Ordering the Chaos (2020 Review)

Personally, the year I find most similar to 2020 was 2011. In both of these years, for various reasons the range of activities I could choose from was externally limited (in 2011 because of National Service and in 2020 the coronavirus), and I thus focused strongly on work and personal development. In 2011, this amounted to reading up a fair bit on mathematics and computer science before I started full-time at Imperial, which made most of first year fairly light; I caught up on the gaps in my knowledge from not having taken Further Mathematics in my IB years, and also learned a fair bit about Java and discrete mathematics. In 2020, I think I improved a lot at work in terms of how I manage a team, to the point where I’d question a good number of the approaches I used in 2019. I also spent more time than I expected refining my German skills.

Software Engineering

Transactional Storage

I am still a software engineer, still at Palantir, and still working on databases. This marks 4.25 full-time years there, 4 of which were spent on AtlasDB or Transactional Storage. I’ve continued to find working on the team interesting and challenging, both as the team’s remit has expanded (especially this year!) and my responsibilities in keeping things running and improving, as other engineers decided it was time for them to move on. I also became a people lead; in a sense, it’s nice to be a relative junior again, in this way. A junior senior, if you will.

AtlasDB merged with another internal team to form Transactional Storage; this made sense as there was often work crossing product boundaries, and also support rotations could be handled with a bit more breathing room. There was a double lead structure which was interesting; we were certainly in a place where we trusted each other to work on projects, and at least on my end I was comfortable with maintaining very minimal context on any projects the other lead was overseeing. That said collaboration was a bit more limited to support and a project at the end of the year. He then decided it was time to move on, which means that treating ongoing projects as such isn’t really an option any longer.

The company has started to become a fair bit more critical about individual growth and performance, and we’ve had monthly reflection exercises about this (for me this entails sessions for each team member with them, plus sessions with my lead and group-leads about myself). I’ve found these useful in terms of forcing myself to sit down and write feedback, though the monthly cadence can be quite a lot of work since I have to do five of these reflections each month and I generally try to put a lot of effort into them.

Working from Home

A big change this year was of course working from home; we started at the beginning of March (three weeks before the UK lockdown) and that continued throughout the year, even through the summer (and of course the fall/winter with the second/third/Nth waves). Before this I had a fairly full calendar, with a lot of syncs, coffees and external events; many of these were cancelled or stopped, and a consequence of this was I spent a lot more time working directly with the team. This involved daily pairing sessions with the newest member of the team as well as less regular sessions with others; if I was to help guide a team today this is how I would probably do it, unless everyone on the team was very experienced and independent. It wasn’t something I anticipated last year, perhaps because I thought everyone on the team was very experienced and independent anyway; I had the shortest tenure in terms of industry experience at that time!

Working from home was probably not good in terms of raw efficiency when I was working on a specific task – there, the comfort of the office, the quality of the network connections and the availability of external resources where needed was high. Of course, collaborating on an original design was going to be a lot more difficult, which I think meant that for some features a lot more needed to be written down in RFCs, or I ended up doing more of the scoping than I would like. However, it did serve as a forcing function to strip away things that may or may not have been necessary, and in this way I think I still reached overall 80-90% efficiency compared to what I would have in the office.

I am skeptical of the claim that this exercise reflects which activities are necessary and which are not, though; until late in the year when we started to create more opportunities for social interaction team morale felt much weaker than it was in the office, and even now I don’t think it’s on the same level. Furthermore, on a technical level I think there’s value in being able to have spontaneous in-person discussions that may run across teams, which hasn’t really been the case here.

Recreational Pursuits

Logic Puzzles

I’ve continued to participate in puzzle contests. On the Sudoku side of things I’m pretty consistent; in the Sudoku GPs this year I moved up from rank 66 to 52, and also finished at rank 28 in the UK Sudoku championships. The more general Puzzle division is much more unstable with a Puzzle GP rank of 89, which is an improvement over last year’s rank 92 though it hardly feels like one. There are certain types which I can do probably just as quickly as Sudoku (like TomTom or Star Battle), and then some which I just don’t seem to get (like anything involving drawing snakes). I used to practice puzzles a lot more aggressively, often burning through multiple 90- or 100-minute puzzle sets each week. I still find them fun and relaxing, though I don’t do as many as in the past. This is probably a balancing issue: my hobbies were previously primarily sudoku and puzzles, while now they have to contend with Deutsch, board games and quite a few more things.

As in previous years, I also wrote a puzzle for the Palantir Puzzlehunt. This year’s puzzle Salt Fat Acid Heat was unfortunately a fair bit easier than originally intended: it involved three connected 6×6 Thermo Sudokus. My goals here were twofold: (1) have a puzzle without numerical clues, and (2) have connected puzzles that were only uniquely solvable together. I thought I reached a nice design which had a deduction involving all three grids at the same time, though this proved too hard in playtesting and was thus scaled back. Feedback on the puzzle was expectedly divided, with praise for elegance and complaints for being too classical, and remarks on the difficulty (even though it was already simplified).


In diesem Jahr war die Entscheidung, Deutsch weiter zu lernen, schwieriger als im letztes, weil ich selbst bezahlen muss. Wegen des Coronavirus hört Palantir meinen Sprachkurs auf. Ich setzte meine Deutschstunden mit Katja fort, denn ich finde, dass sie gut und schnell lehren kann. Ich weiß nicht genau, wie gut mein Deutschkentnisse jetzt sind. Ich bestand die A1-Prüfung mit 100 Prozent, und beim Deutschstunde beendeten wir das A2-Kursbuch. Ich las Bücher und machte Ubungen auf dem B1 (z.B. Deutsch für Besserwisser B1, Hueber Lesen und Schreiben B1), und ich probierte (und bestand, aber nur mit 70 Prozent!) die B2 Leseprüfung. Ich bin sicher, dass ich B2 nicht erreichte. Auch nicht zum Lesen: ich kann nicht auf Deutsch “zeitgenössische literarische Prosatexte” verstehen. Ich bin wahrscheinlich B1 für Lesen, Hören und vielleicht Schreiben, aber nur A2 für Sprechen.

Ich bin viel besser in der Grammatik als noch vor einem Jahr. Damals lernte ich gerade wie man “weil”, “deshalb” oder “ob” benutzen – jetzt verwende ich diese Konstrukte leicht. Es gibt Dingen, z.B. Passiv und Adjektivendungen, das ich noch ein bisschen komisch finde. Aber wenn ich einfache Dinge sagen will, ist es kein Problem.

This year, the decision to continue learning German was more difficult than last year. This was because I had to pay out of my own pocket. Palantir stopped the language courses because of the coronavirus. I continued my lessons with Katja because I find she can teach well and quickly. I don’t know exactly where my German skills stand. I passed the A1 exam with 100 percent, and in the German lessons we’ve finished the A2 book. I’ve read books and done practice exercises at the B1 level (e.g. in Deutsch für Besserwisser B1, Hueber Lesen und Schreiben B1), and I even tried and passed the B2 reading exam (though with only 70 percent). I’m sure that I haven’t reached B2 though, not even for reading. I cannot read and understand “modern literary prose” in German. I’m probably B1 in reading, listening and maybe writing, but only A2 in speaking.

I am much better in grammar as compared to one year ago. At that time, I had just learned how to use “because”, “therefore” and “whether”; now I use these constructs easily. There are things like the passive voice and adjective endings that I still find a little unusual. However, when I want to say something simple, it is not a problem.

Board Games

One of the activities that I’ve been doing somewhat more this year is playing board games with friends. Expectedly, most of this has been remote on Tabletop Simulator. There have been two games that have been particularly successful in the gaming sessions I’ve had with my friends:

  • Spirit Island, a cooperative game about repelling invaders from a natural island (or, depending on how one sees it, optimally crafting and scheduling database transactions), and
  • The Crew: The Quest for Planet Nine, a cooperative trick-taking game where the goal is usually to ensure that specific nominated players win specific cards.

I’d played some Spirit Island before this year, but I had more sessions this year and have become a lot more comfortable with the game mechanics. I used to stick primarily with a few spirits that I knew how to play quite well (River Surges in Sunlight, Lightning’s Swift Strike, A Spread of Rampant Green) but have had a lot of fun with some of the newer spirits with the various expansions. Stone’s Unyielding Defiance feels very powerful, even if the gameplay is a bit onenote. Playing Shifting Memory of Ages is exciting, if at times very dependent on major power card draws, and to me feels like a more fun version of Vital Strength of the Earth. Downpour Drenches the World is enjoyable in its own right, essentially denying invaders progress and then repeating a power five times. I think these three are probably my latest set of “fun” spirits. Some of my teammates used to play the game a lot in the office; more recently, we’ve had a few sessions remotely. The game has a lot of moving parts, but is essentially a big and, to me at least, engaging optimisation puzzle (that admits pretty suboptimal solutions, fortunately).

The Crew on the other hand has been primarily played with friends from Imperial. It has elements of Hanabi in that communication is limited, and thinking about player intentions is critical. James and I have played through and beaten a number of the hard missions, though the 2P variant with Jarvis hasn’t actually been that difficult, I think – we both play quite logically. I think some of my experience with playing out a bridge hand (not so familiar with the bidding, more so with the play) is relevant.

Travel and Exploration

What is travel?

More seriously, there was one weekend trip to Switzerland this year and that was about it, though I did spend more time exploring my parts of my local area that I did not know as well in the past (such as St James’s Park and some walking routes in London). Eat Out to Help Out and the various Amex Shop Small promotions did also make me look around for new deals and offers.

I think if the situation abates in 2021 (like it did in the summer), even if in an inconclusive way, I will be travelling quite a lot (taking the relevant precautions, of course, and the vaccine as it becomes available). I can’t remember if I even travelled overseas at all in 2011, so this might strangely be more than that still.


I started getting back into DDR late this year, though I’ve had to play at home for the most part because of the coronavirus. Some of the songs also make good background music when I’m working on something, especially tasks that are “just work” as opposed to requiring a lot of deep focus. Hopeful Frontier is probably a good example; the track runs at a crisp 200 BPM and has fast 16th and even 24th note streams. I can’t step that quickly in DDR terms (that’s 13.33 and 20 steps per second, respectively) – and in any case it’s not actually a DDR song, instead originating from jubeat, one of the other games in Konami’s rhythm game series. Paranoia Survivor is slower but has a bit more personal significance; it’s the first level 15 song I passed on DDR probably about 10 years ago. Coming back to DDR after not having played for so long, it’s hard but not ridiculous (unlike the level 17s which I don’t have the stamina or footspeed for now), and being able to almost automatically pivot my body to do the crossovers has been nice.

Some of this is probably coronavirus-influenced but this cover of I Will Remember You is probably the most-listened song outside of the aforementioned rhythm game music.


Distinct Positive Outlooks

2020 was a tough year for many developed and developing economies. It does feel kind of perverse that this year was financially successful for me, amid broad destruction both in terms of economies as well as many individuals’ finances. I’ve been fortunate to remain at Palantir, and I do produce good enough work so that the people there are willing to keep me around. Palantir had a direct public offering (DPO) at the end of September. The stock opened with a volume-weighted average price of $10.12 in the first trading session, and then rose further, eventually going above $30, which was well beyond what I expected. The tax rates on the earnings here are pretty high (having to pay the employer NI contributions puts the marginal tax at around 54%) but I don’t find them unreasonable given equity-based compensation is a privilege.

Budgetary Changes

This was the year where I spent more on groceries in April than I did for the whole of 2019. Obviously, working from home influenced expenditure patterns; there were large increases in Food, Coffee and Groceries. However, especially for groceries working from home alone does not explain the magnitude of the increase. Grocery expenditure in 2020 was 9.8 times that in 2019. Previously I would usually have 2 meals at home each week, while now it is probably more like 10. Thus, the “equivalent factor” should be no higher than 5, and actually given that the office was still open in January and February, 4.33 is probably the correct multiplier for the year. The implication then is that I am spending more per meal, let’s say p times. Then, we want to solve \frac{1}{6} + \frac{5}{6}(5p) = 9.8 \leadsto p \approx 2.31. That probably checks out; I have gone to M&S quite a bit more than I used to. I’ve also spent a fair bit more on fresh food and the occasional specialist Japanese groceries as a treat. This was accompanied by massive drops in transport and travel (as expected), and noticeable increases in Learning (paying for German lessons, buying more books).

Investments and Philosophy

COVID-19 has made me re-think my approach towards money. A more formal way of saying it would be that I’ve realised that having money is not a sufficient condition for being able to redeem it for something useful or enjoyable; a more causal way would be the “you only live once” (YOLO) phrase that Drake invoked. Having non-essential retail and almost all restaurants close during the first lockdown (even for take-away: while it was legal to remain open for take-away, they had logistical issues to sort out first) made the limitations of having money or some other store of value apparent. There was still online shopping for goods (and, as it turns out, services) that one could have done, but it was relatively constrained. There is a point where optimising for today becomes irresponsible, but having an excessively long-term mindset can also be damaging.

We’ve had many studies on the subject of delayed gratification and how the child that was able to wait for his marshmallows tends to perform better in life later on, but never eating the marshmallows doesn’t seem to be the solution. I’ve discussed this in a blogpost before, but if we consider the question of “would you take N dollars today or 2N dollars in a year’s time”, I’d used to always have opted for the future option, but nowadays depending on N I might opt to take it today (generally if N is very large). For some reason, I have a natural instinct to save aggressively. This period has taught me that there are limits to this, and perhaps along the lines of what the teacher says in Ecclesiastes, the fruits of one’s labour are to be enjoyed.

Interestingly, in 2019 I claimed to have lost some interest in my portfolio; that’s even more true in 2020. I’ve taken quite a bit less interest in my investments this year, perhaps because of this lesson. I continue to do the monthly motions, but otherwise haven’t looked much at my portfolio. My understanding is that performance for this year is broadly flat for a GBP-based broad-based global equity tracker; I do remember pushing a bit more money into the market in March when the market fell, and did the Bed and ISA, but otherwise don’t really remember much.

Leave a Reply