Dixie’s Science Roundup: Artificial Life, Artificial Societies, and Poverty Traps

       P   O   S   T   E   D       B   Y       D   I   X   I   E

The Associated Press reported recently that we’re three to ten years from creating artificial life.  Science writer Seth Borenstein cites three major hurdles that we must first overcome:

A container, or membrane, for the cell to keep bad molecules out, allow good ones, and the ability to multiply.

A genetic system that controls the functions of the cell, enabling it to reproduce and mutate in response to environmental changes.

A metabolism that extracts raw materials from the environment as food and then changes it into energy.

If I may be permitted to paraphrase, what stands in the way of creating an artificial cell that can transport materials in and out of its cell wall, reproduce, and metabolize is the creation of a cell that can transport materials in and out of its cell wall, reproduce, and metabolize.

Not to worry Seth: we’ve all had to write under deadline.  At least your article has the ring of scientific truth.

Seth didn’t tell us that scientists plan to introduce these artificial cells into a synthetic Earth currently under development outside Los Alamos.  By adding special growth factors, they hope to coax these cells to quickly evolve into higher life forms that will understand the “impracticality” of banning Islam.

While biologists are pondering Seth’s tautologies, Joshua Epstein is busy modeling societies in silico.  Epstein has shown that by using colored dots on a grid to represent various human agents, and programming these agents with a few very simple rules, he can model complicated social phenomena like rebellion against a central authority and ethnic cleansing.

Epstein is currently modeling a community of nonprofit consultants who turn on one another during an especially severe shortage of flipchart paper.

Finally, in a recent Scientific American opinion piece, Jeffrey Sachs, director of the Earth Institute at Columbia University, asks what we can learn from the geographic distribution of poverty.

Poverty is most strongly correlated with areas of the globe that struggle with low food production because of irregular rainfall, degraded land, and other factors.  Some parts of the world, like the tropics, are home to diseases more easily controlled in temperate areas.  A region’s physical isolation can also abet poverty by limiting access to sea-based trade.

According to Sachs, these geographic deficits produce “poverty traps,” areas in which humans struggle to subsist.  While these poverty traps tax human endurance, Sachs is convinced targeted investments can overcome them, and at “surprisingly low cost.”

Clearly Dr. Sachs is convinced that wealth is better than poverty, if only for financial reasons.*

_____

* Thanks to Woody Allen for this conceit.

Comments

[erasmus]

The Epstein article is way cool. Where do you dig up this stuff? Was he one of your profs?

[Albert Ruesga]

I saw a review of Epstein's book in a recent Scientific American article (a gift of my in-laws, God bless them). I don't yet know what to make of his argument. The outputs are amazing and beautiful, I think, and clearly worth studying further. Having worked with software models before, I also know that a researcher might spend a lot of time tweaking the parameters until an interesting or recognizable pattern emerges on the screen. There's that effect to consider.

Epstein's research reminds me in a big way of Conway's Game of Life.

[Pam Ashlund]

I now officially worship at your feet Dix.
I seem to have lost by copy of Steven Levy's "Artificial Life", but if memory serves it emerged (wink) in the early '90's.

If you live long enough (and I'm working on it) these predictions can get pretty funny. You don't hear very much about Artificial Intelligence anymore and that was around the corner, any day now, etc. in the 1960's.

==Pam

"I've been rich, and I've been poor and rich is better"

[Albert Ruesga]

I'm no expert, but even when I was in school back in the late 80s to early 90s studying this stuff, AI was getting snagged on basic human knowledge ("When I put my left arm behind my back, my left wrist is located there too") and natural language comprehension. By contrast, so-called "expert systems" (for disease diagnosis, for example), were fairly easy to program. Have you kept up? I'm not at all current.