1) I am my own greatest enemy
Through grad school and my postdoc, I always wanted to work as much as possible. Research felt a bit like a competition and I figured the more time I put in the more I would get out. I wasn't necessarily wrong, but I did get a bit burned out at times. The most difficult feeling I've encountered since having a munchkin is that it's inevitable to feel like you aren't accomplishing as much at work as before. Truth is you probably aren't. I quickly started to view this feeling like I view my own impostor syndrome though...that little voice is always going to be there no matter how much I work, so I might as well be satisfied with what I do get done. Each institution has its own bar for "good enough". My guess is that this bar is lower than the one in my own head. This has by far been the best year I've ever had for publications and grants, and that has to be OK. If it's not, then c'est la vie.
2) Efficiency and saying no are key
I had to be more efficient to survive if I was only on campus for X hours a day. I figured out what I could get done at home, and what I could only get done at the lab. I learned how to work from home and became quite good friends with Arizona's VPN system. While I highly valued walks around campus for getting my thoughts straight, I don't really have time for this anymore and found other ways. I became a much better planner than I was before. Although I spent less time preparing for lectures, I still managed to relay all the information I wanted to (I think). In class, I tried to limit the emotional energy spent on students to a realistic level and set up firm boundaries (no responding to emails after 6PM or before 9AM). I gave more online quizzes to limit my grading by hand. I actually said no to some reviews. There is only so much time in a day and I learned how to cut unnecessary expenditures, because that's what had to happen. You have to make sacrifices, the key is learning and cutting out superfluous actions.
3) Your community is key
I won't go too much into it, but both my wife and I worked full time during this last year and somehow managed to avoid daycare. She is, quite frankly, a superwoman.
I also wouldn't have been able to survive the last year without the support of my department. They value my contributions, and understand my need for work/life balance (as far as I know, check back in two years when I'm up for tenure). Arizona even started granting paternity leave last year. I don't know how it is other places, but I can only hope that the tide is turning like this.
I've also been able to survive because I've lucked out and had very good people in the lab. I've tried to embrace an environment where the only metric for success is having a set of goals and hitting certain checkpoints rather than working a specific number of hours. I don't question vacations or time spent with families, and magically enough we still come up with really good data. I fight for them whenever I need to. They know what I expect from them, and I know what they expect of me. I give everyone a lot of rope and foster independence. This certainly doesn't work for everyone, but it's the only way I can survive as a PI.
4) Sleep is not overrated
When there is an infant in the house, your sleep patterns go out the window. This translates to a lot more spelling mistakes in documents (and lectures!) and a much lower tolerance for the everyday BS you deal with in academia. My advice is to proofread documents first thing after sleeping or a nap (or get others to proofread). My other advice is to try the best you can to avoid saying things you shouldn't while sleep deprived, even if you're teaching a class of 100. You'd be surprised how some of your comments are actually interpreted by the students, so it's better to just shut up.
5) The fringe benefits of a munchkin are awesome
Research is hard. Our lives are filled with rejection day in and day out. There is no better cure for rejection emails than coming home and playing with my daughter. Absolutely nothing. Her little smile puts everything else into perspective.