I am beginning to think of myself as less of a science-fiction writer and more of a sciencey-fiction writer. I realized this a while ago as I was talking to a friend who was agonizing over the mechanics of dimensionality in her current book (she was flipping frantically through Stephan Hawking's A Brief History of Time and watching Michio Kaku videos about string theory.)
My advice to her? Read and watch all the stuff, make your book sound science-y with the things you learn, but don't shred your brain trying to solve a problem that even Michio Kaku hasn't solved.
My thought is that readers and viewers know the reality of things right now; they know that extra-dimensional travel is not (currently) possible or that Mars's lack of a magnetosphere makes terraforming the planet a non-viable prospect, but that's not why they bought the ticket. They bought the ticket because they wanted to imagine for a while that it is possible.
I think science-fiction needs to look forward, to be creative and spark imagination--and that's not always possible to do while incorporating every known obstacle. Space operas have to project to a future time when deep space travel has become feasible, books about robots have to imagine that the thousand million problems hindering actual AI have been dealt with. (Doctor Who, of course, is the best example of being science-y--and nobody cares because it is so delightful.)
So there it is. Be creative, be inspired, look forward. Use what is known now, but don't be afraid to envisage a world where some of the peskier problems have been already solved. You'll need your energy for the writing all the love scenes in your robot, terraforming space opera.