Scientists are just like Machine learning pixie dust. It's fun to say, but you'd be hard pressed to find a small business that needs an astrophysicist. I don't mean to pick on astrophysicists by the way, pick any random phd, a small business probably doesn't need that.
Maybe I'm biased because I did research in Particle Astrophysics and now work as a developer for Particle physics experiments, but the problem with most small businesses is not that they don't need any random PhD, it's that they lack the foresight to understand what a person trained in particle astrophysics can actually bring to the table. In modern day terms, a degree in astrophysics is a degree for the polymath.
The valley and wall street are a bit different... people are willing to give you a shot at almost anything in any seemingly unrelated field because they think you'll probably be able to handle it, because you've solved hard problems that are often similar in scope. Predicting the market is pretty similar to reconstructing a collision, folding proteins, or predicting molecular orbitals. Cloud computing isn't that different than grid computing. Statistics are statistics and models are models, and companies are finding out new ways they can benefit from information processing by utilizing the people who pioneered it.
I do agree though... the Main Street small businesses, the ones that I think the GOP often fantasizes about, aren't going to usually need someone with a PhD in whatever.
A very large number of those 2 million unfillable jobs mentioned above are for basically "Someone who knows linear algebra and statistics really well."
Scientists fit the bill -- any physics phd who can't find a job can join the finance or ad tech industries by sending out a single resume.
Intel (a small business in 1968) was started by two physics phd guys. Adobe (a small business in early 1980s) was started by two phd guys. Google (a small business in 1998) was started by phd-dropouts. I think there are many companies started by phds.
If a small company is trying to solve sufficiently complex problem, like what Intel (in 1968), Adobe(in 1980s) and Google(in 1998) were trying to solve they need a phd. If a small company is building a new social network or social app it might not need a phd.
Right, because PhD is like pixie dust you can sprinkle on a problem.
Sorry, the point i'm trying to make is it's easy to be intellectually dishonest about the complexity of problems people are facing (sorta like the ML article earlier). If your small business needs new research to survive, you've lost. Very few small businesses are limited by complicated third order effects. they're limited because the owner doesn't feel like doing inventory (or writing more tests, or calling more potential customers) over the weekend.
Sure, you can find counter examples here and there, like special hedge funds or old hippies with geology shops, but most small businesses know 5 things they could fix to make it better.
If they're really that super-efficient and organized, they're either a large business that turned into a small business because the field is dying or they're not a small business.
Well as Aaron Hillegass says in one of his books, a degree in astrophysics may not be all that practical, but it helps you overcome challenges where doubt might creep in about having the intellect to take on a difficult problem.
1) I am having trouble solving this problem, I must not be smart enough.
2) Wait, I have a degree in astrophysics, so I am smart, this must be really hard.
