Clearly, coming up with ideas and trying them out is not something you can only do in code, but when you do code you do that all the time. Every day, every hour, maybe even several times an hour.
Ironically, a self-described non-technical "idea guy" has relatively little experience in coming up with an idea, implementing it, testing it and iterating on it.
I'm speaking in general terms of course, some "idea guys" are simply brilliant and have an amazing vision. But they're the rare exception.
Often the focus of their idea is the consequences of their new fangled web service once it already exists and have already succeeded. They tend to focus on how cool it will be when it exists, when their idea has already materialized, has made a dent in the world and has a great amount of users and a good reputation.
Rarely do they focus on the difficult part: how to get there from here. How to solve the chicken and egg problem. And because they don't focus on that journey, they tend to miss things, like they say the idea will attract people of type A, but when you ask them about a problem they introduce attribute X that clearly isn't attractive to type A, and yet they don't tweak the idea to fit the new circumstances. This is because they are bad at making mental models, unlike programmers they don't make or use mental models every hour of every day.
But it's the idea they think is great, they usually don't claim to be the next Steve Jobs.