Maybe my brain is easily tricked, but I've found that switching jobs lets me make-believe that everything is great for awhile, and the reality of building another CRUD app doesn't set in until a few months to a year later.

The reality is, 99% of companies aren't actually working on something interesting. They're just making incremental improvements on something that's been done many times, and 99% of the work is grunt work. The kind of research-driven code crafting us developers wish for simply doesn't exist in the programming world. Maybe someone out there is working on self-driving flying cars, but I willing to bet even that isn't as glamorous as it sounds. It's probably mostly turning academic algorithm code into something usable, a.k.a. rote transformation of math into code.