Discussion about this post

User's avatar
Beancounter4beantraders's avatar

Speaking of AI. Ran the stay article through Claude and this is what I got:

Claude responded: The author's bottom-line prediction is: NDASA likely gets standing, and if it does, a stay is likely to be granted (and ultimately the rule reversed).

The author's bottom-line prediction is: NDASA likely gets standing, and if it does, a stay is likely to be granted (and ultimately the rule reversed). Let me weigh how solid that chain actually is, since each link carries different risk.

Standing: this is the load-bearing assumption, and it's shakier than the author treats it. The author concedes NDASA's argument that the rescheduling order will cost its members money is, by his own description, "ridiculous" once you do the math (about $1,000 per employer), and the medical review officer argument is similarly attenuated (an indirect, market-mediated economic harm flowing from employers changing testing behavior, not direct regulatory injury). The author still concludes this shows NDASA's members would suffer a specific, concrete injury, but standing doctrine generally requires the injury be not just concrete but also fairly traceable to the challenged action and not too speculative or attenuated. An indirect economic ripple effect through third-party employer decisions is exactly the kind of theory courts often reject under modern standing doctrine (think of cases like Clapper v. Amnesty International or the general skepticism toward "competitor standing" theories that aren't squarely on point here). The author's own framing — calling the argument "ridiculous" while still endorsing it — should make you discount his confidence, not just note it as color commentary.

The merits analysis cuts against his own prediction. This is the more important tension. The author spends a long stretch arguing the D.C. Circuit's reasoning in the NORML case was simply wrong, that the statute's plain meaning assigns rescheduling authority to the AG alone, and that the court "legislated from the bench". He also notes significant intervening factual developments since 1977 — the 2020 international rescheduling of cannabis, the Epidiolex scheduling history, and the 2023 HHS recommendation — that undermine the factual premises the 1977 court relied on. If he's right that the precedent is both wrongly reasoned and factually outdated, that's an argument the government has a real shot at distinguishing or limiting NORML on appeal, not just losing if the case reaches the merits. Yet his conclusion leans on the idea that merely forcing the case to be heard on the merits is "more likely than not" fatal to the rule. Those two threads don't fully cohere: if the precedent is as weak as he argues, "reaching the merits" doesn't automatically mean the government loses.

The stay-specific reasoning is the weakest link. His stay analysis barely engages prongs 3 and 4 of the four-factor test (harm to other parties, public interest) beyond a rhetorical aside ("what's a few more months?"). That's doing a lot of unacknowledged work — patients and operators who already incurred 280E relief and reliance costs, plus the public interest in regulatory stability, are real countervailing considerations the author doesn't seriously weigh before concluding a stay is likely. He also leans heavily on a declaration from Dr. Madras likely being persuasive to a court that has historically sided against agency overreach, despite himself describing her evidence as cherry-picked and exaggerated — again, citing weak evidence as likely to be persuasive isn't the same as it actually compelling a stay on the merits.

Net assessment. I'd treat "more likely than not" as too confident a framing for what's presented. The standing argument he calls weak is doing critical work in his own chain of reasoning, and his merits discussion arguably undercuts (rather than supports) the idea that simply reaching the merits dooms the rule. This reads less like a dispassionate probability estimate and more like an advocacy-adjacent piece where the author wants the rule to fall (he says as much through the entire "what DEA should have done differently" section) and is reasoning toward that outcome rather than from neutral priors. That doesn't mean he's wrong — D.C. Circuit panels can be unpredictable, and the precedent he criticizes is real and binding unless overturned — but the specific "more likely than not" claim isn't well-supported by the reasoning actually laid out in the piece itself.

Beancounter4beantraders's avatar

The problems with this is 1) adoption, 2) distribution, and 3) consistent availability (go to brand or off brand) over time which is impossible in this industry.

4 more comments...

No posts

Ready for more?