My girlfriend right now is trying to figure out her postdoc in neuroscience.
While I fully support her passion, the industry of academia seems completely irrational from almost every sense and full of traps. While the end goal of a tenured professor is exceedingly stable, getting there requires heroic acts of risk, going with the flow and instability that even when compared with the world of startups seems insane.
After 4-7 years on a PhD (which is fraught with the possibility of your PI losing funding, vets shutting you down for weeks on end, the university messing up shipping of things that need to stay frozen, strange policies of academic journals, endless bureaucracy, the university threatening to cancel health insurance or not pay you for a few months despite them wanting you to work) you exit with the ability to make less money at a Postdoc than you had with just your undergraduate degree, and certainly less than you'd make if you had just stayed as a lab tech for that term.
In admission to a Postdoc program, you've gotta deal with PIs who seem to check their email in more archaic ways than RMS and seem to play games with you with their intermittent responsiveness. The chances of you getting into a place doing the research you'd like in a geographic region that has any semblance of culture or livability. Of course, all the pitfalls of your PhD are still present, as funding is constantly on the brink and your PI might die or retire without notice- which essentially ends your academic career.
Then maybe after a postdoc (or two), you find a tenure track position (again, where this is... you seem to have little control over), which also seems to pay crap considering the amount of knowledge and experience you have.
All of these funnels seem to have a 20:1 (or worse) completion rate. Something always seems to screw up. Fortunately, my girlfriend is almost done with the PhD part, and the Postdoc part is looking more promising than it is for most.
Whatever relationships, life, family or culture you'd like to maintain through this seem nearly impossible. Now, I'm aware that 'softer' subjects are a little more flexible on parts of this. You don't need a lab with a half million a year funding to write the next great american novel or study the culture of people on 4Chan.
In comparison, startups seem much more certain. Move to SF|Boston|NYC, program awesome things, get funding, etc... no one's going to ask you suddenly to move to Alabama to work with the one startup in the US that does Haskell programming, but in science academia that's entirely possible. Worst case scenario, you go work for Google or similar. And after working for 10 years, they aren't going to pay you barely livable wages.
Unless you're running a center (which probably less than 0.1% of people entering into a PhD program ever will), the monetary reward almost never catches up. The risk of failure is high, and the alternatives for leaving are grim. Whereas if you leave your CS PhD, you can get a job at a startup... there are few places that are dying to hire PhD dropouts.