[GeorgeKangas]

I can see two ways around the low probability problem, for RNA abiogenesis:

1) The crystal gene hypothesis of A. G. Cairns-Smith. As a clay crystal grows and splits, the info in the crystal's defect structure is replicating with impressive fidelity, and those defects also interact with the surroundings. So you get the Darwinian game bootstrapped pretty much for free. Later on, the crystals start using organic polymers; later still, the polymer technology is developed well enough to take over from the clay. So this might make abiogenesis reasonably probable on one planet.

2) An observable universe is just any epsilon size patch, on an inflationary universe. The space-time curvature of our whole observable universe is too small to measure, hence the radius of our inflationary universe is a large multiple of the 13 G-lightyear radius we can observe. So abiogenesis could be highly improbable in any observable universe, answering Fermi's paradox, yet be probable within the much greater volume of an inflationary universe (maybe this is what TFA said? TLDR [Edit: yeah, it says that right in the abstract]). And there could even be a large number of inflationary universes, for all we know.

[scarmig]

One nice thing about 1) is that it is more falsifiable than 2). 2) basically gives us the ability to explain any almost arbitrarily unlikely event (at least unless we figured out a lower bound on the space time curvature of the observable universe). You might as well say that life is vanishingly unlikely but a series of vanishingly unlikely many-worlds quantum coincidences happened to result in the particular universe we see.

[sandgiant]

There is a sort of philosophical upper bound on the total size of the Universe in an eternal-inflation scenario. This is the Boltzmann brain paradox. The argument goes that if there are infinitely many universes it is infinitely more likely for you to be a brain randomly fluctuating into existence in the vacuum, rather than being an actual human being evolved through evolution.

This conclusion is obviously absurd, so the argument goes that there can't be infinitely many places in which random things can happen in the Universe. Note that the size of the Universe required to produce a Boltzman brain is much larger than that discussed in the article. It's still nice to think that there might be an upper bound for us to avoid some of the more daunting implications of the anthropic principle.

[benlivengood]

Your second point is probably the most critical for this article; all we can conclude (eventually) is an upper bound on the probability of abiogenesis for a certain volume, total mass, age, or number of stars. There is literally no lower bound; there may be uncountably many other inflationary universes without life.

[Enginerrrd]

Yeah I think the idea that the first self-replication was RNA-based to be really really dubious.