Is there some greater context to this question? how often does someone promise to extend the life of someone else outside of familial relationships?
"It's Better To Save Infinite Shrimp From Torture Than To Save One Person"
Infinite is doing a lot of heavy lifting in this argument, I feel. You conclude your argument with
"I don’t know exactly how many shrimp being tortured is as bad as one human dying, but it’s probably less than 75 million!" I believe this is the key point around this charity, about which people argue. What is the shrimp-to-human ratio that they are willing to trade off? This is only compounded by our inability to properly conceptualize the difference in magnitude between large numbers, making this an extremely challenging way to convince others of your position.
Personally I find the arguments using finite numbers more convincing. i.e. number of shrimp killed * probability of shrimp feeling pain > 1 human suffering. As well as arguments arguing that shrimp welfare is more effective and neglected than animal welfare.
"From my perspective, the victory condition was that the kid can tread water, jump in, and climb out."
I would consider adding an additional condition of being able to swim 30-50 meters without aids. It addresses the potential risk of falling into the water in a location where it's challenging to get out.
Don’t think this is a terrible idea but worry that the limitations of implementation make it unworkable.
For instance, physical dead men switches would only work in the case where the models are stored/run on devices with the switches implemented. This means that models cannot be open, and that they cannot be shared to non-Deadman switch infrastructure, which in the case of a hostile AGI, means not allowing the model to copy itself offsite, and preventing it convincing humans to allow it to be copied off site.
If we can guarantee these facts we have probably solved alignment and no longer need the Deadman switches.