A recent PhD graduate whom I know well, who has been doing quite well in graphics research, just took a non-research job in a non-graphics position, among several postdoc offers from top research institutions.
The reasons he cited are numerous, but the 2 main ones are:
. He likes graphics research, but he is getting bored and tired of the SIGGRAPH game, including deadline crunch and dealing with reviewers.
. He would like to learn something new and try a different life style.
In a nutshell, I think he is burned out. I hope I have done my best to help him achieve work-life balance (I did that quite well myself, even when I was a grad student), but I guess it is just too hard for normal people to be much disciplined.
Then I realized that I probably also had some kind of burnout around my graduation. I took a non-research position as the first job, even though it was in a graphics company (NVIDIA). I also wanted to try learning new things (hardware architecture) and living a different style (engineer).
So I guess it is probably OK. People are not meant to be doing the same thing all the time. This is also why I like to try different job sectors and geographical locations.
There are two things to watch out, though, all based on my personal experiences: passion, and rust.
Passion
I have a very simple rule to choose jobs: do what you really like, and be very good at it.
Sometimes, when people get burned out, they might temporarily settle for something that they neither really like nor really be good at. But eventually, you will know if the job is not for you. I did not realize how much I like doing research until after I was not been able to spend enough time on it for about 3.5 years after my graduation.
The important thing is to get out there as quickly as possible. Otherwise, you will eventually become one of these people who are not really happy or good at their jobs but also cannot quit.
Rust
People tend to get rusty for skills that they have not practiced for a while. This is particularly so for advanced skills, like research.
So, make sure you do whatever you can to be active in research during your non-research job. Otherwise, you might not be able to come back, even if you want.
I have been trying my best to be engaged in research during my NVIDIA days. I even managed to publish a single authored graphics hardware paper. But it still took me about 2 to 3 years to get back in shape for SIGGRAPH after joining MSR. The difference between SIGGRAPH and other graphics venue is like the difference between playing professional sports and working out in a gym.
I guess SIGGRAPH is probably an extreme case, but I hope you get what I mean.