This is a good example of how causality is not always straightforward.

Serotonin reuptake inhibitors do work and work better the better in people with more severe depression symptoms.

However they generally can only be expected to work after around 2-4 weeks where as the serotonin increase is on day 1.

The mechanism isn't exactly "more serotonin = more happy" but that's a bit complex for most people.

I think a new working hypothesis may be that these drugs end up up regulating the synthesis of BDNF over a longer period of time, so while it does increase serotonin, that may be unrelated to why the drugs actually end up helping people in the sense it’s not the serotonin that is decreasing the depression, but something downstream

That was something that was know for decades. They don't know exactly what role serotonin plays in depression. Within the last few months they just published a paper about how they have been able to for the first time have been able to measure serotonin in the human brain. It even showed that people who were depressed showed lower levels of serotonin. This will bring new research to find exactly the role of serotonin plays in depression and hopefully develop more effective medications. That study you referred to also left out the fact there are other neurotransmitters involved including norepinephrine and dopamine. Ruoxinlin works on all 3 neurotransmitters. The only medication to do this so far has been nefazodone (weak SNDRI due to low effect on dopamine).