For me, the weird thing about the double-slit experiment for matter is that the excitations of the detector are discrete, i.e. we see individual particles arriving at the detector, as expected under the classical model of matter behaving like particles.

It’s only when we send many particles over time, that we can tell that we’ve been sampling from a distribution given by the interfering wave-functions.

This suggests that the particle is interfering with itself, thus behaving like a wave, but since we only detect each particle in a single place, it’s still behaving like a particle in some sense.

This is very different from the classical wave phenomena seen when we do the double-slit experiment in a water tank, where there is constant interference across the whole axis the whole time (there is no particle-like behaviour at all).