I disagree. Say there's a 1 in 100 chance someone acquires COVID because of a delayed vaccine (maybe conservative in the US where spread is still quite high) and a 1 in 150 chance they die from it. That's a 1 in 15,000 chance vs. 1 blood clot death in 7,000,000 doses so far. https://twitter.com/kakape/status/1381962983929049093

Now of course there's a lot of stuff we haven't considered. But it cuts in both directions. The incidence of blood clot deaths may be higher that indicated so far given we're not systematically monitoring for them. OTOH, we also have to consider the base rates in the population.

We also have to consider that the person who gets COVID because of a delayed vaccine will pass her case along to an average of 1 other person given where Rt is in the US right now. So the death rate is actually twice as high as I indicated above.

So I don't think it's a particularly close decision. Even if blood clot deaths were 10-fold higher than observed so far, which is certainly possible, it wouldn't be a close decision. And that's before considering the knock-off effects on contributing to vaccine hesitancy.