It's pretty difficult to generate a coherent vision by committee - you usually end up with something that nobody finds compelling, not even the people who collaborated on it.

And of course the image of an all-knowing visionary CEO is a manufactured image! It's not like people can actually predict the future. Rather, it's selection bias at work. Those CEOs whose vision doesn't resonate end up getting replaced, or their company goes bankrupt, and then they get written out of the history books. The thing is, you need to have a vision to get people excited and actually rowing in the same direction (nevermind the right direction), so without that you're dead in the water anyway.

Sure visions aren't usually created by committee. But aren't visions usually starting very raw, and not really that valuable or unique("let's build a computer for every pocket") - and to get to a coherent and unique vision, you need some research and exploration, which is usually a team effort ?

No argument there (and I think that the success of Musk's companies is largely because of the faceless engineers he hires that toil away to make it a reality).

My point is that left to their own devices, most of those engineers would rather toil away on their own startups or hobby projects, and that convincing them to all work on a tiny piece of a big project requires someone who's both charismatic and has a finger on the pulse of what problems can excite people. Musk has been uniquely successful in aligning people to work on big, hairy, audacious goals.