computer models
can we trust computer models to predict the future? heh. yes and no. relatively recent research strongly suggests that efforts to predict the stock market based on past performance is doomed to fail. does that mean that *all* computer models suck? no. engineering models of bridges and airplanes work incredibly well. no one builds real aircraft to see if they'll fly. they build a computer model. if the computer says it flies, then they'll build scale models. then they'll build the real thing. building the real thing first would be stupid. so *all* computer models are perfect? heh. no. see above. in 1972 mit published some predictions on the limits to growth based on computer models. how'd they do?
see for yourself. on one hand, it's pretty good. on the other hand, we haven't gotten to the interesting bits yet. though looking at which direction the miss is trending, i'd say the "inevitable" economic (and population) collapse is further out than was predicted 40 years ago. and mostly because technology has extended our reach into the ground to pull out more oil profitably.