My Bayesian prior has been sufficiently updated to effectively know: I don't know a single mother with a restraining order for a son who doesn't smoke too much weed. I have met a bunch of mothers and sons in my life. In a Popperian sense we can all only assume so stop splitting hairs.

Clearly, smoking a shit ton is bad for people with marginal mental health or marginal emotional health.

Couching it in pseudo-scientific babble doesn't change the fact that it's an isolated data point, and you're vying for the olympic goal for jumping to conclusions.

Anecdata.

According to Scott Alexander, I only hate the grey tribe outgroup ergo with enough e-risk paperclip minimization, we can drop p(doom) with lowered replacement fertility levels via lethargy from cannabis consumption, meaning marijuana is a hedon machine and positively utilitarian. Legalizing marijuana is effectively altruistic.

Clearly, weed is good

So you agree as Bayesian prior’s are literally assumptions.