While it raises a few valid points, this essay is pretty bad overall. Apart from distorting the history of SSRIs (one can't meaningfully tell the story of the rise of SSRIs without talking about the prior successes and drawbacks of tricyclic antidepressants and MAO inhibitors), it pushes a false narrative characteristic of politically-motivated drug-bashers: there is a single "physical theory" of depression, centered on two or three molecules, that was advanced through commercially-motivated distortion of science and later debunked. The basic monoamine hypothesis has been known to be faulty for decades, nearly as long as monoamine reuptake inhibitors have been in widespread use; the psychiatric research community has long since been studying other possibilities.

Studies do show that, on average, antidepressants barely outperform placebo. However, this is because the effects of antidepressants are correlated to symptom severity, so averaging the entire population together shows virtually no difference. This certainly supports less widespread use of antidepressants than has been common in recent decades, but not that antidepressants are useless. It's not like the evidence generally shows that psychotherapy is much better, either. It's prescribed so much less not out of some ideological antipathy for introspection and hard work, but because it's not generally any more effective, costs far more, has much worse compliance, and is less readily available (especially in smaller cities and rural areas) than generic drugs. The dirty little secret of psychiatry isn't that SSRIs don't work; it's that everything sort of works, but any single thing doesn't work very well for most patients, and there's no validated model that predicts which treatment is likely to work for which patient.

In my opinion, the biggest problem with studying depression is that the diagnosis of depressive disorders is tremendously unreliable. Virtually all of their symptoms overlap with "sickness behavior", which is triggered by dozens (if not hundreds) of different physiological diseases. Implicit in the diagnosis of a mood disorder (actually explicit in the DSM, but people are rarely told this) is that those diseases have been ruled out as the underlying cause. In practice, this rarely happens. Doctors will do some perfunctory screening for things like hypothyroid and hypogonadal disorders, but those screening tests have very poor sensitivity and only cover a handful of common causes of depressive symptoms. If you're lucky, you might get an actual formal screening for neurological disorders. Initial screening for sleep disorders is mostly done with crappy questionnaire scales that disproportionately focus on superficial aspects of stereotypical presentation ("Do you snore?"). In a nutshell, being diagnosed with a depressive disorder has little inherent meaning beyond your GP/PCP running out of ideas or patience. In turn, any given study population of "people with depressive disorders" is unlikely to actually be homogeneous in the origins of its depressive symptoms.