You say "less risky" but does either method produce a probability with its forecast? Or does probability remain a secondary, fuzzy, human, meta-judgement?

That's a tricky question. Based on recent production history, a 2% growth rate is very likely. However, an exponential growth cannot go forever in a physical world so the exponential model is very unlikely on a large time scale.

I agree. In fact, sustained exponential growth is not sustainable in most systems on even a medium or short time scale. Overshoot and collapse is an all too common response in everything from prices to populations.

I worried that someone would take the idea that 2% can't continue "forever" and flip it to a psuedo-proof of collapse.

The fact is, it isn't a proof of that or anything else. There might be plateaus, there might be soft landings, there might be a "long tail" ... who knows?