A PhD should be about improving society, not chasing academic kudos

> AFTER a PhD is finished they can go on and try to influence society using their skills, or not.

Exactly this. The primary skill developed during a PhD (for me, and those around me), is knowing the difference between:

1. solutions that can be pulled from literature (bake),

2. solutions that should be easy to assemble from existing solutions (buy),

3. what technology requires a small-to-medium delta above and beyond the bleeding edge (build)

All those are research, but far too often (3) is what everyone targets.

Basically every human endeavor "should" be about putting the general good over the individual good. After all, it's in one's own selfish interests for that to be the state of affairs; there are far more people not-you doing things than people who are you doing things, so you stand to benefit more if all those not-you people are doing work that indirectly improves your quality-of-life. ;)

The (sometimes literally) million-dollar question is: how do you align incentives to bring about that desired end-state? Ph.D's are not unique in struggling with this age-old conundrum; if the government is willing to pay for robotics research for drone warfare but nobody is putting up the cash to pay for nursing-care robot research, then we can make some pretty firm predictions what kind of robots we'll see mature and reliable in twenty years.

Disciplined Minds is a great read.

Early on the book discusses some data that indicates more highly educated people tend to be more supportive politically of the government than the population in general. (E.g. during the Vietnam war lawyers defending anti-war activists in the US were trying to figure out how to pack juries with people who would be favorable to their clients and discovered that roughly the more educated someone was, the more likely they supported the government and would disapprove of anti war activism)

The book makes the argument that gaining professional credentials is less about gaining knowledge/skills, but more about demonstrating that one can be trusted to work in an organisational and confirm to the assigned goals/ideology of that organisation.

One rough argument made in the book is for workers performing roles with work that can be completely specified by superiors, it's less necessary for workers to demonstrate they can conform to how their employer wants them to think, provided they do the work. For professional roles involving more intellectual work where an employee's day to day tasks cannot be completely specified in advance, it's important for such employees to demonstrate they can think how the organisation wishes them to think, so an important part of training for these roles is essentially political training.

> > The high level of dropouts is similarly worrying

Why? Some things are hard and those things will by definition have higher dropout rates. The fact that PhD programs have high dropout rates is not necessarily an issue unto itself.

> You're going to need 2-3 years of experience actually doing research to know if it's something you'll enjoy long term.

Is it realistic to expect a college sophomore to know whether or not they want to pursue a PhD program? I wish more undergraduate programs focused on original research but if yours doesn't you may not get any exposure to it until your senior year (if even then).

> Don't you need a master's degree to be eligible for a doctorate program?

There are 8 or 9 former classmates of mine with PhDs and only one went out of his way to get a Master's prior, and only because it was a much more prestigious institution (Georgetown) than our undergrad (unknown liberal arts college). Everyone else either entered their respective programs directly from undergrad or had industry/life experience first.