> There's also hope that neuroscience can start to supplant at least parts of psychology in a few decades.

That definitely does happen, but so far it happens glacially, I assume because doing groundbreaking neuropsychiatric research is considerably harder and more expensive work than "evergreening" existing drugs or reading stuff like Meditations and Walden Two [1] while sipping expensive wine and then pulling an allegedly-modern psychosocial theory out of one's nether regions. My current favorite reference point is anti-NMDA-receptor encephalitis, a (class of?) autoimmune condition that was only definitively identified ca. 2007. One must assume that there are centuries of breathless accounts of "demonic possession" or "latent schizophrenia" or people who were "on drugs" (colloquially often referring to effects of heavy doses of ketamine, PCP, and other potent deliriant/dissociative agents, which tend to be NMDA antagonists) that are probably explained by this condition. And who can say what percentage of today's "schizophrenics" have arrived at a similar place along an as-yet-unidentified path?

[1] Not to say those works aren't worth reading; they're iconic for legitimate reasons. I'm just saying that context and proportionality matter in ways that are often conveniently ignored.

Indeed, we need better tools and models to explain mental phenomena. It's getting better with models like predictive coding, but it's mostly still folk psychology at this point, just splitting progressively finer hairs.