An advanced trick (not sure why, it seems easy enough) is to point someone in the general direction, wait a day, then sit down with them for 60 minutes and ask/talk to them about what they're planning on doing. Then you get the creativity without the surprise. Maybe even check in on them a few times as work goes on.

The bad engineering managers I've seen seem to suffer from anthropophobia when they aren't giving orders in a meeting room.

Fully agree. Unfortunately, not everyone wants a general direction. A list of things to do and by when so they can plan out their day and only do what’s necessary yields baseline expected results and frees you to give more attention to team members who do want the freedom to choose their tasks and contribute to strategy, etc.

Worst are the people who both don’t like the “I need you to …” task giving and do absolutely nothing when left alone or only given a goal and a direction. How do you deal with it if you are working on an industry where performance adjustments and feedback can take years and you can coast without doing anything four more than a couple of years?

I count being horribly disappointed as being surprised, just with the sign reversed. That's the risk you have to take as a leader. But, yeah, if it's a high stakes, do or die situation, do what needs to be done, even if that's something like getting 3 different groups to work on the problem independently, or something.

Long term, if you have the right team, the positive surprises will more than cancel out the disappointments and the "meh" results. But, if it's a life or death type matter, fuck the long term -- you have to survive the short term before you start worrying about that. You can't manage or lead an organization that's at war the same way you would one that's at peace, and nor should you try.