A strong enough signal will show in a sample of one.

With a correct model, cause and effect can be predicted without any samples (scientific progress often occurs this way).

And some people’s intuitions are remarkably astute. Albeit there are a lot of people who have incorrect intuitions and don’t know it.

A cannabis example would be strong paranoia after one sesh. More ingrained paranoia (qanon?) that I have seen linked to regular heavy usage doesn’t feel to me to be a stretch of the imagination.