Wednesday, June 16, 2010

Nick Bostrom's Patch

The Phantom Blogger has notified me of a new paper by Bostrom meant to address some perceived difficulties in his argument. Interesting stuff!

In fact, very interesting, because I think Bostrom's "patch" does a great job of illustrating exactly how much stress he places on "ancestor simulations". So much so that it introduces a problem with his argument - when Bostrom says "ancestor simulations", he really means it. These aren't even general simulations of an evolved world, but simulations specifically of the ancestors of the very post-human population in question. It's hard not to get the impression that Bostrom thinks that technologically advanced civilizations may or may not decide to run ancestor simulations, but the one thing that's certain is they never decide to run any other simulations of note.

I also note that any civilization technologically advanced enough to run such simulations is apparently "post-human" by necessity. Some conversation could be had there, but really, I'm content with rolling my eyes, grinning, and saying "Transhumanists. Whatcha gonna do?"

Still, continued dwelling on the conversation leads me to ask another question I think would be awkward for Bostrom, as well as others:

Would an "ancestor simulation" be an example of Darwinian evolution?

Now, this doesn't seem to be the case. Every ancestor simulation is expressly designed and put into motion by intelligent agents. Every one of them has real teleology - extrinsic teleology, no less, the kind ID advocates in particular zero in on. Indeed, their results are all predetermined by their programmers (remember, these really are "ancestor simulations" according to Bostrom - they don't simulate AN evolution, but A SPECIFIC evolution.)

But this threatens to recast Bostrom's argument such that one of the following must be true:

(1) The human species is very likely to go extinct before reaching a posthuman stage
(2) The fraction of posthuman civilizations that are interested in running a significant number of ancestor simulations is extremely small
(3) Darwinian evolution is false for far and away the majority of conscious beings.

Of course, one could always bite the bullet and say "Well, a simulation of evolution can still rightly be called Darwinian, even if the results are predetermined and even if evolution was fully guided, complete with the outcome(s) being known." But then, there's no barrier to being both an ID proponent and accepting "Darwinian" evolution. What's more, actual "randomness", "lack of guidance", "lack of foresight", etc all become inessential to Darwinian evolution.

So, here's the million dollar question: Is Nick Bostrom a "Darwin Doubter"?


Ilíon said...

Do you not mean "descendant," rather than "ancestor?"

Ilíon said...

... it's that "post-human" business that prompts me to ask.

Crude said...

No, Bostrom is dead clear on that one: Ancestor simulators, ie, 'simulating their (the programmers') evolutionary past'.

Like I said, after thinking about it on and off for a few years, I finally realized one thing that was bugging me about the argument. That weird stress on 'ancestor simulations' in particular. And I think I figured out why he's going that specific route.

Ilíon said...

I suspected (but wasn't clear) from your wording that that is what Bostrom is calling it. But also, as I mentioned, that "post-human" business was throwing me.