Actually, I think you a) misunderstood the claim in point 2, and b) take issue with the assumptions - not the math/statistics - in point 3.
your math in point 2 assumes you will certainly find yourself in a life threatening situation while carrying the device.
In point 2 the likelihood of a life threatening situation was not assumed or calculated (that's point 3). The conclusion, restated: If there's a 1/3500 chance you'll be able to effectively use the device to save your life, you're breaking even while mitigating risk.
And your point 3 probably vastly overestimates the fraction of life-or-death calls.
You've stated that I've made assumptions and you're right - which is why they're labeled as such. Feel free to suggest yours.