I don't think absolute infinite choice is necessary for a sandbox. Everyone knows there is limited time and resources. The discussion about Chult should happen in session 0 or even earlier -- the "what game are we going to play" discussion. Not that it isn't possible for a campaign to take a hard left turn, but if it does everyone needs to be okay with the GM taking a few weeks off to prep/research/brainstorm the new situation.
I'm running two "sandbox" games right now. And they're a mix of what is being talked about here. In the first, knowing these players, they need that first "adventure" to get them moving. Once they're moving, I can sprinkle hooks and rumors here and there, and then they are free to pursue whatever they want.
The second group is doing a hexcrawl/sandbox. We used Beyond the Wall for character creation, but OSE for classes, etc. The BTW backgrounds gave each character ideas for things they want to do in the world, regardless of hooks. One player wants to be a merchant. So he spends all his time talking to NPC traders about routes, goods, practices. Another is seeking to reclaim his family's lands lost to invasion. Their starting village has several locations nearby they could visit, and they have a map of the environs and what they would likely know. If they want to wander off, they can, to wherever.
To
@Reynard s comment, something like Chult might not even be something the party may have heard of. Depends where they are in the realms (unless they hear about it at a port or some such, at least thats how it would run in my game).
I run in Greyhawk, which has the benefit of most of my players knowing nothing about it, the places, the countries, etc. Sure, they could look it up, but they don't. But the FR they know all about and have a general sense of who and what is where. That makes it a little harder.
I'm also using this to try to be a little more parochial. No world ending events, no doomsday clocks, just exploration and the players doing what they think their characters would want to do.