You're absolutely right that there's a lot more that goes into a clinical diagnosis of a psychological issue. It's important in that setting because it informs the likely treatment and it's very important to use all the information at your disposal to arrive at the most complete diagnosis.
But that doesn't mean that it's not valuable to use incomplete information to speculate on a diagnosis in other settings. Historians do this when they can use psychological theory to explain the actions of the people they're writing about. The example that stands out in my memory was Alice Miller's exploration of Hitler's possible psychological profile in "For Your Own Good." Just because she wasn't working from a complete medical profile doesn't devalue that exploration, it just means that the conclusions reached should be interpreted differently. Her diagnosis wasn't aimed at crafting a course of treatment for the individual since he was long dead. Instead, it was meant as a way to understand how a set of normal and understandable circumstances can yield an individual so far from normal that he could otherwise defy understanding.
It seems quite natural that Steve Jobs would be psychoanalyzed in much the same fashion. He was someone who was highly successful and impactful, so to the extent that his psychological profile can be mined for traits and practices that led to that success, it should. And he was highly polarizing, so we can also find areas of his profile that current and future leaders should be aware being possibilities in themselves as well. But since Jobs is dead, the need for a clinical accuracy in the diagnosis is gone. A more probabilistic diagnosis that looks at what was likely going on rather than concluding what is going on is all we can do.
Trump, while still alive, is also a candidate for this type of psychoanalysis. Like Jobs, no part of the diagnosis will lead to any course of treatment. In fact the most common diagnosis people seem to claim (narcissistic personality disorder) would mean that Trump would be very likely ignore any of that speculation, no matter how accurate it is. But where this type of analysis can be useful is in guiding the rest of us in how we respond to Trump's actions. Again, it's not important for us to be perfectly correct, it's only important to inform our actions based on the most likely diagnosis. Because the consequences of a "we can't know, so we shouldn't guess" are just too high. This is a man that can start wars, launch nuclear missiles or exercise any of the myriad powers granted to the US President. Even a limited understanding of the thought process of the man wielding those powers can be vital to opponents and allies in the US and other world leaders abroad.