For example, running very large trials in a short time is very high effort.

a) Risk vs. reward -- not that I'm saying we're making the right balance decisions, but the correct speed depends on that balance in general.

b) The current administration is extremely unlikely to make vaccine approvals faster. If anything their instincts are the other way around.

Because it's inordinately more expensive.

We're computer people, so we have a good analogy here; the COVID vaccine did speculative branch prediction. They basically operated _as if_ they would get approval at all stages where they could, parallelizing much more of the process at the cost of a _very_ expensive branch fail if something went wrong.