While that was an early hypothesis, I'm not sure that has really been proven. A lot of the early data wasn't really "initial" viral load, but viral load when they presented in the hospital. That's not necessary an indication of initial viral load.
So, initial viral load is not COVID specific. Also viral load at first detection is a different yet important criteria for EBOLA, COVID, and other diseases. https://www.jci.org/articles/view/83162
Your paper is about viral load at diagnosis--not infection. Diagnosis happens after your immune system has been fighting a while.
The 1 vs 50 difference is known to be based in part on the innate immune system not just the adaptive immune system. Most of the time you don’t need to create specific antibodies etc. If only a single virus enters the body it’s likely for you to fight off the infection without noticing though with the downside of failing to develop long term immunity.
https://en.m.wikipedia.org/wiki/Innate_immune_system
The 50 vs 50,000 is a related though less proven idea. Extreme exposure can simply overwhelm the immune system and extreme cases directly result in septic shock, but the middle ground is harder to test.
Just try it for yourself, put a N95 on you face and don't even press the metal nose thing, and then see if you can fit a surgical mask better than this unfit N95.
This is another meme which needs to die.
I have, with a proper fit tester. If you use a half mask respirator with cartridges, sure, the rubber seal is better by default. However, most N95 masks I've seen in public are are the cheap rigid masks leaving large gaps all around the circumference, whereas the surgical masks/bandana cling to the skin better.
Also, it’s important to realize that initial viral load makes a significant difference in how sick people get. Simply reducing exposure can therefore reduce how sick someone becomes.