But where do you draw the line at "overthinking"? I agree that "don't help the homeless guy down the street in favor of funding AI alignment research" is a bit unintuitive, but keep in mind that "don't help homeless guy in favor of helping 100 random guys in africa" is also unintuitive (at least to the extent that we needed a whole movement to popularize it). I'm not saying that AI alignment research is actually the most worthwhile cause to fund, but "convincing people to donate to unintuitive, but theoretically greater utility projects" is basically the reason why effective altruism even exists. If people already naturally donated to the highest impact charities rather than donating to their alumni or the local opera house, we wouldn't need EA because that would be the default.