But do agree in optimizing the use of both time and money.

Often the things that give us the greatest satisfaction are some of the hardest things to do.

My hope is that with boxed serious work time, and a clear sense of freedom outside of this time, it will be quite easy to find sufficient balance and diversity in my life. Let’s see how this experiment goes…

I think another element that you touch on in your comment is the need for balance. Just because something isn’t work doesn’t mean that I shouldn’t limit myself in the activity. To be effective and to function well I have come to realise that I need balance between the different areas in my life.

The “like to do less of” metric seems a bit odd in some cases. For example, I’d like to do less “work”, but at the same time I’d also like to do more “work” due to benefits longer term. Maybe that’s not a very good example, but the point is that I often have opposing desires about things. Imagining myself to be Gates and asking what I’d still do seems easier to me.

Applying the metric to myself, it’s clear that some of my research activities are “work”, and some are not. The work stuff consists of things that I have to do to keep myself in this line of business now and hopefully into the future. If I was Gates I wouldn’t care about these things. Writing on my blog about AI is not work, even though it’s very close to what I do for a living. Taking the bus on a rainy morning is work (if I was Gates I’d take a cab). Walking to work on a sunny morning isn’t work. Going for a run isn’t. Reading a book on neuroscience on the couch in the evening with a warm drink isn’t work. Hanging out with friends on the weekend talking about the future of AI isn’t work. Tango isn’t work. And so on…

Ok, great, so what’s the point of all this? I think it’s important to distinguish between when I’m working vs. when I’m not. Both for efficiency and for my own psychology. For efficacy I need to schedule, say, 5 hard hours of work a day with the idea that long term I should hope to ideally reduce this down to perhaps 1 or 2 hours a day. On the other side, a key thing when I’m not “working” is to be aware of this fact so that I’m sure that I don’t stress or somehow feel that I am working. Most of the really interesting things that I’ll do will most likely be during non-work time, including my most interesting bits of research. I’ll also need to fit a variety of non-research things in their for my own psychological well being.

I currently have a “to do” list for things I want to get done, and a “done” list for what I actually did each day. It works ok. I’m now going to try 3 lists: “Work”, “Play” and “Did”. Scheduled work time is only for things on the work list that I wouldn’t do if I was Gates.

Will all this theory translate well into reality? Ask me in a month

One of the hardest parts to this for myself is simply defining what is work vs. what isn’t work. Tango, clearly not work. Writing parts of this finance paper, clearly work. But what about replying to comments on my blog that are about one of my research areas? Work, sort of? It’s a cold rainy Sunday in London, what if I read an interesting book on machine learning. Kind of maybe work? This might sound like a minor issue, but I’m finding that it’s a serious problem. With fuzzy boundaries semi-work things tend to creep into both work time and non-work time, causing the whole thing to come unstuck.

I’m not sure what to do about this. Maybe I need 3 categories? “Core work” which is stuff I really have to do: writing papers, trying to solve equations, reading papers relevant to my current research tasks, writing experiments for my research. “Play work” which is stuff that is good for my work, but that doesn’t have a deadline and is more fun and not stressful: writing blog posts about research, reading an interesting book or online article about some AI research that isn’t directly relevant to what I’m currently trying to do, dreaming up new research ideas and where I want to go in my research in the future. These things have long term importance, but whether I do them today, tomorrow or next months doesn’t matter much. Rule of thumb: if it feels like a weight on my shoulders, it’s not in this category. Finally, stuff that is clearly non-work like going to the movies or for a run. Then try to time box the “core work” part, perhaps?

The other thing I wanted to say, before this turns into a book, is about the 80-20 rule. The problem with this is that you can have a highly non-linear return with respect to achievement in some situations. Take our second year programming course. We made a computer calculus system and I made a neural network. These things went beyond what was required. When the CS department wanted to employ an undergraduate student to do machine learning for the summer, I was top of the list. As a function of the time that I spent to go from good to great to being the best candidate, I had a large long term pay off. I think this happens in quite a few situations. Sometimes, “going the extra mile” is enough to make the difference between getting the deal/promotion/new job/etc… and getting, well, often next to nothing.

Concerning time-boxing, that is something that I have thought a bit about too though in a slightly different context. If you consider the traditional project management triangle, the constraints are Schedule, Scope, Cost, and Quality. If a project begins to slip then one of these must slip. So one concept of time-boxing is to fix the schedule and then ensure that it is the scope that slips. Well, I’ve been thinking about that and myself personally. I’ve come to the late realisation that I’m actually quite pedantic and have prided myself on “if something is worth doing then it is worth doing well”. Therefore for anything I’ve done personally, I’ve tended to fix the scope and quality and I’ve allowed schedule (time used on the activity) to slip. However, I think I’d be more effective if I time-box myself. Applying the 80-20 rule, I can achieve 80% of my objective with the first 20% of effort and maybe I should just stop being so pedantic and be more effective.

Obviously this isn’t something I’ve learnt to apply yet. This comment has taken me a while to think through and type and maybe by time-boxing this activity I could improve my effectiveness!

My main problem seems to be a lack of time. There is just so much that I want to do and so little time each day to do it. Besides trying to shut more of the world out, or “selective ignorance” as you call it, I’m currently experimenting with something called “timeboxing”.

The basic idea, at least in my version of it, is that things have set periods and amounts of time. For example, work is 9 am to 12 pm, then 1 pm to 7 pm every day. No more, no less. Then, maybe, 7 pm to 11 pm is strictly non-work time. 11 pm to 7:30 is bed time. At a finer grain, a task might get a specific time slot. Then I just do the best job I can in the period of time.

For me, this seems to help counter act the problem where I don’t have a clear work vs. non-work distinction. Partly, I suspect, this is because I don’t really “work” in the sense you advocate as my job is also what I like to do. Anyway, the point is that relaxation time tends to bleed into work time, and work time tends to bleed into non-work time… to the extent that I feel like I’m not being maximally efficient in either. Or put another way, it’s better to work 9 hard hours and then head out for tango and drinks with friends, rather than 11 hours at work sort of working and sort of chilling out and mucking around.