The Grand Theory of the Unified Executive

Somewhere out there, in some splendid ivory tower, a Professor of Theoretical Law is working on a new Grand Theory of the Unified Executive, which he hopes will overturn the increasingly obsolete Standard Model of three independent sources of Law, Justice and Executive Power (who ordered those anyway?).
This will simplify Law and Order tremendously, and make actual enacting and implementation of law much simpler, showing finally the underlying Unity of All Law.

To be this man is what all pre-law and law students aspire to. It is the career track that essentially all law schools aim their students at (well, all the first rate ones anyway), it is what inspires all law students to struggle through the large lecture classes on Common Law and Legal History 101; and, while the disappointingly large fraction of law students who fail to get on the theoretical law academic track may find sub-fields where they make use of their legal training, maybe even make more money than Professors of Law, they will always feel like their dreams were shattered and their time in law school a waste.

Of course I am being unfair to lawyers - while there really are some Professors of Theoretical Law, most law students do not aspire to this, and aim to and are content to become applied lawyers.
You can to a large extent divide them by sub-field; there's your contract lawyer, tax lawyer, your real-estate lawyer, the criminal defence attorney, the prosecutor, ambulance chaser etc. Some aspire to be judges, or lawmakers, or professors of law, but of the half a million or so lawyers in the US, only a small minority does so.
Most are content to to work in rote jobs earning $80-120k per year, doing essentially the same work again and again - you check deeds, fill in escrow account forms, file motions, check on subclauses of corporate law and verify contracts for loopholes and inconsistencies. Nice work if you can get it.

Now, in some ways lawyers have it easier than scientists: the laws are known, mostly. You can go look them up and they're all there.
On the other hand, the laws of law do not obey the Equivalence Principle - not only do they change with time (take that searchers for variation in the fine structure constant, hah!), the laws can change with position (if you live in Europe, you should not rely on "Law and Order" for your understanding of the local law!). No conservation of energy for lawyers. Worse than that, the change can be rapid, you can't even make adiabatic approximations. Worse still, they're discontinuous - C¹ approximation fails. Aargh.
Lawyers can even create their own novel legal concepts de novo, through sheer force of reason, personal charm and powers of persuasion - "not guilty by reason of temporary insanity", anyone?
Fortunately Q-Lawyers are rare.
And people think physics is hard.

The same, to an extent applies to medical doctors (and here I risk bringing the wrath of my fellow science^bloggers upon me - who are disproportionally at the research end of biomed).
For the most part pre-meds and medical students are employed as practising medics. It is what they aspire to - to have predictable, well paid careers that while challenging and intellectually rich, deal mostly with the "known". They mostly deal with a constant environment - the laws of infectivity are mostly immutable (hm, prions anyone; and retroviruses give you a wee twist), and broken bones do obey the laws of physics, including the Equivalence Principle - and we know at some level what it is MDs do. We can explain it to our great-aunt who wants to know "what you are going to do with all this studying". MDs do a lot of application of their knowledge and principles to a fairly deterministic series of actions - notably diagnosis is heavily reliant on differential diagnosis through what are essentially deterministic binary trees; and treatment is then either deterministic or at worst trial-and-error probabilistic. Or wrong, but we've already talked about what lawyers do.

And, yes, much of this work can be applied to determistic automation or look-up tables - you see this happening, consider "tax preparation software", or "do your own real estate transaction packs".
There are still complications, and someone with the experience and background knowledge will almost always outfperform the amateur, but that does not change the underlying reality.
Interestingly an acquaintance of mine at a major biomed corporation supervises a project on automated diagnosis algorithms - any decade now...

In contrast, physicists are not so easily pinned down. (BTW note there are only 15,000 in the whole USA; the median salary is close to that of lawyers, but the high percentiles are significantly lower - lawyer salary is skewed high. Of course that is in part because very highly paid physicists are usually classified as "management").
Interestingly, at a glance, adding up the top legal subfields accounts for ~ 90% of lawyers but the top employment sub-fields for physicists only account for about 70%.
Even more interestingly, all the top paying physics categories are medical physics!

This is partial source of the perception of uncertain career prospects for (physical) scientists.
Just WTF do we do? I mean apart from the 1/3 or so who do pure research (of the few who actually make it all the way through) and the ~ 1/7 who teach at university (and some fraction who teach K-12).
It is not like there are a lot of jobs where you show up at 8:45 and some people come in off the street and ask you to find them a bound solution to the Schrödinger equation (with novel boundary conditions and potential, of course, can't just look it up!) - are there?
Some physicists are in occupations where they are basically a knowledge repository - what material can satisfy these specs, is there anything that can do such-and-such; is this possible in principle. But the sheer difficulty of describing such a job, and the absence of an immediate example, illustrates the difficulty. And why US corporations increasingly eliminate or retire-out such positions in favour of engineers and MBAs.

What scientists do is ill defined. There are few of them. Seriously, in Life, Physical and Social Sciences the only subcategory to break 100,000 are "market research analysts", followed by psychologists (just under 100k).
There are only 1000 astronomers in all of the US! (Coincidentally, the population of Iceland is almost exactly 1/1000th that of the United States...)

So we leak, and we lose people, and it is mostly a one way street, there are English Majors who became top research physicists, but not many... but I know physics majors who switched to english. Some even become market research analysts, if only from sheer force of large number statistics.