In fairness, psychiatry is totally different today than in 1973. The obvious change is that a huge reduction in the number of inpatient beds, combined with increasing demand, have created huge pressures to admit only the most obviously unwell patients and discharge them as quickly as possible. Most psychiatric inpatient stays are just a few days - just enough to get a patient through a crisis, revise their medication and (hopefully, but not always) arrange for appropriate outpatient care and support. The downtown of most US cities is a testament to the fact that, in 2023, under-treatment of severe mental illness is a far greater concern than over-treatment.
On an ontological level, psychiatry made a huge leap forward in 1980 with the publication of the DSM-III. One of the core goals of the DSM-III was to address the concerns raised in the Rosenhan experiment, making diagnostic criteria more robust and reliable. While there are still many controversies and shortcomings - most prominently regarding the over-diagnosis of less severe conditions - we now have a suite of reliable, validated diagnostic instruments for most serious conditions. For the most part, we aren't diagnosing or treating patients based on the gut instinct of an individual practitioner; we're using objective criteria with proven inter-relater reliability, guided by the over-arching principle that, regardless of symptomatology, no-one is mentally ill unless a) they're experiencing distress and/or b) they're causing significant harm to others. There are many shortcomings in how psychiatric medicine is practised today, but the era of locking people up just because they behave strangely is definitively over.