This is related to what Kent Beck calls the "Red Test Pattern". When doing Test-Driven Development, don't stop when the last test is green and refactored; commit/push your code, and then write that next test, the one for the next functionality you want to realise. Check that it fails. Then shut down.
Next day, you'll be sucked right back in because of that damn test failing.
For me, it's also a good way to keep habitually doing TDD.