Transcend Or Die

Follow up to Robin’s Biting Evolution Bullets:

Robin quotes Cirkovic as saying:

Biological imperatives, like the survival until the reproduction age, … will become marginal, if not entirely extinct as valid motivations for individual and group actions.  Let us, for the sake of elaborated example, consider the society of uploaded minds

I believe that this line of reasoning is deeply wrong, and for important reasons.  Robin touches on them briefly…

But there can be genes without DNA, and selection pressure without violence or great expense.  And the fact that Egan did not talk about selection effects does not even remotely suggest they are absent in the situation he describes.

…but I think that this is an important enough point that it bears repeating more loudly and directly[1].

The flaw in the view that we may become "post-biological" is that evolution works even if you don’t believe in it (to borrow Eliezer’s phrase).  Sure, some members of some cultures in some species may become post-biological.  Much like those who commit suicide, such individuals eliminate themselves from the biological rat race without having the slightest effect on the viability of the race[2] as a whole.  All they’ve done is to select against themselves.

Evolution is not a game you can win by forfeiting.  Those who do so merely cede the race to those who don’t.  If 99% of a species becomes post-biological, it will not be many years before they are the 1% minority.  That is the inevitable, bullet-biting logic of evolution.

The cliche of an old, wise culture (or species) trying to convince a young, expansionist culture (or species) to chill out is a cliche for a reason – it is an inevitable conflict.  Outside of fiction, it is also rather hard to win, because to beat an expansionist, you must either:

  • Convince the expansionists to change their ways
  • Kill them

In the former case, you’re fighting biology, which is tough but doable.  In the latter, unless you have overwhelming technological superiority, you may need to expand to get the resources for your genocide.  Which rather defeats the purpose.

Furthermore, the logic of exponential growth dictates that both of these require 100% success.  Leave a mere 1% of expansionists (people, cultures, races…) un-transcended and alive, and 6.64 doubling periods later they will be as numerous as before.

Hence the only future populated entirely by the post-biological is a ruthless version of David Brin’s Uplift universe, in which the law for younger races is "Transcend – Or Die".  And this law must be enforced swiftly and thoroughly, because the enforcers have chosen to forfeit the upper ground.  Miss one upstart race for one doubling period, and they will have twice the resources with which to defend themselves.

It’s not a pretty picture (bullet biting rarely is).  Neither is the war-torn universe of conflicting expansionists that appears to be the only alternative.  But the supposition that those who play to lose will beat those who play to win is just not plausible, heart-warming though it would be.

[1] And I didn’t see it when I skimmed the comments – apologies if someone made it already.

[2] In both meanings.

GD Star Rating
Tagged as:
Trackback URL:
  • Grant Gould

    Of course that’s assuming that the non-biologicals don’t also have means of reproducing that expand exponentially. If the uploaded intelligences have a Moore’s Law analog with any reasonable time constant then it is the biologicals who must kill of the expansionists or die.

    The question of whether a fast-reproducing or a slow-reproducing species dominates is not simply a matter of reproduction rates, or else mosquitoes and ants would have overrun us long ago. The question is how efficiently they can cultivate and use the resources that constrain their growth.

  • Julian Morrison

    You give too much credit to evolution and too little to mind. Evolution is the strictly slower and more limited process.

  • Patri Friedman

    I was using biological as a rough synonym for expansionist, as in the quote, which referred to “biological imperatives”. Silicon-based intelligence that expands exponentially is obeying the biological imperative.

    My claim is that the biological imperative to expand is one that life is very unlikely to outgrow.

  • Utilitarian

    1. The ‘exterminate or non-coercively persuade’ dichotomy is lacking. An advanced civilization with widespread presence may simply intervene to limit the development of particular capacities by junior species. In a universe where Earth is a wildlife preserve there might be superintelligent probes on Earth surrounding and observing every potentially threatening development, and able to neutralize it without significant collateral damage.
    2. If the region can sustain 40 doublings from a junior civilization’s level, the senior may simply keep a baseline of 33 doublings above that, occupying less than 1% of its maximum niche while retaining an overwhelming lead to squelch competitors and the ability to credibly threaten to saturate the environment.
    3. There may be advantages from high local density, but if probes are widely distributed throughout the reach of the senior civilization, they may observe a potential high density development throughout its entire development, producing a nearby buildup as needed.
    4. It seems that generally the most important kind of ‘expansion’ is in the volume of space reached by probes, not in the extent to which that space has been exploited for resources. An advanced civilization can control the development of competitors with minimal resource exploitation within its explored space, but competitors that arise beyond that space will be encountered in a resilient form.
    5. If the expansion of the universe results in event horizons, as the expansion of space increases the distance to resources faster than they can be approached, then the total fitness differences between civilizations that ‘burn the cosmic commons’ and those that do not may be reasonably small, or at least small enough that for utility functions that are linear with resources used commons-burning is not an efficient strategy even when external commons-burning civilizations. Imagine that propulsion technology enables small civilization seeds to expand at 99.9999999 of c, and that such seeds can convert a solar system into millions of additional seeds within a year: under these conditions the visible universe would be exhausted before commons-burning (for colonization speed) would likely result in more than a rounding error for total resources accumulated by a civilization.

  • Nick Tarleton

    Neither is the war-torn universe of conflicting expansionists that appears to be the only alternative.

    Third alternative (IMHO the most likely): we really are alone, and there won’t be other expansionists to conflict with unless we’re stupid enough to create them.

    The use of “biological” to mean “expansionist” is terribly unhelpful. The desire to reproduce would have no problem carrying over into non-organic substrate. Also, innate reproductive/expansionist motives aren’t needed. With a low enough discount rate, a society with a static population would have a motive to grab as much constantly disappearing negentropy as possible for future use (i.e. turn off the stars), and/or to protect itself from hypothetical competitors (Utilitarian’s point #3). Expansion is probably also a convergent subgoal for unFriendly superintelligences, for all of those reasons.

  • Adam Ierymenko

    I wonder if you could “transcend” evolution without actually doing so. In other words, I wonder if you could have evolution without death as we understand it.

    Consider a living system in which organisms are capable of splicing out and replacing their genetic material while they are alive. Now imagine that there exists a market in which individuals trade genetic material. The value of genetic material is somehow set in response to its fitness effects on the individuals that possess it.

    This would create selection at the gene level without requiring the deaths of gene carriers.

  • Adam Ierymenko

    “You give too much credit to evolution and too little to mind. Evolution is the strictly slower and more limited process.”

    I’m not sure if I accept that. I don’t have any proof otherwise, but I also haven’t seen anyone prove this assertion.

    What if you had a computer with the same processing power and parallel throughput as the human brain running a good state of the art genetic programming engine? Would it be as “smart” as a human programmer for instance?

    We don’t know the answer to that question since nobody has come close to quantifying such notions as intelligence abstractly and we don’t have a computer that powerful to try it with.

  • Frank Hirsch

    Nice article & discussion!

    I’d like to invoke Orgel’s Second Rule… =)

  • Ben Jones

    Evolution is not a game you can win by forfeiting.

    Why do you think post-biological implies post-evolution?

  • Dihymo

    Dialectics again. Oh boy.

    Ok look here’s how you beat the expansionists:
    Self-cloning. The only thing bad about expansion is that people die, are harmed, have their wealth stolen. Oh and the fact expansionist primitives meaning the half-attended thoughts that lead to it suck up all the time one might use to do other things. Expansionism is an addiction because it leaves no resources for any other activity.
    So you expand not by getting resources but by converting fence sitters and the neglected hopeful sane, slightly irrational, person.

    You do a complete bypass of the whole thing.

    But the conversion itself cannot be expansionist. You convert someone not by giving them a task that you want performed but by enabling them to have time left over during the harmless tasks that they wish to perform. You also need to chastise them by suggestion against expansionist desires.

    But wait it gets better.

    Why is it called expansionism when in fact it is extensionism? That is a horizontal increase meaning quantity. I would think expansion would be a vertical increase meaning quality. And I don’t mean riches. I mean time reasserting itself as a common currency.

  • Michael Anissimov

    The drive to expansion could be canceled by a singleton.

    For the reasons Bostrom articulates, a singleton seems likely.