The point I was making was more that trying to proactively shape society for some goal will always miss something critical and fail in a spectacular way. Look at Communism or the neocons of Bush Jr era with the Iraq invasion. It sounds smart on paper but when you execute it then it falls apart with tremendous human cost, and the people who are doing it refuse to acknowledge it until they are physically removed from the levers of power.
Now that I think about it though, it is more to do with inflexibility of the plan, rather than having a plan itself. If you are working off of a ideological commitment, rather than setting an end goal with a fuzzy time frame and a loose path to get there, then that's when you land in trouble.
Adaptation is good but it might turn reactionary if not populist. You also want to limit experimentation to measure results. Like when you change few or many lines of code and performance changes.
Now that I think about it though, it is more to do with inflexibility of the plan, rather than having a plan itself. If you are working off of a ideological commitment, rather than setting an end goal with a fuzzy time frame and a loose path to get there, then that's when you land in trouble.