Tuesday, December 09, 2008

Narrative and Simulation

I can't help but notice that, as I watch the little evolutionary simulation I just posted about, I seem to get invested in the success or failure of the little animated car the algorithm generates. I cheer for it as it makes it over the little bumps and ridges of its simulated environment, and feel disappointed when it flips over and the simulation ends in failure.

So, out of the simulation, I construct a sort of narrative. My colleague Dave suggested that this is a common feature of human interactions with simulation in a conversation we had a week or so ago. For Dave, if I understand his argument correctly, this suggests that authors who argue that the simulation has superseded the narrative as a form (N. Katherine Hayles, for example) aren't quite correct.

Here's the thing though, I may indeed build a narrative out of the interactions I observe between the simulated car and its 2d environment, but that narrative is completely without agency. The algorithm doesn't care what I think, it just evolves the best solution to the problem based on the results of its trial and error process. Granted, the simulations we see today are based on the assumptions embraced by their creators - on the implicit narratives in their heads. The designer of my example had a story of car and some terrain and a desired outcome (car travels across terrain) in mind when he/she wrote the simulation.

That's all true today, but the question is: will it be true tomorrow? Will we see simulations feeding simulations? Will the "man in the loop" (to use the term of the mid-century Strategic Air Command) be finally displaced?

I don't think it is possible to answer that question. Certainly there are some very smart people invested in "Strong AI" (there is pretty good evidence the folks at Google are heavily committed to this project) and all that entails. It is also true, however, that very smart folks have been working on this project for decades with little to show for their efforts.

It seems to me at least possible, however, that simulation may some day be firmly in command. That Tomorrow Google may count the fall of every sparrow, just as Yesterday God did. If that is the case, the very idea of human agency may prove to be a brief historical aberration, a few short centuries of what will surely seem to be madness by the beings inhabiting the later regime.

What will human beings do in such a world - if any should survive? The same as we have always done. Create narratives about the inexplicable world around us.

No comments: