PhD grind part X of N: Results from a two week time diary

I'm TAing with Jen Mankoff now, who's somewhat of an enthusiast for time management, so I'm trying to learn a few things from her. One thing she suggested was a "time diary" - just write down everything you do, so you can find out what's actually taking up your time. I did it for two weeks, and here's what I got per week (average of the two):

Research: 19 hrs
TAing: 12 hrs
Class: 4.8 hrs
Email/logistics: 7.6 hrs
Socializing: 6.7 hrs
Waste (checking the internet, etc): 4.8 hrs
Other (walking between buildings, lunches that didn't fit in other categories, fighting with the internet when it went down, WC breaks, etc): 2.5 hrs
Total: 57.2 hrs

Things I learned from this:

• I'm not doing so bad, part 1. I'm putting in a lot of time into research. (This was on the week before and after the CHI deadline, so "research" is a little higher than normal. So if you're reading this and you think "oh, I'm such a lazybones, I only work 50 hours a week", two things: 1. these are perhaps abnormally high weeks for me, and 2. count it out yourself, you may find you work more than you think.) But I'm at least putting a lot of time into "research", which is good.
• I'm not doing so bad, part 2. There's not a ton I could cut out. I guess I could cut the "waste" time down, but I don't think I'll ever hit 100% efficiency anyway, so 90% seems not so bad. Maybe I could cut down email/logistics, but there'll always be a need for some of it. I guess I could cut down socializing, but that... seems wrong. This "socializing" is the all-important "networking" if you want to be super utilitarian about it - this all may further the all-important career. Attending department lunches or lab group meetings, meeting visiting profs, hanging out with PhD friends and chatting, whatever. And no, I don't actually think about it as "networking" while I'm doing it.
• I thought class was a big time suck. Maybe not. (though, again, CHI weeks, I really pushed class out of the way. I spent a ton more time on class the week after I time-diaried.)
• Grad school is a great environment to do research on anything you want... after you finish your required stuff. And there's ~35 hours of required stuff a week. ("research" counts all the time I spent on research, including filling out IRBs etc, so even some of that wonderful 19hr chunk is not so wonderful.) Which means, if you're a 40 hour worker, you'll have 5 hours to do RESEARCH, and you'll be frustrated. If you're a 70 hour worker, you'll spend half your time doing RESEARCH, and it'll be great. ... Be warned.

Edit: related is this post where this guy Togelius comes to the exact same conclusion as me, but frames it as "increasing marginal utility." I guess it's an optimistic way to look at the same thing.

Or... I could also get a job writing dumb software for 35 hours a week and then do research in my spare time, and be equally effective. (and paid a lot more, and I get to crank out some dumb software in the meantime.) Well, except I wouldn't be equally effective; in those 35 hours, I get these 3 benefits:

1. I learn something from class, and something from teaching, and I guess something from filling out IRBs and stuff

2. being in the university gives me access to papers, conferences, research funds to run a study, etc.

3. I make actual friends while "networking".

1 is true, but I'd learn something from writing dumb software too. 2 is true but unfortunate; I mean, the system shouldn't exclude people who just don't happen to have a university affiliation. 3 is true, ok.

It still feels like a lot of waste and frustration for those three benefits.