Thank you for this interesting post, even though I don’t agree with your conclusions.
I believe one key difference between killing someone and letting someone die is its effect on one’s conscience.
If I kill someone, I violate their rights. Even if no one would directly know what I did with the invisible button, I’d know what I did, and that would eat at my conscience, and affect how I’d interact with everyone after that. Suddenly, I’d have less trust in myself to do the right thing (to not do what my conscience strongly tells me not to do), and the world would seem like a less safe place because I’d suspect that others would’ve made the same decision I did, and now might be effectively willing to kill me for a mere $6,000 if they could get away with it.
If I let someone die, I don’t violate their rights, and, especially if I don’t directly experience them dying, there’s just less of a pull on my conscience.
One could argue that our consciences don’t make sense and they should be more inline with classic utilitarianism, but I’d argue that we should be extremely careful about making big changes to human consciences in general without thoroughly thinking through and understanding the full range of the effects of these.
Also, I don’t think use of the term “moral obligation” is optimal, since to me it implies a form of emotional bullying/blackmail: you’re not a good person unless you satisfy your moral obligations. Instead, I’d focus on people being true to their own consciences. In my mind, it’s a question of trying to use someone’s self-hate to “beat goodness into them” versus trying to inspire their inner goodness to guide them because that’s what’s ultimately best for them.
By “self-hate,” I mean hate of the parts of ourselves that we think are “bad person” parts, but are really just “human nature” parts that we can accept about ourselves without that meaning we have to indulge them.
Do you know about The Dignity Index? Might be interesting to team up with them/get their input.