Euh. Okay, less conjecture, slightly more experience here, I hope... I have a clean set of findings here (I'm just sifting through threads looking for stuff I can contribute, so mind my intrusion here!). I haven't set-out to prove or disprove any theories or concepts in this thread, but just sharing my experience on the subject so far - maybe to provide new ideas to try or chew on.
The car only ever "sleeps" on the 12v battery. Nothing ever "runs from" it. When asleep, only the bare essential functions are awake. This draw is very, very minimal - and the car wakes so easily and stays awake for so long, it's not likely draining the 12v battery more than a few percent.
There is a current sensor on the 12v battery. One of the car's controllers tells the PCS (power conversion system, or DC-DC converter) to adjust its output voltage - which actually ends up changing the entire 12v bus voltage - to maintain a 10-amp charging rate. Yeah, this is totally practical and it actually works, even if a little rough, because the 12v battery only consumes as much current as its voltage (state-of-charge) dictates. So, simply modulating the PCS voltage is all it takes to limit the charging rate to 10 amps.
Once it reaches 14.4v and the amperage is around zero for a while, it'll linger there for a bit (presumably as an equalizing effort) and then back off to around 13.5v where it's considered "done" and will allow the car to sleep. If the battery hasn't gone through this complete charge cycle, it won't sleep 'til it's done.
During the charge cycle, it seems to also characterize (in new firmware around 2020.28.x) the 12v battery by considering its voltage vs. charge current compared to what a "good battery" should look like, and it'll give that "12v battery needs to be replaced" warning if it doesn't like what it sees. This is why even a brand-new lithium 12v replacement battery can set-off the 12v warning in new firmware.
It's also been observed that the car will "play yo-yo" with the 12v battery a bit, lowering the PCS voltage below the 12v battery's resting voltage and letting the car be powered by the 12v for a while, back and forth, discharging and recharging it, maybe to test and characterize the battery, maybe to exercise it.
My experience with lead 12v batteries has been that "if it's just sitting there doing nothing, it'll last forever". I've got a big blue Optima spiral AGM battery that was made in 9/2010, and it's been sitting on the ground most of its 10-year life, picked up and started using recently, and it behaves 100% like brand new. It was resting at about 11.6 volts when I saw it - almost completely discharged. However, I've also seen UPS batteries that were never used - and were kept fully charged their whole life - and they were basically inert bricks because the UPS was constantly maintaining a charging voltage on them. It seems like it's possible to abuse them in both ways - to leave them fully charged is just as bad as with a lithium battery, but to constantly cycle it is also bad as it'll wear them out. I think Tesla is walking a fine line here and doing their best to manage the 12v battery as well as they can.
FWIW, my mid-2018 Model 3's 12v battery performs today just as well as I'd expect it to when it was new. Quite impressed.