In fact, you are awfully close to many worlds. The only difference between your view and "many worlds" is that certain universes have to disappear whenever there are conscient observers in other terms of the wavefunction, and that looks to me like a very strangely implementable thing: how can the arrangement of certain atoms in a certain term of the wavefunction, giving rise to a concient being (be it a microbe), suddenly make disappear other terms that describe entire universes billions of lightyears big ?
Isn't it more reasonable that these other terms just continue their happy existence ? No observable fact of their existence can in any case influence the conscient configuration, so why do they have to disappear ?
I think that from the moment you can accept "collapse" (and hence will run in any case into difficulties with relativity), then the most obvious candidate of a physical process that leads to collapse is not "conscient being" or "microbe", but rather gravity. It's actually Penrose's idea: from the moment that we have a superposition of states that have a significantly different gravitational interaction, the superposition fails and collapse must occur.
Again, from the moment that a collapse happens, you have a problem with relativity, because a collapse has to happen on a specific spacelike slice, and can't hence be a lorentz-invariant process - that's why it was easy to find a counter example with the two microbes, each one being "earlier" than the other one. So IF you can accept collapse, and hence IF you can accept a serious problem with the principles of relativity, then I think that Penrose's idea is very interesting. I'll tell you why: it avoids a fundamental difficulty with general relativity.
If you have a quantum-mechanical superposition of two gravitationally different systems, then those two states have two different spacetimes. But if you have two different spacetimes, then it is not clear at all how you are going to define a unitary time evolution operator over it! The times themselves become part of the superposition!
This is not a problem with special relativity: the spacetime in special relativity is the same for all terms, it is flat Minkowski space. So it is possible to define a unitary time evolution operator over this shared spacetime on superposed states. But that doesn't work anymore with a wavefunction in which each term has a different spacetime associated to it (because each term describes a different mass configuration). It could even be that each term by itself evolves unitarily, but with its own unitary operator, different from term to term. THAT has then the possibility to give rise to a non-unitary overall evolution, and maybe an objective collapse. That's more or less Penrose's idea.
But again, it will run into conflict with relativity, in one way or another, as ANY physical collapse scenario will. But if ever collapse is physical, I'd seriously bet on Penrose's idea.
I find it a way more attractive approach than any "living observer collapses the universe" approach.