I agree that average over the kilowatt-hours nuclear might be safer than coal. But the problem is that the risk is HIGHLY concentrated.
I can self-insure my computers because if one ever fails I have the cash to handle it. It's not a financial crisis. I don't self insure my car (for collision/comprehensive) because it's too big of an expense for me to absorb at any one time without severe hardship.
Nuclear reactors (as they stand today) are more like cars than computers. They tend to rob a few people of their whole lives and really eff up the ones that don't die. Coal plants have the "advantage" of robbing many people of a portion of their lives in a more subtle way.
LFTR reactors have physics on their side in that if containment is lost, so is criticality. Boiling water reactors, pressurized boiling water reactors, etc in my mind can't be made safe because they require power to stay safe, even if they're not generating power.
You've suggested that so long as the generators and whatnot arrive eventually, things will be fine. The problem is that you've made the assumption that somehow, some way the generators will arrive and the reactor can be made safe again over the course of some weeks or months. Nuclear reactors are one of the few pieces of infrastructure that can't simply be abandoned if something bad happens, or else the problem gets worse and worse and worse. This is a positive feedback cycle rather than a negative one. Positive feedback often "blows up" (a term we use in electrical engineering) or grows without bound. Nearly everything else in modern society functions with negative feedback and that kind of intrinsic safety is what most people want.
That's not to say that ALL nuclear reactors are by definition unsafe; just all the ones that are currently operating commercially because it's such a nightmare to permit new designs and the existing manufacturers have no interest in inviting competitors into their market.