The original herd immunity "estimates" of 60% were 1-1/Ro. This never considered heterogeneous spread and it assumed everyone was equally susceptible.
It's hard to estimate prevalance because PCR testing has a large false negative rate. It depends highly on which day of the infection you're tested and even then the peak rate is still quite miserable for trying to find it.
Serorevalence is also tricky as we trade off false positive for false negative and the antibodies decline after a few months. That's not to say that memory B cells can't spin up production on demand though. Then there's the whole t-cell response which may play a bigger role in asymptomatic cases. All in all, its understood that antibody testing is underestimating infection rates.
Regardless of what values we think herd immunity should have, it's pretty clear that places like Sweden have seen the death peak and then tail off without any major change in distancing policies. If anything, people are relaxing more with time.
edit: I should mention that since march, we do have better treatment options and we have stepped up our game protecting care homes all over the world including Sweden which saw a majority of deaths from those facilities. This is why it's important to look at trends in other countries as well.