www.technotheoria.org/p/nootropics-a-fraught-but-promising

This claim is contradicted by the existence of standard blinded clinical drug trials.

For example, the FDA/NIH DailyMed label for WAKIX (pitolisant) states that its efficacy in narcolepsy was evaluated in multiple “randomized, double-blind, placebo-controlled studies.” Pitolisant is a real prescription drug, so these studies are direct counterexamples to the statement that blinded studies with real drugs are "not possible."

Whether blinding is perfect for every drug is a separate methodological question. Some drugs produce noticeable side effects that can partially unblind participants or investigators. But that is very different from saying blinded studies with real drugs are impossible. They are routinely conducted and are a standard part of drug evaluation.

The claim says arsenic is considered carcinogenic in humans but not animals. That is not what major carcinogen assessments say.

The International Agency for Research on Cancer (IARC) states: “In view of the overall findings in animals, there is sufficient evidence in experimental animals for the carcinogenicity of inorganic arsenic compounds.” The same evaluation also states that arsenic and inorganic arsenic compounds are carcinogenic to humans.

So the article's contrast — carcinogenic in humans, not animals — is incorrect for inorganic arsenic compounds. Historically, some older literature emphasized how hard arsenic was to model in animals, but the current IARC evaluation explicitly says the animal evidence is sufficient.