* Antidepressants are barely better than placebo.
* Benzos are a short-term bandaid, best reserved for panic attacks.
* Antipsychotics have truly horrifying permanent side effects.
Amphetamines are a borderline miracle. Again, comparatively. Nothing else available today comes close in effectiveness at treating any psychiatric disorder. They definitely have their shortcomings, but for people with ADHD they're such a huge boost to quality of life and lead to better overall life outcomes.
There's a chance pharmacological depression treatment will catch up to where ADHD is at if/when ketamine become commonly available. Same with PTSD and MDMA. The future of anxiety treatment looks a bit more bleak... but there's experimental stuff out there with HDAC inhibitors and I think some other stuff causing fear extinction.
On average, the acceptable "separated from placebo" margin doesn't have to be that high, but the range of effect is absolutely enormous, even within close relatives.
Antidepressants that did nothing positive or negative for me or some family members were either extremely helpful or extremely harmful to other family members of mine. (Almost never vice-versa; my biochemistry appears to be quite stubborn about responding to a great many classes of treatment, well _or_ poorly, but at least one medication just made me feel completely numb while being extremely helpful to another relative.)
It can also be a weird multidimensional variable in ways many people don't expect - one medication I'm currently taking for depression is 4 pills twice a day (very short half-life), because I was gradually tapering it up. We tried switching to the same manufacturer's single pill that was 4x the dose, twice a day, and immediately I started feeling extremely nauseous whenever in a moving vehicle, which has never happened before. Switch back, problem goes away.
(I know the above sort of outcome is not exclusive to mental health medications, but in my personal experience, I've never had any other medication type vary in outcome based on whether they gave me one or multiple pills adding to the same dosage.)
As far as I know, LSD and psilocybin both have a roughly two week tolerance curve.