It's unlikely that earthquake prediction will soon advance to the point where it can limit our earthquake exposure risk. So it is best to invest in adaptability, flexibility, resilience, etc. Matt Khan makes this point with respect to climate change.
Nassim Taleb takes this thinking a step further by suggesting that we seek "antifragility". It is great to bend with the wind, but even better to be strengthened by the stress. He summarizes this argument in "Learning to Love Volatility: In a world that is constantly throwing big, unprecedented events our way, we must learn to benefit from disorder," in the WSJ. He alludes to his new book which I have not yet read, so I can only respond to the essay.
Taleb suggests "five policy rules that can help us establish antifragility as a principle of our socioeconomice life." They include "Rule 1: Think of the economy as being more like a cat than a washing machine;" "Rule 2: Favor businesses that benefit from their own mistakes ...;" "Rule 3: Small is beautiful, but it is also efficient;" "Rule 4: Trial and error beats academic knowledge;" "Rule 5: Decision makers must have skin in the game." These rules are all good conversation starters and prompt the question of how they relate to each other.
For example, take #4. Real "academic knowledge" is by definition unsettled and open to trial-and-error learning. If not, what is it? Blind faith? Take #3. The author attacks blind faith in economies of scale. Again, blind faith, where it exists is always a problem. But I think that most people understand that there are also diseconomies of scale -- and finding a sweet spot where each are contained is a great idea.
We can always expect problems where error-correction is not possible. Where we engage in one-off projects, we take the biggest risks. Get the Chunnel wrong and what do you do? Manhattan Projects and NASA projects do not lend themselves to cost-benefit analysis. No competition means no error-correction possibilities. One way to heed all five of Taleb's admonitions is to minimize the number of mega-projects in our lives. Start by forgetting about high-speed rail.
Cost-benefit analysis is not even applied in the simple cases. The Nov 20 WSJ includes "California Spurs Electric Cars ... Chrysler Group LLC next week plans to unveil an unusual automobile -- an electric car that doesn't stand a chance soon of turning a profit and is unlikely to draw many customers. So why is the company doing it? It has to. California requires it."