Which is like a plaintext equivalent, so if the clip is easier to guess spammers will just do that instead.
(I've seen this explained based on accounts of a public library built in the US which had stairs, but no elevators, for a specific section - this was still ADA compliant as they had staff on hand to serve readers in that specific section)
So a 30s delay or whatever it takes to limit spammers is probably not entirely unthinkable.
Remember that the whole trick of a spammer is that you aren’t able to see that his many actions are coming from one source. Otherwise it would be trivial to block. So there’s not a lot stopping a spammer from running many 30 second delays in parallel.