Maybe it's better to let people do explicit side projects as a creative outlet so that they don't shoehorn them into production projects.
I completely agree. I've worked at a company where we had explicit weeks taken out of the schedule to do "app building" where we were expected to do a mix of kicking the tires on the product and experimenting with random tech we wanted to play with (subject to the Legal Overlords and their hatred for OSS, at the time). Speaking with my manager hat on, it worked quite well for exorcising those demons.
I've had good luck in the past doing these kinds of side projects as tools -- they're useful, but don't rise to the level of Mission Critical. This avoids the problems mentioned in the parent. (Organization becomes dependent on bleeding edge languages/libraries, etc.)
That said, I definitely support his statement that people should have some hacking / learning time budgeted, even if only to avoid the scenario I mentioned above.