When I see two spaces, especially in proportional text, I use it as evidence of the writer being a luddite or "past it". Only one piece of evidence, not conclusive. Probably most of its explanatory power comes from it being an indicator the user is old, educated in typing with a typewriter.
. . . or a programmer, accustomed to using monospace fonts, who has learned to touch-type and hits the space bar twice as a matter of habit.
When I see one space, I tend to suspect the writer is not a (very good, anyway) programmer, and is not particularly concerned with how text looks in reader software that uses monospace fonts -- and probably learned to type on a cellphone. I also expect to see such people spell "you" with the first two letters missing at some point.
No, I don't think a Bayesian evidence update implies confirmation bias; but your comment directly implied confirmation bias, i.e. explicitly looking for a specific piece of evidence on the basis of a different piece of evidence. That behaviour will tend to multiply the effect of any given piece of evidence, because while you're busy looking for that confirmation, you'll be less open to seeing the other side.
Ironically, the easiest way to end a sentence with a period and continue the next (after a single space) on an Android phone keyboard is to tap the space bar twice.
