Maybe one way to solve this problem is to define several milestones, and a percentage of the funds get released once milestones are reached. At each milestone, each funder can decide whether or not the person has achieved their stated goal for the milestone.
If the person who puts money into the project gets cold feet, they can back out at any time, but any money they have committed up to the current milestone goes to the founders and the rest gets returned.
Problem is that oftentimes most of the money is needed before a first prototype. And many concerns wont emerge until the prototype comes out. So effectively what will happen is that the milestone system will be gamed so that most of them are "paper" or "cad" which dont take much effort.
well, with these projects that are vastly overfunded, like the case example, it seems like most of the money could have been safely put in some sort of escrow so that the founders wouldn't get lazy.
And if there's a proper description of the milestones, it seems like gaming might not be so easy.
For example, if there were a "test the metal case for signal loss" milestone, and if the founders only gave a half-hearted attempt, then a bunch of people could have had the option to pull out at that point.
Basically the founders need to create a proper plan, with reasonable milestones, and then the funders might be more willing to give more money. I think after a while, people would be able to flag a gamed milestone.
You can't necessarily fault the founders for not thinking through all of the potential pitfalls. A proper plan will not cover all of these details. Take the apple death grip issue -- they had a plan, it was produced, and the issue was really only emphasized after it was produced and sold to customers.
If the person who puts money into the project gets cold feet, they can back out at any time, but any money they have committed up to the current milestone goes to the founders and the rest gets returned.