One point he doesn't mention is that there are many interesting problems in the "real world". Lots of academics just point to industry in general and say, "no freedom, no thanks." I've held a few jobs and all of them presented with unique, interesting, and challenging problems. I've had the freedom to choose my own approach to solving problems and met up with other academic-minded people to have good lab-meeting style discussions about how to tackle a project. Industry positions can be pretty attractive.
The thing about the software industry is that "interesting" tends to be defined down -- for example, you'll be working on a CRUD app, but the "interesting" part is that you're doing it "at scale". Or you'll be doing some "interesting" refactor of a hard piece of software, that isn't interesting in any other way. Or most insidious of all...there's nothing really "interesting" in your job, and you just get inured to the day-to-day nature of the work, which always tends to look the same. The difference is in the big picture, and for most software jobs, the big picture just drains your soul.
I'm pretty lucky in that the project I'm working on now has a truly "interesting" technical component, but it still doesn't compare with the idea that I could be working on an interesting technical problem that also might lead to a new antibiotic or vaccine...or just discovering something new. I miss that part.
Every now and then, I casually leaf through job ads and I have no idea how to tell which are which. I think that's what scares me. I get this feeling that when it comes time to join industry, it's just going to be a crapshoot---maybe I'll get lucky and maybe I won't.
(Note that this isn't a criticism lodged in comparison to academia. I've already made up my mind that I'll be leaving once I graduate.)
I worked with a huge crop science company for four weeks and got some valuable insights (mostly, I had to justify every little step, and the local boss was a huge fan of micromanagement: meetings, meetings a few a day. But they did have a lot of money and were willing to throw it at possibly not profitable projects; and people did proper 9-5, not 9-5678910 like academia)
I know there are places that are more open-minded towards publishing your results in the open literature, but the majority of companies seem to put roadblocks in the way. I know some people who have done very interesting work (one at Exxon, for example) and will never be allowed to talk publicly about it, because it's their company's trade secret now.