Okay, 'heresy' is overstating things, but this post on the "Unqualified Offerings" blog makes an excellent point, which can be summed up thusly: Attempts to get new people into physics, whether women, members of racial/ethnic groups under-represented in the physical sciences, or just your average white guys, are undercut by the fact that it's not a promising career path--if you measure it by how many people land good jobs directly in the field.
Yet unemployment rates for people trained in the physical scientists are quite low, and their salaries aren't bad (though they're not great, either, if you compare them to people with similar or less education in fields like business and clinical medicine).
'Thoreau', the blogger, is correct that a degree in physical science can lead to a wide range of opportunities. He's also correct in saying--or at least implying--that those opportunities are a bit off the radar and not so easy to find:
The reality is that the real value of a physics degree is similar to the value of a liberal arts degree: In all likelihood you won’t spend the rest of your career making direct use of the concepts and calculations mastered in physics courses. However, a physics education gives a person (1) a very fundamental view of science and technology (2) an introduction to an experimental culture that is very DIY and clever with indirect measurement, and (3) an introduction to a mathematical culture that is able to blend high-powered computing, back of the envelope estimates, very fancy pure mathematics, and the workhorses of standard applied mathematics. This sort of training generally leads to a lower starting salary than engineering graduates (fewer entry-level jobs that precisely match the major) but long-term prospects that are comparable to engineering grads (and often more flexible).
I'm not sure he's right about the salaries being lower than in engineering, but he's certainly right about long-term opportunities and the professional flexibility that physics training provides.
I get annoyed when people argue that we need more and more scientists (whether it's in physics or in other fields) while research positions are absurdly competitive. I get even more annoyed when they justify this with the hand-waving argument that "scientific training allows you to do all sorts of things"; it's annoying even though it's true. I get annoyed because it's damn hard (usually, though not always) to make the jump from physics to one of those "other things," especially if you want to earn a decent salary.
Employers tend to take a short-term approach when hiring (for a variety of reasons), with job requirements that are narrow and task-specific. Because these narrow, specific requirements rarely touch on quantum chromodynamics or the thermodynamics of black holes (except for research positions), a degree in physics--especially an advanced degree--is a very tough sell.
Those low unemployment rates would be a good bit higher if postdocs didn't soak up so much of the excess. Yet it's true that most scientists, physicists and otherwise, eventually find good work. In terms of earning potential relative to length and difficulty of training, you can definitely do better. But if you enjoy solving technical problems, you could also do much worse.
The problem is that latching onto a career after training in physics is much, much harder than it ought to be. Since the mid-'90's Science Careers (nee, Science's Next Wave) has tried to make it a little easier. But it's still hard