2020-21 New Year review

This is an annual post reviewing the last year and making resolutions and predictions for next year. 2020 brought a combination of challenges from living in a pandemic and becoming a parent. Other highlights include not getting sick, getting a broader perspective on my life through decluttering, and going back to Ukraine for the first time. (This post was written in bits and pieces over the past two months.)

2020 review

Life updates:

Janos and I had a son, Daniel, on Nov 11. He arrived almost 3 weeks later than expected (apparently he was waiting to be born on my late grandfather’s birthday), and has been a great source of cuddles, sound effects and fragmented sleep ever since.

1 week old
6 weeks old

Some work things also went well this year – I had a paper accepted at NeurIPS, and was promoted to senior research scientist. Also, I did not get covid, and survived half a year of working from home (much credit goes to the great company of my housemates). Overall, a lot of things to be grateful for.

AI safety research:


Working from home was definitely a productivity hit. I was mostly focused on urgent tasks, such as conference submissions and reviewing, and didn’t get much research done. I did 366 hours of “deep work” (1.95 hours per work day) this year, compared to 551 hours (2.4 hours per work day) in 2019. This includes theory work, reading papers, writing papers and code, but not editing text or debugging. I got back into using work cycles, which added some helpful structure in the home environment.

I was very grateful to be living in a group house during the pandemic. While it was a bit tricky to have 5 people sharing the space when working from home, it was awesome to have an in-person community and not feel completely isolated from the world. It was also much easier to do nonzero exercise when I had someone to do it with, e.g. running in the park together.

Spending a lot of time at home inspired me to do a lot of decluttering. In particular, I went through all my old notes, got rid of most of them, and gathered the ones that still seem interesting and relevant (notes from rationality workshops, Hamming worksheets, reflections and so on). I put these into a binder for easy reading, and found it useful for getting a big picture sense of how my attitudes and problems have evolved over time. This has been particularly helpful during the pandemic, when my life has often felt small and repetitive.

I got a UK driving license for automatic cars, which took a surprising amount of practice given that I already had a US license. There was a lot to get used to with the left side of the road, the narrowness of the streets and frequent maneuvering – I spent a number of lessons just on getting the positioning right. I did the theory test in March and planned to take the road test in early summer, but then the driving schools closed for lockdown, and I ended up starting the lessons in July. I took the road test in September and didn’t pass because of “undue hesitation” at a busy roundabout, so I had to repeat the test in October, two weeks before Daniel was due (thankfully, I passed this time and could forget about driving for a while).


Physical health has been pretty good this year. Last year I had 7 colds, while this year I was not sick at all – probably due to social distancing and taking zinc regularly. Thankfully, recovery after the birth was relatively quick, feeling mostly normal in around 2 weeks, though it will take some time to get my core muscles back online. I’ve been getting some back pain from lifting Daniel (now at 6kg thanks to his voracious appetite), which makes it all the more important to rebuild core strength.

The second half of the year came with pretty bad sleep – a lot of insomnia in the last trimester where I woke up at 3-4am for no discernible reason and couldn’t fall asleep again, followed by fragmented sleep after Daniel was born. Living and working on 4-5 hours of sleep before I went on parental leave was surprisingly ok, probably because I was waking up on my own rather than being woken up in the middle of a sleep cycle. On the other hand, being woken up by a hungry baby definitely feels more meaningful than waking up at 3am for no reason and not being able to go back to sleep.

Rate of insomnia by month
Average hours awake at night by month
Average hours of sleep by month

This year has been pretty hard on my mental health due to a number of ways that the pandemic interacted with having a kid, and various problems that I considered solved have made a comeback lately. I spent most of the year at home without many forms of self-care, such as my usual exercise, sufficient sleep, or nice things like going to the sauna. While the birth went well, there were a lot more stressful interactions with the healthcare system than I had hoped. After that, there has been a combination of sleep deprivation, limited daylight, a mostly empty house after some housemates moved to the countryside, difficulty with meeting friends outside because of cold weather, making increasingly modest plans only to have them shot down by the ever-changing lockdown rules, and the exciting new covid strain we have in London that calls for high levels of caution and isolation. Thankfully, my mom was able to come stay with us for a few months to help with the baby, make food and keep us company.


In January I visited Ukraine for the first time since I emigrated 17 years ago. I saw my dad and aunt, as well as my niece and her kids who live in a remote part of Canada but happened to be in Ukraine for the winter. It was an interesting experience to navigate around Kyiv – I no longer had a map of the city in my head, so I recognized some familiar places but could not recall where they are relative to each other, so this felt like visiting a new city with a lot of deja vu. I was pleasantly surprised by the large number of Georgian restaurants in Kyiv, which we made sure to frequent and were not disappointed.

In March we did a week-long meditation retreat at MAPLE in the US. We hesitated whether to go ahead with this plan given that flights might get canceled, but ultimately decided to go. The retreat was in a remote location in Vermont that seemed pretty safe from a covid perspective. I was advised to follow an equanimity practice that worked pretty well (focusing on acceptance rather than observation of things that come into my awareness).

Janos in his element

We had a peaceful week meditating among the snows, which unfortunately became less peaceful at the end when some people started coughing, so we spent the last couple of days meditating in masks and gloves, and left as soon as the retreat ended. Our flight back did get canceled, but we were rebooked on another one for free. Upon returning home, we self-isolated from our housemates and acquired covid tests, which were thankfully negative. We later learned that several people at the retreat tested positive for covid, so this was a close call.

In August we went camping in North Wales at a friend’s cottage (after the first lockdown was lifted). The cottage itself was abandoned (and a bit spooky), but we could stay on the adjacent land and thus avoid crowds at the newly opened campsites. We enjoyed a lot of swimming in cold waterfalls and a much warmer Atlantic ocean. We also hiked up a nearby mountain Arenig Fawr (Daniel and I were taking it slow).

View from the summit of Arenig Fawr

In September we had a vacation in Madeira, where everyone was tested for covid on arrival at the airport, and there was no community transmission at the time. Madeira is a volcanic island that is basically one big mountain, and we had an interesting time driving around it (on the way to a hike, our car refused to go up a very steep road and we took a taxi the rest of the way). The terrain was a great combination of ocean and mountains.

We enjoyed large quantities of Portuguese food (as we soon learned, it did not come in small quantities). Our special favorites were local rock mussels called limpets – we ate enough of them to have a shell stacking competition.

The winner built a tower of 22, which collapsed before we could take a photo, so here is a tower of 13.

In December we did two short hikes near London, and verified that the basic algorithm of putting Daniel in a car seat and then in a carrier for the hike seems to work pretty well (he mostly sleeps through all this). I’m glad to be able to visit some nature during these strange times, it makes the world feel just a bit less small.

Iron age hill fort in Epping Forest

2020 prediction outcomes

Unsurprisingly, some predictions for the past year were messed up by the pandemic.


  • Author or coauthor three or more academic papers (3 last year) (70%) – yes (3 papers)
  • At most 12 non-research work commitments, such as speaking and organizing (10 last year) (80%) – yes (5 commitments). Easy, since a lot of events got canceled.
  • Meditate on at least 270 days (290 last year) (80%) – no (244 days). The past month I only managed to meditate on 9 days, and this has not been good for me, so I need to do better next year.
  • Read at least 7 books (5 last year) (70%) – yes (9 books). Human Compatible, The Precipice, Secret of our Success, Ender’s Game, In the Realm of Hungry Ghosts, Watching the English, Positive Birth Book, The Gardener and the Carpenter, Raising a Secure Child.
  • At least 700 deep work hours (551 last year) (70%) – no (366 hours). I found it much harder to do deep work at home, and was on parental leave for the last 2.5 months of the year.


  • I will write at least 5 blog posts (60%) – no (3 posts)
  • Eating window at most 11 hours on at least 240 days (228 last year) (70%) – no (131 days), since I stopped doing intermittent fasting this year
  • I will visit at least 4 new cities with population over 100,000 (11 last year) (70%) – no (2 cities, Birmingham and Funchal). Much less travel than normal this year.
  • At most 1 housemate turnover at Deep End (70%) – no (2 housemates). One housemate moved to live with parents in the countryside who would have likely stayed in London under normal circumstances.
  • I finish a language in Duolingo (60%) – no, though made some progress on Mandarin (completed checkpoint 2)

2021 resolutions and predictions


  • Avoid catching covid (90%)
  • Author or coauthor three or more academic papers (3 last year) (70%)
  • At most 7 non-research work commitments (5 last year) (80%)
  • Meditate on at least 230 days (244 last year) (70%)
  • At least 450 deep work hours (366 last year) (70%)
  • Do 4 consecutive chinups (70%)


  • I will write at least 3 blog posts (3 last year) (60%)
  • Janos and I will get vaccinated for covid by the end of June (60%)
  • Daniel will get to meet all of his grandparents in person in 2021 (70%)
  • I will return to avoiding processed sugar by the end of the year (60%)
  • I will finish Hungarian Duolingo (complete checkpoint 5) (70%)

Past new year reviews: 2019-202018-192017-182016-172015-162014-15.

5 thoughts on “2020-21 New Year review

  1. Tommy Lees

    Could I ask what the probabilities assigned to each resolution refer to? Are they a statement of how likely you are to succeed? Or perhaps some measure of importance?


  2. Pingback: 2021-22 New Year review | Victoria Krakovna

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s