I do agree with several contributors here that where possible/conditions permit, raw data flying either from take off to FL200 (for example) or from FL100 to landing should be practised, at least once a month. It may or may not have anything to do with this alleged incident, but it is a very useful exercise, and if done reasonably well, can act as a confidence-booster as well as improving one’s handling and scanning skills (and it keeps the PM awake …)
I appreciate some companies frown or don’t allow this - I think that’s a real shame, and I think it’s wrong. If you can’t fly at all without automation I would suggest something is wrong.
Besides, those of us that fly (much) older Boeings will have probably experienced the FDs (pitch and/or roll) mysteriously disappear at some point during take-off!