In the past, I’ve been reviewing myself on a quarterly basis to track my progress against my goals. Again, I’m here to inventory my goals and projects, explain how I’ve accomplished them (or fallen short) throughout 2015 Q2, and then elaborate plans, goals, and projects for 2015 Q3. This is largely in response to my previous review.
Note that despite this review being posted so late in July, it only covers the time period of 1 Apr 2015 through 1 Jul 2015. I have not added any additional information about things that have happened after 1 Jul. I’ll endeavor to start doing these earlier, in the future.
My Day Job
I still work as a data scientist at a start-up in Chicago.
My core goal was to think harder about improving as a programmer. I did that in “One Year Out: How I Became a Better Programmer”.
What Went Right
As of the end of 2015 Q1, I’ve now been programming professionally for one year -- three months as a software engineering intern and nine months as a Data Scientist.
In the past reviews, I’ve always looked back and admired how I’ve been improving at becoming a better programmer. This quarter is no different.
I studied writing good code more seriously this quarter, doing many of the things I suggest in “One Year Out: How I Became a Better Programmer” -- reading books and pair programming.
I also started branching into learning the beginnings of new programming languages, as seen in my attempt to solve Project Euler problems in many different popular programming languages. I’d eventually like to get to the point where I can start using programming languages because they’re the right tools for the job, rather than just because I’m comfortable with them.
What Went Wrong
I spent much of this quarter caught up in a project that was supposed to be three weeks long, but dragged on to take over two months. I still firmly believe the reason for this is not my fault (and, to be fair, no one at the company seems to put any blame on me, even though people are annoyed). The task just required a lot more infrastructure changes than we first anticipated, and I fell down quite a few long roads of things that ended up not working.
Every step it seemed like this would be the last hard step and the rest would come quickly. And I would be pretty good about not wasting too much time on any individual step and making sure to get help. But sure enough, once that step was complete, the next step would also be really hard. I didn’t really figure out the whole meta-system until it was too late.
Ideally, had I recognized this earlier, I could have substituted the project for something much simpler to “buy myself some time” to figure out the more difficult stuff. However, this wasn’t as bad as the two months wasted from Q1, since (a) this hasn’t been as long and (b) it hasn’t been nearly as avoidable in retrospect.
To dwell on things that are under my control, I think I’d like to focus more on getting back on the road and doing things. I can handle projects that slog on, but I’m far happier if I also allow time to do projects that go quickly and carry with them opportunities for innovating technically and solving challenging projects.
I’ve introduced a metric called “number of times a technical innovation has taken place within my day job work” and I’m going to aim to aim for at least three a week. I don’t know if you can force innovation by making a goal for it, but it does seem a lot better than measuring hours and I think it will help me fight the tendency in myself to allow my days to slog on and just become a day of meetings where nothing of substance gets done.
Another thing I’d like to do is learn more about how we put predictive models into production. This seems like a useful skill with a lot of opportunities, but I don’t understand it as well as I should. Now that I’ve done a lot of work on gathering data, munging data, and fitting models to data, I want to complete the last part and earn the title “full-stack data scientist”.
I’m also still going to try to learn more about programming, but I’m moving that into my next section about learning goals.
Secret Entrepreneurship Projects
Since Sep 2014, I have been working with two friends, fiddling around with several different project ideas, exploring entrepreneurship. By Q2, we had two core projects to focus on: (1) a framework in R for developing predictive models more quickly and (2) another product. I’d say more, but I still legally can’t.
Launch the modeling software project as an invite beta ASAP [fail]. We did not successfully launch the predictive modeling framework project as we ended up prioritizing the other project instead.
Launch the modeling software project as an open beta before the end of Q2 [fail].
Be more confident that the other project is actually launchable [success]. We ended up getting a real lawyer and paying a lot of money to find out that this sort of thing seems reasonably possible. I still have no idea how far it can go, though.
Move total revenue to >$1K [success]. We met this goal, but I’m not really in a position to say how or why on a website that anyone can read.
What Went Right
During Q2 we really hit our stride with being able to devote a significant amount of time every week to moving our projects forward -- it’s just that our projects ended up taking more time than anticipated.
What Went Wrong
Nothing really went wrong this time around -- I think we worked on our projects to the best of our ability given our various constraints. For the predictive modeling project, we made the choice to do more programming work on it before deciding to release it and then we made the choice again to put it on hold to try and get the other project to the point where we could see if people are interested.
Now that the second project is getting to the point where it needs more business development than technical development, we can circle back and try to finish the first project. And if it makes sense, we might even try to finish some projects that we considered earlier than that.
- Launch the modeling software project.
- Figure out whether the other project will work.
In Q2, I’ve been learning R programming (to get better at my job), mathematics (to get better at my job), and Chinese. The Chinese is a fun hobby and helps me bond with my girlfriend who is fluent in Chinese.
Finish Advanced R [semi-fail]. I finished 55% of Hadley Wickham’s Advanced R during Q1 and I wanted to finish the other 45% within Q2. Unfortunately, I did not realize how long the last 45% would take or how little time I’d have in the end of Q2, so I only made it an additional +16%. I’m still happy with my progress here, though.
Master >75% of Khan Academy’s Algebra II [semi-fail]. By Q1, I had “mastered” 40% of Algebra II on Khan Academy. I thought that with enough of a push, I could finish the remaining 60% and move on. Unfortunately, this turned out to be far from the case, given how often Khan Academy forces you to review. By the time I had “mastered” 50%, the review became unbearable. I also started feeling that I was losing out conceptually and not spending enough time on the bigger picture. So I decided to move onto learning Calculus instead.
Have >300 words in long-term memory on Memrise [semi-fail]. I ended up abandoning Memrise. I decided that I’d rather just learn on Skritter, because the mobile app was much better and the two were overlapping more than I had thought.
Learn >75 words on Skritter [smashing success]. In Q2 I actually reached 112 words learned, 86% more than I thought I would learn.
Finish the Chinese 101 Coursera class [semi-fail]. I ended up finding the course too basic and I kept getting annoyed at Coursera’s new verifications process. Instead, I got enamored with Growing Up with Chinese, in which I completed the first fourteen lessons.
Complete >5 more lessons in Rocket Chinese [semi-fail]. I only completed three additional lessons, not six.
What Went Right
The biggest success was switching to learning Skritter on the mobile app rather than the web application, which both expanded the amount of time I could spend practicing Chinese and expanded my motivation to actually do so. Now I can mess around with the mobile app when I’m waiting in line or on the subway.
Another thing I was happy with was figuring out what the best resources are for me, so that I spend more time focusing on good, non-overlapping resources and less time focusing on bad resources. For example, the Coursera class was good in the beginning, but quickly became too basic.
Overall, this created a good workflow for me where I could listen to Chinese audio lessons on my walk to the subway and use Skritter while actually on the subway. Given a commute of 30min each way over five days, that would be five hours of Chinese practice I could do each week without sacrificing any “productive” time. In practice, though, I tend to be less motivated to actually do this every commute, especially in the evenings.
Similar logic applies for my goals of learning math or programming, though these are not possible to do while computing, because they require enough space on the subway to sit down and open a laptop, which I rarely have.
What Went Wrong
Again, my progress was not particularly steady. I guess this is to be expected as motivation comes and goes and other pressing life events come in. Toward the end of Q2 I moved to a new apartment, which was quite draining across the board in allocating time to learning. Also, as I got more and more tied up in my current work project that was dragging on, I spent less and less time at work learning, eventually dropping things down to zero.
Lastly, my goals for Q2 ended up being too ambitious, again. I think this time I’m going to set my official goals at 50% of what I think I can do. I expect that once reality sets in, those 50% goals will turn out to be stretch goals. ;)
Finish Advanced R. For real this time.
Learn Haskell. A co-worker of mine is offering a twice weekly Haskell course at our work and I’ve been enjoying it so far. It should complete by the end of Q3. I’ll also supplement by skimming “Learn You a Haskell”. Together, these two things should set me up to try professional Haskell development by Q4.
Learn Statistics by reading through “Introduction to Statistical Learning”. I think I can get through Chapter 4 by the end of Q3. That would put completing the book by the end of 2016 Q1.
Learn Chinese. I’ll try to learn 40 more words on Skritter, complete three more lessons on Rocket Chinese, and complete through lesson 17 on Growing up with Chinese.
Learn Cooking. I’m not sure how I’ll go about doing this yet. I’d like to explore a bit before setting concrete goals. But I’m sure I’ll enjoy reading The Four Hour Chef.
Personal Health / Exercise
More important than nearly anything to a healthy and productive life is really nailing the core three: exercise, sleep, and food. Unfortunately, I’ve struggled with these three so far my entire life.
Get to work by 9:30am more often than not. This didn’t happen, but I’m not convinced that’s a bad thing.
Go to gym at least once a week and running at least once a week. This did not happen either due to one week being very dominated by moving and another week being very dominated by trying to move the start-up idea forward. Overall, going to the gym and running both increased significantly, however.
What Went Right
For the first time in a long time (perhaps ever), I’m cautiously happy with my diet, sleep, and exercise habits.
For diet, just like last quarter, I’m still very happy with the health of my current, vegetarian diet. I think putting it on Beeminder helped me keep it consistent and see that I do have some weakness in not eating enough fruit, which I think I can easily address.
For exercise, I got to the gym and I went running much more than I did last quarter, though the habit still was not that regular and not as consistent as I would like. I think the time I go to the gym matters a lot for whether I’ll actually do it. Previously, I would try to go during the middle of my workday, which didn’t work very well because I would often get stuck in an unexpected meeting or feel like I should finish a particular project instead of leaving. So instead I moved my gym time to the end of the workday. This ended up being phenomenally successful until I started getting in the habit of staying at work until after the gym closed or going immediately from work to working on the start-up. I’d love to try going before the workday, but I’m not that motivated in the morning...
For sleep, I finally got to a position where I was not constantly oversleeping and feeling crappy in the morning all the time. However, I still have not fallen into a consistent sleep schedule. Instead, I’ve had two sleep schedules -- one really late, going to bed around 1:30am and waking up around 9:30am, and one really early, waking up around 6:30am and going to bed around 10:30pm. The early morning schedule feels the most virtuous to me and syncs better with the people I care about (and work), but -- as much as I hate to say it -- the late schedule is overall more productive for me and feels better. I’m not really sure why that is.
More generally, I think committing to less ambitious goals helped, since I ended up overshooting my less ambitious exercise goal all the way to be being better than my previous best. And I think Beeminder helps a lot for goal tracking and keeping the goals in my mind. What I did have trouble with, though, was maintaining a consistent definition of a goal, especially with regard to my shifting sleep schedule.
What Went Wrong
I wouldn’t say anything in particular went wrong, though it would be nicer to have a more consistent sleep and exercise schedule.
I bought a wake-up light, but it didn’t seem to help much at all beyond the first week. I think I just adapted to it. Eventually I moved into a new apartment with a window with actual light from the outside and I think that’s been helping more recently.
I didn’t end up trying to go to the gym regularly with a friend and I didn’t put any additional effort into improving my diet by preparing food in advance.
Honestly, if I could keep things where they are at or improve on them even a little bit, I’d be pretty happy.
.impact and Charity Science
Charity Science is a Canadian-based non-profit that aims to increase public awareness of GiveWell’s recommendations and to figure out the best ways to effectively fundraise on their behalf.
.impact is a decentralized network of volunteers working on EA projects.
What Went Right
My goal for Q2 was to not get too involved in Charity Science or .impact so I could keep my focus. I succeeded in that goal and both organizations ended up doing very well anyway!
The biggest victory here was delegation. I helped find and coordinate new volunteers to work on EA projects, including the EA Forum. I created a list of possible volunteer projects we’re still looking for people to cover. Ozzie and Haseeb ended up organizing a workathon structure on Sundays for people to contribute their time and we’ve seen a large spike in people’s contributions. Huzzah!
Charity Science ended up doing cool things too, but I don’t really have much basis to take credit for any of it. I’ll let them tell you more about it soon. It also involved doing very well with delegating to volunteers.
What Went Wrong
No particular mistakes to note on my behalf, mainly because I didn’t do anything personally other than delegate. ;)
I moved to a new apartment; my third apartment in twelve months.
What Went Right
This move was a good one for my long-term productivity and social happiness, because I now have a space that is large enough I don’t go stir crazy and that I can work from a desk and a chair instead of from my bed. I’ve already noticed a good deal of productivity increase since I finished my move and it allows me to have guests over, which is pretty awesome.
What Went Wrong
Unfortunately, however, the flip side of this was that I needed to devote about two full-time weeks to (a) looking for apartments (including conducting a wider apartment search than I’ve ever done before to see what I really want in an apartment), (b) signing the lease (which ended up having complications), (c) packing up all my belongings, (d) moving all my belongings, (e) unpacking all my belongings, and (f) setting up the space (including furnishing it with more furniture than just a bed!).
I’m happy with how all these steps went. In particular, I was grateful to help from my girlfriend for help with looking for apartments, unpacking my belongings, and helping me set up the space. I’m also grateful to my friend Josh for help with moving my belongings.
I’d like to not move again for at least another year.
I had three goals mainly focused on consolidating knowledge I already had.
Expanding summaries of books that I’ve read and liked [goal abandoned]
completing an organized document with all of my link posts [goal abandoned]
completing more links posts [partial success] -- I finished one.
What Went Wrong
I didn’t focus much in this area, instead allocating capacity to my other areas which needed it more. In Q2 I read five books, all on Audible (down from 11 finished in Q1). I generally found myself less motivated to read books as I found I was hitting diminishing marginal returns. Instead, I allocated my listening time toward Chinese lessons and listening to podcasts.
I had hoped to spend a good deal of time working to consolidate knowledge from the books and links I had already read, but this turned out to require more productive time than I anticipated (I couldn’t just do it casually while commuting) and I was already stretched for time as it was.
There’s a large unsolved problem in how to actually apply the skills that you learn, so that they’re integrated in your life in a deeper way than just being something you know surface-level. I think Paul Graham is somewhat right that this happens automatically, but I still can recall many times in my life where I’ve failed to apply something I had already learned and would have benefited had I actually applied it.
Given that my capacity is already stretched and that I don’t have a lot of ideas at the moment, I think I’d like to spend another quarter just ruminating generally on how to approach this problem, with the goal of producing concrete Q4 goals in this area.
I want to include reflecting on my social life as a part of this guide because it’s important to me, but I don’t want to say too much here in order to respect privacy and because reviewing it this way still makes me feel weird.
My big goal here was to figure out a system for improving my consistency in communication with everyone I want to be in touch with. I hastily put one together at the end of the quarter. It’s working reasonably well so far, so I call that a success.
What Went Right
My girlfriend: I think our relationship overall continues to grow stronger. Soon she’ll be moving closer (but still long-distance) and I should be able to see her every other weekend, which will be great. We’re also seeing each other a lot over the summer before she starts her job.
My parents: I’m still calling them regularly. I am excited to visit them in person for the 4th of July. I hope that they’ll come up and visit my new apartment sometime soon.
My brother: I’ve been keeping in touch regularly with my brother as well. I will visit him on the 4th of July and I will be seeing him again soon at Gen Con.
Local friends: Also going well, but overall a dip from the previous quarter, due to me being busier. I have new friends who have recently moved to Chicago who I hope to see as well. I’ve been in touch with some of them to schedule a time to hang out. My Q3 goal will be to hang out with all of them.
Distant friends: This has been going better as I’ve been trying to keep more in contact with people over Facebook and the phone.
What Went Wrong
I want to involve my girlfriend more in my day-to-day life and I’m struggling to (a) remember that when I make plans and (b) figure out how to do that.