[edited to clarify.  apologies for oversimplifying. ]

Right.  If your goal is to generate discussion for its own sake, it's less likely to be welcomed.  You need a reason for wanting the discussion, and that will determine how and whether to go about it here..  If your goal is to get some help in your understanding of the world and finding whether and how this idea fits into current knowledge and models, then shortform is a good start, and based on engagement (or not), you can expand to a longer post highlighting how it differs from current models and when it's helpful.

in summary: this is not a place to proselytize or promulgate ideas. It's a place to cooperatively explore what is true and how we know it.  There are LOTS of exceptions and subtlety in specific topics that are already in the Overton window around here, and I wish there were fewer, but for new/unpopular ideas, start with curiosity and learning for yourself, not with pushing or convincing others.

Edit: also, if it's unpopular/criticized due to complexity of long trains of inference, or large inferential distance from the more popular ideas/models, it's a VERY good tactic to break it down into smaller pieces, which you can discuss independently.  This is not "write more, in a series that can't be understood until complete", it's more "figure out the cruxes and individual atoms of disagreement/unpopularity, and resolve them in isolation". 

This may be valuable in less-than-adversarial complex equilibria.  Even if things aren't controlled or predicted from outside, they contain lots of forces that are pushing toward over-simple optimization (see https://www.lesswrong.com/w/moloch).  Pushing away from optimal can add slack (https://subgenius.fandom.com/wiki/Slack).  

Only skimmed, but I think you need to include COST of early action times the probability of false-alarm in the calculation.   

The high number of false positives silently trains us to wait

For me, the high number of false positives loudly and correctly trains me to wait.  Bayes for the win - every false alarm is evidence that my signal is noisy.  As a lot of economists say, "the optimal error rate is not 0".
 

How can someone inside a universe tell which type it is?

Also, a lot of thinking about paradoxes and extremely-unlikely-foretold-events misses what's likely to be MY motivation for testing/fighting/breaking the system: amusement value.  I find unlikely events to be funny, and finding more and more contortions to be adversarial about a prophesy would be great fun.

I don't know the reference myself, and I'd probably recommend against using insider shortcut phrases with people who aren't already aware.  For most people who don't already have the background knowledge to understand it from your explanation, a link isn't going to help them much.

For the kind of person who WILL benefit from a link, I'd recommend a more general one - perhaps LessWrong overall, or https://www.lesswrong.com/w/probability-theory.

A bias against drugs is very different from "drugs are always bad".  It's very reasonable to say "I'd prefer not to mess with my body via fairly blunt chemical intervention, but there are lots of exceptions for specific cases where the risk is worth it".  

Not taking drugs IS better, in the median case of a drug being offered to you.  It's just that the variance is wide - sometimes it'd be extremely bad (say, narcotics before a road-trip), sometimes it's quite good (antibiotics when you have pneumonia).  Often it's less clear, and having a default position against this kind of intervention is probably OK.

This generalizes quite a bit: Simple moral strictures should very rarely be the first or second consideration in your life-optimization decisions.  They certainly CAN be tiebreakers when it's a relatively close call.

For the specific, I don't actually know anyone who thinks all drugs are automatically bad.  I do know people who sloganize this way, but when pressed they tone it down to recreational and self-prescribed palliative drugs are bad.  Prescribed drugs (and, depending on the vigor of your counterpart in the discussion quasi-legal off-brand uses for a specific reason) are generally well accepted.  

Neither pro-drug nor anti-drug are coherent positions - there's just too much variance across drugs and patients to have any simple rule.

Separately, the topics of mental health and neurodivergence are not well-formed in our culture(s).  That deserves a discussion very distinct from drugs generally.

"supply and demand" is correct, but like the "calories in-out" theory of weight loss, is missing a lot of important causality about the supply and about the demand.  There are a LOT of textbooks that go much deeper - George Stigler and Milton Friedman are well-known authors on this topic.  

There's a wide range of behaviors and responses that are better framed as "protect yourself" than "seek justice" or "punish defectors".  I'd argue that the majority of thinking (for non-universal, non-government topics) should be framed in terms of exclusion to avoid costs, rather than punishment.  

Amazon should check if you're producing fraudulent products and ban you. This is because they're unusually skilled and experienced with this kind of thing, and have good info about it.

Amazon could be skilled at this kind of thing, but they're famously frugal and are optimizing for throughput, not for justice or even safety.   They do, in fact, ban sellers and customers who are significantly negative-value.  But their precision-recall balance is -way- different than a criminal investigation or personal decision of retribution would have.

Transit systems should ban non-payers, not to punish them, but to save the expense and hassle of trying to monitor them, and to prevent the waste of resources in having more people in the system who aren't contributing.  (IMO, first, ban anyone who reduces value by acting badly on a bus or train, even if they paid).

Likewise for infrastructure - the first goal is not justice, or even fairness.  It's protecting the infrastructure itself.  If someone is harming your mission, exclude them.   At some scale, if the infrastructure is an effective monopoly and is necessary for life, then the simpler exclusion mechanisms become infeasible, and more legible/coercive mechanisms (law enforcement) comes into play.   

This is one reason to prefer that infrastructure is distributed and no single piece is critical and irreplaceable for people who won't cooperate with the complex of expected behaviors in that community.  It makes it possible to exclude people, and they can find other places where they fit better (or if they piss off EVERYBODY, then maybe it's ok they don't get many services).