1. Ideas are a dime a dozen. Your idea is most likely not nearly as unique or revolutionary as you think. Most people you tell it to will probably be neutral to negative, and the chance that you will tell it to someone that both thinks it is an excellent idea and have the resources in time and money to follow through on it is exceedingly slim.
2. You need the invaluable input from other people. Maybe your idea is crap, maybe it has already been tried, maybe no one is prepared to pay to solve that particular problem, maybe someone else is already doing whatever you want to do successfully and are years ahead of you. You better get to know these things before you spend 18 months coding in your basement.
And both those arguments apply to people that don't know how to do due diligence or lack significant experience in the industry that they are addressing. Somebody could be working on your idea in total stealth and unveil it tomorrow. It's impossible to have perfect due diligence in a non-transparent world. Yes, if you lack experience, you should definitely talk to a lot of people, especially prospective users.
What we are working on currently is a spin off of an existing company that we already run and we've spoken to numerous other users about it. It's a solution to a common problem in a particular industry that nobody has appropriately addressed. Until we launch, I do not see the wisdom of telling strangers what we are working on. There are existing very large companies that have certain pieces in place and if they had the insight that we do, could devote resources and beat us to the punch. Sometimes there is value in blindsiding your competition.
I simply disagree with complete openness as a universal rule. Maybe because of egos it's better to be open more often than not but it should be a carefully considered modus operandi.
We will be launching within 6 months and already are alpha-ing with users. You can take a measured approach to openness. It's not just complete openness with everyone you meet or complete cryptic, insular stealth mode.
2. You need the invaluable input from other people. Maybe your idea is crap, maybe it has already been tried, maybe no one is prepared to pay to solve that particular problem, maybe someone else is already doing whatever you want to do successfully and are years ahead of you. You better get to know these things before you spend 18 months coding in your basement.