Do the hard stuff first. Then everything else looks easy. Make a mess of the code, then fix it and make it efficient.

I lose interest after doing the hard stuff

I'd add: if the task is a MUST - do the hard stuff first

If you're working for a company that gives you a project "kind of like an MVP" absolutely do the low hanging fruit first. As it might be scrapped before you even get to the hard part.

I was thinking of the learning bit.

Agreed. Solving the hard stuff first may influence all the stuff around it. And helps you get a clearer overview of what still needs to be done.