Shiny things and organizational volunteer work seem to become incresingly important as companies grow in size, the volume of internal projects increases, and thus it becomes a lot harder to gain consistent internal visibility. So working on a "shiny" project becomes your best (if not only) shot at getting noticed. The truly Machiavellian will recognize this environment for what it is, and just pitch shiny new projects if none are immediately available -- the necessity of those projects usually being a secondary consideration.

The problem often grows worse, not better, when you introduce concepts like top-down goal alignments and stack ranking. These can easily backfire by forcing savvy employees to scramble for maximum-impact projects and deprioritize all others. You end up with a handful of hero projects and a whole bunch of misfit toys that nobody wants to touch.

I'm not sure any company has ever truly solved this problem at scale. Obviously some company cultures are better at it than others. (I've never worked at Amazon, so I can't speak honestly or credibly to its culture.)