You're reading the public-facing archive of the Category Theory Zulip server.
To join the server you need an invite. Anybody can get an invite by contacting Matteo Capucci at name dot surname at gmail dot com.
For all things related to this archive refer to the same person.
Why is Rezk completion usually specified as localizing with respect to the projection from the walking isomorphism to the point (by nature an acyclic fibration) rather than one of the injections of the point into the walking isomorphism (by nature an acyclic cofibration)? It seems like they should be generating the same strongly saturated class by 2-out-of-3, but for reflections the class of morphisms that gets squashed is naturally cofibrantly generated rather than fibrantly generated so it feels like using generating acyclic cofibrations should generalize better to similar situations.
It's much easier to answer questions like this if you give specific citations to papers that use terminology in the way you're asking about rather than using vague and arguable words like "usually".
Hm, trying to come up with the references you asked for I discovered that some sources including the nLab use the projection, but Rezk's original paper https://arxiv.org/abs/math/9811037 used an injection to define the model structure for complete Segal spaces, so it's evidently doable.
That's another good reason to track down references -- sometimes you discover that what you thought was "usual" isn't at all.
One reason that sometimes people choose to localize at a map to the point is that localizations at maps to the point, i.e. "nullifications" at objects, have some better properties than general localizations. For instance, in HoTT a nullification is a modality, whereas a general localization is only a reflective subuniverse. I don't know if this is anyone's motivation in the case of Rezk completion though.