>When crossing the street, there's a tiny risk of death, but a fairly good chance of a somewhat desirable outcome (getting to the other side.) If you considered your death to have negative infinity utility, and the desirable outcome to have finite positive utility, then you would opt to not cross the street ever,

Weighing probabilities is one thing, and assumes humans are capable of rationally weighing these outcomes (which isn't clear) but even then which part of our scientific understanding of death precludes death having negative infinite utility?

Still, this doesn't disprove death as negative infinite utility, as the risk of death also exists from _not_ crossing the street; one isn't preventing the risk of death by refusing to cross the street. In fact, staying permanently in one place, never crossing the street, and doing nothing indefinitely out of a fear of death, would simply guarantee one's death.

>and the desirable outcome to have finite positive utility

Extend the timescale and the desirable outcome leads to the same outcome as the "undesirable" one.

Crossing the street ... is also a study in how rules change. A distractible American(-Brazilian) arriving in the UK will have a few heart-stopping moments when they look left, and on the verge stepping into the street, take that precautionary glance right and realize they did it wrong.

The "West" is possibly on the verge of some large rule changes.

at some point the chance of dying not crossing the street and crossing the street are equal.

I agree the characterization isn't very useful, as you point out - but that's the tricky thing with infinit(ies), isn't it? When discussing human behavior and introducing an infinity concept, the final result is the same for all people in all cases.