That is a valid question but another thing to note is that you wouldn’t be giving this drug to the vast majority of people who do not get any symptoms from a Covid-19 infection. Nor would you be forcing people to take this drug.
Considering not everyone who gets infected gets symptoms, would it be illogical to not take the vaccine, just in case you get infected and get symptoms, but to take the drug once you do get infected and get symptoms?
To determine if this is a rational strategy or not, we'd have to get real numbers (is it really likely that 1 in 100k people die from taking the vaccine?), and compare that against the reduction in probability of dying from COVID by even combining both treatments.
And please, I just made up these numbers to answer the posed question:
> how can "I'll take experimental treatment B instead of experimental treatment A" be an argument
The numbers could lean either way and would be very sensitive to variations in the probabilities involved - I'm sure it would be very hard to reach any form of consensus on "probability of dying from taking the vaccine". It's also worth addressing wasn't even making the point of which (so-called) experimental treatment has a better likely outcome but rather addressing criticism at (so-called) experimental treatments in general.
From an "unknown risk" perspective, you'd also have to consider that COVID itself could have yet-unknown long-term risks.
That would not factor into a correct analysis: the unknown risks of covid are the same whether or not you get vaccinated (or any other treatment) because by definition the vaccine has not been shown to mitigate the unknown risks.