatory innovations that mainly let one group gain at the expense of other groups.  So this data allows the possibility of very large predatory innovations.  But if you are thinking about non-predatory innovations, history suggests you should be most concerned about a single new meta-innovation. 

GD Star Rating
loading...
Tagged as: ,
Trackback URL:
  • Tim Tyler

    Re: “Yet we know of perhaps four innovations responsible for deviations that were not only identifiable, but overwhelmingly so” – What? no mention of the evolution of sex – or the evolution of cooperation?

  • http://hanson.gmu.edu Robin Hanson

    Tim we just don’t much data on the immediate impact of the sex innovation, and I’m not sure what you mean by “cooperation” here.

  • http://apperceptual.wordpress.com/ Peter Turney

    In evolutionary theoy, this is called the evolution of evolvability. See:

    Dawkins, R. 1989. The evolution of evolvability. In C. G. Langton, editor, Artificial life, the proceedings of an Interdisciplinary Workshop on the Synthesis and Simulation of Living Systems. Addison-Wesley, Redwood City, CA.

    I’ve written a couple of papers on this topic. Here are some more.

  • http://apperceptual.wordpress.com/ Peter Turney

    we just don’t much data on the immediate impact of the sex innovation

    In the theory of genetic algorithms, we have Holland’s schema theorem, which is about the power of the crossover operation (i.e., sex) as a search operator. In biology, we have the Red Queen hypothesis.

    I’m not sure what you mean by “cooperation” here

    I believe Tim is alluding to Axelrod’s The Evolution of Cooperation.

  • http://hanson.gmu.edu Robin Hanson

    Peter, your computer model suggests a steady acceleration of growth, but the data suggests growth rates instead make rare large jumps. On sex you are pointing to theories, not data on the actual immediate impact.

  • http://apperceptual.wordpress.com/ Peter Turney

    we know of perhaps four innovations

    See The Major Transitions in Evolution.

  • http://apperceptual.wordpress.com/ Peter Turney

    Peter, your computer model suggests a steady acceleration of growth, but the data suggests growth rates instead make rare large jumps.

    My main point is that your post lacks any mention of the huge body of related work in evolutionary theory.

    Whether my model accounts for jumps seems to be a minor side issue. However: (1) My model is a very simple model. It’s purpose was merely to show that largest-scale trends are compatible with Darwinian evolutionary theory (contrary to common opinion among evolutionary theorists). It was not intended to account for all evolutionary phenomena. (2) The steady acceleration is probably due to the fact that the simulated environment was externally manipulated to vary at a steady rate.

  • http://transhumangoodness.blogspot.com Roko

    “history suggests you should be most concerned about a single new meta-innovation.”

    – like, for example the various forms of superintelligence that have been proposed. These include recursively self-improving AI, brain computer interfaces, or Robin’s favorite: copyable uploaded humans. It just strikes me as odd that few outside of a very small transhumanist/singularitarian community is taking this idea seriously. For example, I am looking for a PhD opportunity in artificial intelligence which might contribute to the goal of a better understanding of general artificial intelligence, and I am finding that there are almost no opportunities.

    Are we missing something? Why has the rest of the world not cottoned on?

  • josh

    does tim tyler mean sexual reproduction, because that would be a sort of meta-innovation that increased the rate of natural selection, wouldn’t it?

  • http://www.acceleratingfuture.com/tom Tom McCabe

    What would you consider a meta-innovation? Are there any modern-day examples?

  • Tim Tyler

    Re: The Evolution of Cooperation – that’s almost entirely about reciprocal altruism. I was thinking of the origin of multicellularity – which was almost entirely kin selection.

    Re: our ignorance about the effects of voluntary sexual recombination – I think we know quite a bit – e.g. from the prevalance of sexual forms among complex organisms, and the distribution of those that are asexual (out on twigs of the tree of complex life).

    I would hesitate before discounting the effect of models in this area. It is from models we know that recombination allows for the concentration of germ-line errors, and how it allows organisms to be pushed around gene land by their parasites. Observations that arise from modelling are still relevant.

  • http://www.iphonefreak.com frelkins

    @roko:

    “It just strikes me as odd that few outside of a very small transhumanist/singularitarian community is taking this idea seriously.”

    It shouldn’t. Please allow me to blunt: until the transhumanists can in general refine their self-presentation, they will be discounted.

  • Tim Tyler

    Re: evolution of evolvability:

    E.g. Dennett lists what he calls “cranes” here:

    * Eucaryotic revolution
    * Sex
    * Multicellularity
    * Language
    * Culture

    http://www.youtube.com/watch?v=11yS_w9487E

  • bambi

    Roko: It just strikes me as odd that few outside of a very small transhumanist/singularitarian community is taking this idea seriously. For example, I am looking for a PhD opportunity in artificial intelligence which might contribute to the goal of a better understanding of general artificial intelligence, and I am finding that there are almost no opportunities.

    You are running into the difference between fanciful speculation and concrete accomplishment. Opportunities abound for things that are actually real, and actually-real things are taken seriously.

    Non-brittle knowledge representation formalisms, languages supporting reflection, extension of Bayesian Networks to cover broader ranges of phenomena, etc, seem like active relevant research areas. Perhaps a discussion of feasible research topics on the “AGI” or “Singularity” email lists would be useful to you and of general interest. Good luck on your research.

  • Tim Tyler

    Nick Bostrom, on progress within science:

    “A third approach is to marry the first two and make a scientific advance that itself expedites scientific advances. The full significance of this third way is commonly overlooked.”

    http://www.nickbostrom.com/views/science.pdf

    ISTM that the idea generalises to technological progress.

  • http://noozit.com/author/tim+lundeen Tim Lundeen

    I believe you have to factor recent human evolution into this list. Raymond Crotty shows in “When Histories Collide” that genes for lactose tolerance made European settlement possible, because milk has much higher energy output per acre than cereal crops in Europe. Gregory Clark suggests in “A Farewell to Alms” that Europeans evolved to have much higher levels of trust and industry, culminating in the rapid growth we’ve seen the last 200 years. There is a very high correlation between levels of interpersonal trust and economic well-being, and between IQ and economic well-being; both of these traits have large genetic components. Finally, there is evidence that we have had high levels of human genetic selection pressure over the last 2,000 years.

  • http://profile.typekey.com/halfinney/ Hal Finney

    It’s confusing to think about meta-innovation, because even regular economic growth requires continual innovation. Exponential growth in itself is a constant battle against diminishing returns. Doubling output every n years gets harder each time you succeed. It’s really quite remarkable that even ignoring these “meta” based jumps, we have managed to maintain exponential growth for so many doublings in each regime.

    I see so many obstacles in today’s world to substantially increasing the growth rate. What would it mean to have the world economy literally double every few weeks? It is almost impossible to picture actually living in such a world. It would seem to require a fundamental change in the human condition, some almost supernatural influence that would totally change how people live and work.

    Imagine how you would live if you knew you were going to be twice as rich a month from now. And then, a month later, twice as rich as that. And then another doubling, every month, from then on. In a year you would be 1000 times richer than you are today. In fact, probably within a year or two you would have to expect another transition to an even more exalted condition.

    I always think of the old joke about scientists designing the first super-intelligent computer. It can answer any question you ask. The first thing they want answered is one of the oldest philosophical questions on record: Is there a God? The inevitable reply, of course: There is now.

    It would seem that some kind of almost God-like intervention will be necessary to achieve the kinds of changes Robin’s formulas predict. Eliezer better get to work.

  • http://yudkowsky.net/ Eliezer Yudkowsky

    That would be overkill.

  • http://zbooks.blogspot.com Zubon

    Hal understates the power of doubling: 2^12=4096. If you started with $1000 in wealth, you would end the first year with $4 million (real dollars, not the standard inflation joke of “just give everyone $1 million”). You would end the second year with $16.7 billion and the third with $68.7 trillion. That is three years from a single middle-class paycheck to the entire current global economy.

    That sounds like a singularity to me. “What would you do if you had all the money in the world?” I am curious about what would be available in the new ~$5 sextillion economy. No, I really cannot imagine that.

  • http://www.mahindrauniverse.com/mahindra_universe/innovation/ aditya

    What do you think has been the best meta innovation? I would like to get your views on innovation on our corporate blog (mahindrauniverse.com).

  • Tim Tyler

    The book presents an empirically based argument that life has developed in great evolutionary waves. These waves have not so much culminated as continued during the relatively brief, though prolific, period of human life. Just as life in general has evolved in a wave-like process of genetic transformation, human life has progressed in very long waves of technological development. The duration of these waves has tended to shorten while their amplitude has risen around an exponentially rising trend of material progress. The period since the industrial revolution has been the latest of these waves of genetic/technological evolution.

    Graeme Donald Snooks, The Dynamic Society

  • Firebrand

    What makes you think that the effect of this economic singularity on wealth will be equally distributed? What makes you think that this rapid economic growth will be the exception to the winner-take-all or winner-take-most principle that seems to operate in times of rapid change?

    I’d say you can look forward to maybe moderate increases in your effective wealth on average, with some people actually losing out, and a tiny few undergoing the sort of exponential wealth increase that you’re talking about.

    Not that this bothers me, just saying.

  • http://hanson.gmu.edu Robin Hanson

    Firebrand, see my posts from four and five days later.

  • Pingback: Overcoming Bias : Growth Mode Jump Example

  • Pingback: Overcoming Bias : Ian Morris on Foragers, Farmers, Industry, & Ems