From the File of Bad Ideas: Banning unknown drugs with "Machine Learning" | Hacker News Reader

From the File of Bad Ideas: Banning unknown drugs with "Machine Learning"

undark.org

12 pointsworik11 comments

Certainly the cannabis analogues which are not similar to natural cannabinoids are big trouble.

That is, if marijuana harms you the harm is pretty subtle and takes a long time to unfold. Maybe it makes you silly, maybe it stunts your psycho-emotional development, from first principles I'm inclined to think it might cause diabesity but you can't blame cannabis for the diabesity epidemic because most of the diabese aren't users.

Cannabis analogues like Spice and K2 really can kill you right away.

After the CBD bubble popped people realized they could convert CBD to valuable cannabinoids such as Delta-8 and Delta-10 THC. I'm no advocate of those substances but they don't seem to be deadly like the structurally unrelated analogues.

You'd think we would have learned from alcohol prohibition that banning a drug a large number of people want popularizes more dangerous alternatives.

Designer drugs/“Bath Salts” are a problem including in many places which have legalized or decriminalized marijuana.

One of the biggest obstacles to preventing them from hitting the streets is the fact that you can legally import them and quite often sell them as they are not restricted substances.

By the time authorities identify the substance and put it on a restricted list a new formulation appears.

Having the ability to predict formulations of these drugs might be quite a useful tool.

Sounds like a very expensive and multilayered game of whack-a-mole.

> from first principles I'm inclined to think it might cause diabesity

Actually cannabis can reduces diabetes symptoms and stoners are thinner on average, I don't see how you can think that

> cannabis can reduces diabetes symptoms and stoners are thinner on average

I am interested in sources for both those.

And please do not use the term "stoner". It many places it is a pejorative term, in my world it is used like racial a epithet, "some one whose welfare we can ignore".

The sources are, um, not high quality.

- https://pubmed.ncbi.nlm.nih.gov/30288813/ - https://www.verywellhealth.com/marijuana-and-diabetes-510517...

Admitting that no set of rules, nor method of generating rules, is of much use in the face of enforcement that is mostly interested in political posturing. (Or in racial targeting. Or in being the top dogs. Or is just plain incompetent.)

...But why, besides "everything is cooler with AI!" stupidity, why are they even attempting this approach? Seems like it's just begging for drug labs to crank out an endless series of drugs du jour. Vs. just saying that anything which is produced with psychoactive intent, or is marketed as psychoactive, or actually is psychoactive is up-front illegal, unless it falls under {Phrase Here describing a reasonable regulatory regime for such stuff}.

If this is a way to get the government to fund machine learning for designer chemical compounds (with possible medical applications as a side effect), I might be in favor of it, at least temporarily... and repeal when the technology is good enough for commercial use without government funding.

Can we not editorialize titles?

Yea. Na!

Reply on news.ycombinator.com