Doctor Zhivago is a relentlessly long read, but is ultimately about the inner life of an Idealist (in the MBTI NF sense) living through the harsh political turmoil of the Russian Civil War.

Quotes I enjoyed

My favourite scene is while Lara and Yuri are hiding in Yuriatin from the madness of the civil war and Yuri’s responsibilities, and just talking with each other. The couple have such a deep understanding and sympathy for each other’s feelings and thoughts, and up until now Zhivago’s rambling inner monologue has not felt received by another character’s with the affordance for him to fully explore it. The dialogue below summarises that:

[Yuri:] Have you ever heard the name of Nikolai Vedenyapin?”

[Lara: ] Well, of course. Before I met you, and later, from what you’ve often told me. Simochka Tuntseva mentions him often. She’s his follower. But I’m ashamed to say I haven’t read his books. I don’t like works devoted entirely to philosophy. I think philosophy should be used sparingly as a seasoning for art and life. To be occupied with it alone is the same as eating horseradish by itself. But forgive me, I’ve distracted you with my stupid talk.

[Yuri:] No, on the contrary. I agree with you. That way of thinking is very close to me.

And later when Lara reflects on their love:

Never, never, even in moments of the most gratuitous, self-forgetful happiness, did that most lofty and thrilling thing abandon them: delight in the general mold of the world, the feeling of their relation to the whole picture, the sense of belonging to the beauty of the whole spectacle, to the whole universe. They breathed only by that oneness.

I think Pasternak illustrated the joy of feeling like your inner thoughts, regardless of their messiness, might be understood by someone. That conversation may branch from a family member to the importance of philosophy.

I had a lot of self-identification in Zhivago’s internal life. He is a man constantly preoccupied with the profoundness of things and what it means for others, what is going on in their life. He spends such long periods reflecting and understanding, and - to his failure - such little acting upon the world.

The uncle’s influence furthered Yura and liberated him, but it fettered Misha. Yura understood what role in Misha’s extreme enthusiasms was played by his origins. From cautious tactfulness, he did not try to talk Misha out of his strange projects. But he often wished to see an empirical Misha, much closer to life.

And Lara in return:

Then untruth came to the Russian land. The main trouble, the root of the future evil, was loss of faith in the value of one’s own opinion. People imagined that the time when they followed the urgings of their moral sense was gone, that now they had to sing to the general tune and live by foreign notions imposed on everyone. The dominion of the ready-made phrase began to grow—first monarchistic, then revolutionary.

I assume Pasternak is kind of like this. Other characters have random asides of esoteric (often Christian) ideas, which show that character as an intellectual. But in doing so they too demonstrate the author’s intelligence, faith, and perhaps a leaning towards the psychological:

Nikolai Nikolaevich was living in Lausanne. In the books he published there in Russian and in translation, he developed his long-standing notion of history as a second universe, erected by mankind in response to the phenomenon of death with the aid of the phenomena of time and memory. The soul of these books was a new understanding of Christianity, their direct consequence a new understanding of art.

And Sima Tuntseva, differentiating “passion” as lust from “passion” for forgiveness and intimacy with Christ:

‘Passion’ in Slavonic, as you know perfectly well, first of all means ‘suffering,’ the Passion of our Lord, ‘the Lord goeth to His voluntary passion’ (that is, to His voluntary suffering). Besides that, the word is used in the later Russian meaning of vices and lusts. ‘Having enslaved the dignity of my soul to passions, I turned into a beast,’ ‘Having been expelled from paradise, let us strive to enter it by abstention from passions,’ and so on. I’m probably very depraved, but I don’t like the pre-Easter readings in that line, devoted to harnessing sensuality and mortifying the flesh. It always seems to me that these crude, flat prayers, lacking in the poetry proper to other spiritual texts, were composed by greasy, fat-bellied monks. And the point is not that they themselves did not live by their own rules and deceived others. Suppose they even lived according to conscience. The point isn’t them, but the content of these texts. These laments give unnecessary significance to various infirmities of the body and to whether it is well-fed or famished. It’s disgusting. Here a dirty, inessential secondariness is raised to an undue, inappropriate height.

“It has always interested me why the mention of Mary Magdalene is placed just before Easter, on the threshold of Christ’s end and His resurrection…Now listen with what genuine passion, with what directness regardless of anything, this mention is made.

The physical world takes up little of Zhivago’s internal monologue and the author’s prose. His escape from Liberius’s camp is an almost non-event, the principle events happening are his thinking about his family and Lara, a brief conversation with a sentry, and a one paragraph montage of his “very long” journey which illustrates more the state of the people than anything Zhivago himself does.

On Style

I enjoyed the rambling conversations of all the characters. There is a lot of dialogue, and it can often be hard to track who is saying what. This creates an impression of ideas being shared and mingled between characters, with the back-and-forth being more central than any one character changing their mind.

I understand it is a common assertion that the book is very long, and there’s an unreasonable number of coincidences and chance run-ins. So I simply echo these.

On Zhivago as an INFJ

As an indulgence to my MBTI brain, I think it is pretty clear that Yuri is an INFJ:

1. Ni - he spends a lot of time thinking about the world, drawing out patterns, and being struck by the enormity or significance of things.
2. Fe - in particular he’s interested in the emotional worlds of people, the ideas and thoughts which are shaping their feelings and actions. He also seems to be cautious about upsetting other people, and it feels like his first and last marriage both kind of “happen” to him based on the expectations of others. And until he meets Lara, he expresses little pain with that.
3. Ti - he is interested in ideas and frameworks, but more as things to play with or explore. His preferred writings is poetry - impressionistic and vibesy - over essays - ideas-based and logical. He has a brief discussion with Pogorevshikh on the train, but not nearly enough to try to argue his case.
4. Se - very rarely does he act decisively, and it seems that actions almost occur to him.

New writing will show up here from time to time.

AI Safety Fellowship

I begun the AI safety fellowship – I met a few cool people, and reviewed basically all the jobs I could possibly work on. The main skill it appears I’m missing is a) research experience, b) proven credentials (like published research, a PhD, etc.). This is similar to what current and interning researchers I have asked have said – and especially the main thing to focus on is research experience.

Now the question of where to get that research experience? I’m not sure exactly. The three obvious routes suggested are: 1) Keep applying to the fairly frequent EA-related programs; 2) Volunteer to work for free for a spell; 3) Go back to uni.

Job Plans

So far I don’t have many good job plans. My ideas of possible routes are:

1. Finity, but focusing on more interesting projects. I need to talk to people doing those interesting projects to know whether I’d like them.
2. AI research, especially on the control problem. I need to gain some experience doing research to figure out whether I enjoy the work or not. There may be other steps before that.
3. EA operations work, for example being in charge of operations at GWWC or a charity. I need to talk to people working in this area to see how useful they think I’d be. I also need to decide whether this is the most worthwhile use of my abilities/energy.
4. Quantitative research at a firm like Akuna. I need to start applications for these kinds of roles.

There’s a lot to do and not a lot of time for me to do these in. Does anyone have any thoughts about prioritising?

Minus One D&D Game

As you may know, I was previously playing in three D&D campaigns concurrently, each taking ~four hours every week. Which was too much! So I’ve sadly had to leave one group. I’m still playing in two (although one is probably three sessions per month, so a little bit lighter).

What I’ve Been Reading/Watching/Hearing:

• It Takes Two (video game) – Bec bought me a video game for Christmas called It Takes Two. The premise is that you and your spouse have been shrunk down, and the only way to return to your previous size is to learn to work together. But it has fun mechanics, where each level is more of a puzzle to figure out how your two characters – with slightly different abilities – can make progress. So it’s quite good for inexperienced gamers like us who aren’t very coordinated.
• Encanto (movie on Disney+) – This is another great movie musical by Disney. The music is fun, and by Lin-Manuel Miranda. The story is great, it points to themes of family and feeling pressure to meet expectations. And I find lots of songs make me tear up, but I think that’s just me becoming more emotional!
• Dune (movie) – I saw this in IMAX, which I’d recommend to all who can. The scenes are great and epic and feel much better on the big screen. I could imagine it dragging on a laptop. I enjoyed being able to spend time with all the characters as it gave extra depth to the world and to Paul. I’m very excited for the next film.
• Matrix (movie) – I also saw this at IMAX. It was… alright. The beginning part was good and fun. But the last third it kind of became worse. It became much less gritty and real, and some of the character decisions and responses felt a bit untrue-to-character. I think it’s exacerbated by the fact that Carrie-Ann Moss and Keanu Reaves and their characters don’t strike me as particularly emotional actors, so it’s hard to evoke the right emotions in the audience.

I spent Christmas Day with Bec’s family, and then went down to Melbourne for a Boxing Day dinner. We had a great meal with a two tasty salads, a turkey, ham, crab, oysters and a range of other foods.

Boxing Day sales

Did anyone get any good deals from Boxing Day sales? I bought a $1,000 mattress topper called an Ooler. It pumps cold water through the mattress at night to keep you cold. I think it’s pretty good, but I’m not sure how much it’s really improving my sleep.

AI applications

If you recall last month I had been rejected from one research program, and had applied to another research program. I was also rejected from the other research program! However I have been accepted into an AI “fellowship” where I’ll basically be working on my own thing, alongside other people looking to get into the area of AI safety. I’ll be working on applying to more things. (Maybe I can apply to a fellowship which will help me apply to fellowships where I can apply for things!).

I was also accepted into a AI Safety Fundamental program, where we’ll be studying about AI and how to use it. Hopefully that is useful!

And if there’s any jobs – AI related, data science related, or just generally maths-y – which you think I might enjoy, please feel free to suggest it to me!

New years resolutions!

Have you made any new years resolutions? I have a few: I want to learn Chinese (I’m having ~weekly conversations with a native speaker; flash cards daily), I want to lose weight (I’m trying to hit 10,000 steps daily; tracking my food; and trying to do pull ups), and scope out new job opportunities (see above!).

What I’ve Been Reading/Watching/Hearing:

• King’s Avatar (TV show, Netflix) – as I’ve been trying to get into Chinese, I’ve been watching TV shows in Chinese. This one is about a professional gamer who gets kicked out of his team because they want him to become a celebrity and he wants to remain anonymous. So he goes across the road to an internet cafe and forms his own team.
• ChinesePod (podcast, Spotify) – 15-20 minute episodes with a short Chinese dialogue, and then some explanation and translation in a mixture of English and Chinese. At different levels (beginner, pre-intermediate, intermediate, upper-intermediate, etc.). I find I’m around pre-intermediate to intermediate.
• Chinese Cooking Demystified (Youtube channel) – I never learned to cook Chinese food, and now I can, via Youtube! This is a husband-wife duo, their episodes are suitably short and to the point, and they include recipes in the descriptions.
• What Colour is your Parachute (book) – This is a famous book about how to find a good job for you. I remember I read it back in 2015 when I was desperately trying to find a job. I had positive memories of it, but upon re-reading it I actually don’t think it’s very useful. It focuses too much on finding a job which is a perfect fit for your unique traits. Especially when you’re quite junior you don’t have that many valuable unique traits! Some of the advice in the first few chapters about how to job search and apply are helpful though.

I’ve started learning how to code websites in React, which is a software package in Javascript which helps you build websites. So far I’ve created two websites: a) one to basically create a browser view of your accounts from the bank Up, b) the beginning of a to do list, heavily inspired by Asana.

Black Friday sales

Did anyone get any good deals from Black Friday? We got a board game called Wingspan, software which helps you play TTRPG’s online, and an air fryer.

AI applications

I’m currently looking to get more involved in some AI stuff, so if there’s any sources which would be interesting please let me know. I have a website which might be interesting for people to see the kinds of things I’ve done, although I don’t maintain it very well.

I’ve recently applied for two AI programs: One would have been a paid trip to SF for a few weeks (I was rejected). The other is an online program that runs until June in which I would be working on some AI safety research (still waiting to hear back!).

Gym-ing again (not quite)

So I haven’t quite made it to going back to the gym yet. Perhaps December will be the month…

What I’ve Been Reading/Watching/Hearing:

• Deep Questions with Cal Newport (podcast) – Cal Newport has written a number of very insightful books about how (and why) to find a good career, develop the ability to focus, and structure your work to not feel overwhelmed. In his podcast he answers audience questions, which I think makes his advice from his books increasingly actionable and easy to understand.
• Therapy from tick, tick… BOOM! (song) – Andrew Garfield can sing! Who knew! I love patter songs, and I love songs with back and forth and I also love songs which sound happy but are actually emotionally painful. The movie (which I have yet to see) was made by Lin-Manuel Miranda (of Hamilton fame) and the songs were written by Jonathon Larson (of Rent and dying when he was in his young thirties fame)

Sydney unlocked!

The COVID lockdowns have finished, hooray! I’ll be travelling to Melbourne soon to see family, after having been unable to for a whole year.

New housemate

We weren’t able to find a housemate during lockdowns, because fewer people were looking. But now with COVID over we have found someone. Good to have that off the list of things to do!

D&D and stuff

I’m currently playing in three Tabletop Roleplaying Games (TTRPG’s): Two games of D&D and one game of Pathfinder 2e. They’re all quite fun, but they take up a lot of time. One game is a multi-year campaign with experienced players and an experienced DM. One game is with friends and allows me to play occasionally. And one game is playing Pathfinder 2e, a much more tactical game which I’ve been wanting to try out. So it’s hard to decide to drop any!

Gym-ing again

Now the gym is open again, I’m going to try to start going more often. I’ve so far been mainly doing bodyweight exercises, but I think the weights are a bit more intensive way of training.

What I’ve Been Reading/Watching/Hearing:

• Ravnica Cycle #1 and #2 (book) – a fun series set in the fantasy world of Ravnica.
• Critical Role Campaign 2 (long Youtube series) – Critical Role is a game of voice actors playing Dungeons and Dragons. Because they’re all actors, it makes for a really interesting improvisational game with deep characters. But it’s also very long! They’re 141 episodes and each go for around 4 hours. I am just about to finish it, after having watched it on and off for the past two years.

In short, the summary was:

• I had already talked too much that day
• I was tired
• I hadn’t planned what I wanted to talk about at the party

I am beginning to suspect that my socialisation energy runs out at about 2.5 hours. That is, I can be social with people for about 2.5 hours before the energy required to maintain my attention on the conversation runs out, and I start disconnecting. There are other factors which hasten or delay this point, which are:

• Talking to people I don’t know (hastens)
• Talking about something I haven’t talked about before (hastens)
• The feeling that I must be quick to respond (hastens)
• Being in a large group with a single conversation, where my attention can drift off (delays, although really this is just taking breaks in the 2.5 hours extraversion lifespan)
• Factors which make it hard to hold attention e.g. noise, being tired, multiple conversations at once (hastens)

I haven’t fully established the most efficient way to recharge. I tried a 5-minute meditation at the party, which made me feel a lot less tired and inattentive, but that also eventually wore out.

Another life hack I have for parties is to plan in advance conversational topics. I imagine this is something that people don’t need to do, but I feel like I have a problem I shall call Pointer Dereferencing Latency, which I’ll take a brief interlude to explain.

Pointer Dereferencing Latency 

I don’t know if this is a common experience, but often I feel like it takes a lot of work to explain “what I mean” to people. A few examples will help.

I’m at work, and I discover these is a problem with one of the models. Sometimes I’ll immediately stand up and go to another analyst and tell them “there is a problem with the calculation”. When asked to explain, I sometimes stumble upon the words to explain, I explain the wrong things in the wrong order, or I miss out on a key explanand. I would say I know what the problem is, it is “the calculation” (a pointer), but it takes time to go from the pointer to the full detail (how to dereference).

Another instance: I am explaining a concept, and I use a word which someone doesn’t understand (let’s say I use “Bayesian”). I understand what it means to be Bayesian. I can correctly point to examples of Bayesian models, and classify an approach as being Bayesian or not. But when asked to explain, it comes out as a series of “um, erm, you have a prior and then you get data and then you update your posterior, rather than being frequentist and…”. The problem isn’t that I don’t understand Bayesianism, it’s that it takes time to be able to explain it. And when I try to explain, I keeping using other pointers (prior, data, update, etc.), as I’m already goddamn trying to dereference one pointer, I can’t dereference all those other points too.

In general, Pointer Dereferencing Latency is when one understands the problem, but there is a huge gap of knowledge between you and the other person, and you are not good at unpacking what you mean without relying on other concepts they don’t understand, and presenting it in an easy to understand order.

This is correlated but not the same as just not understanding. I think a lot of people rely on pointers, and don’t dereference every time they think. Sometimes I’ll be explaining something, and I’ll realise that I don’t actually understand every step, or that the arguments aren’t as strong as I thought. And it’s reasons like this that teaching something can help to consolidate your own understanding.

But Pointer Dereferencing Latency is when I really do understand my argument, but it just takes me time to explain it.

And this is what I think happens when I am at parties, and I don’t have anything to talk about. Because I have a pointer called “interesting things I have been thinking about”, and it points towards a list of really interesting conversations. But in the moment of the conversation when someone asks me what I’ve been up to, it takes too much time to dereference that list into an interesting conversation topic.

Finally, the real reason I wanted to write this post was so that I could write a list of interesting conversational topics which I have been thinking about.

So here goes:

1. What do you think of The Case Against Education?
2. How would you create the internet if you were living in the 50s? Is there anything today which requires heavy investment from government, as it is simply too ambitious for the private sector to do it?
3. How do you relax?
4. Should I have more hobbies, or activities which I use to switch out of a productivity mindset?
5. What would happen if the world were made of mini-states? How much would freedom and diversity win out of over economies of scale? Are social media websites any indication that things would stay the same?
6. Why have we not solved face-to-face interaction? It is still worse than anything virtual.
7. Have Millenials been trained to see anything which isn’t productive as immoral?

My response is three-fold:

• Many of these jobs do add value
• Many of these jobs don’t add value when there isn’t enough economic pressure
• But still I support parts of his sentiments

Other reasons why such jobs might exist, which suggest they do add value:

• Governance: Filling forms out, providing reports, and other administrative tasks allow managers of companies to know what is happening in a company and to have checks and balances that things are happening properly. Bad governance means the managers have a incorrect understanding of their business. Self-service check-outs have resulted in more thefts, so although check out staff are an inefficient use of that labour (though I don’t think Graeber would call it a BS job), it is easier to fool a self-service check-out than a staff member, and hence is a more effective check (no pun intended). To paraphrase the meme: Do you want Royal Commissions? Cos that’s how you get Royal Commissions. Or more positively: We can’t have large international cooperation if we don’t have good systems in place.
• Computers need data: This is actually from Graeber: computers need data so that they can summarise and predict things, and a lot of data is qualitative. So people need to be there to turn them into computer-consumable things, even if those things are mundane. This does not sound like fun for the employee, but it is at least there is a computer which can add up the numbers for you at the end of it.
• Rules are easy, judgement is hard: Burja says bureaucracies are created because the creator doesn’t trust the bureaucrats are sufficiently competent or aligned to act in line with the creator’s wishes. But you need a large company to do things at scale. Enter the bureaucracy.
• Risk mitigation: Graeber talks about one man whose entire job was to fix the air conditioner if it broke. This may be an extreme case, but some people’s jobs are effectively there on retainer because the company desperately needs someone to respond urgently to matters.
• Deep Work Enablers: Cal Newport’s argument: You want to have support staff who do mundane things so that the experts can do Deep Work
• Customers need punching bags: Putanumonit‘s argument that people “below the API” are still there because some people (i) don’t know how to use technology to change flights and (ii) just want someone to deal with it

Around 34% of people believe that their jobs do not serve a purpose to the world. But I’m not sure if people are thinking of all these possible uses. I’m also not sure if people consider arguments of the form: “this product gives people value, and me doing X is part of the process of producing this product,” when evaluating their job.  10% think that their jobs do not serve a purpose for the company, suggesting that around 25% of people think their jobs contribute to the company but not to the world. Either their company is evil (which probably is a problem) or they fail to make the above Part of the Process argument (or they’re a government organisation which won’t collapse, or probably a number of other explanations).

The additional on-face-value-not-good-but-actually-probably-good-for-the-world reason that these “BS” jobs exist is this: A company cannot outright deny people Y, but they want to reduce people frivolously receiving Y. SSC suggests this is why he must fill out a nonsense form saying people can have special chairs at their office. To the BS employee, this looks like their job is to stop people getting Y. But to the company, it stops fraud or unnecessary claims or just ensures people go through the standardised, more efficient process, which means that the company can continue to offer its service at the price it does. (This is unfortunately not very fair to those who are time-constrained, though.)

Not enough economic pressure

I feel like these jobs are more common in government:

Governments have less financial pressures, and so people begin to care about more status, and so want to raise their own headcount. It can also be hard to fire people in government because of HR protocols, from what I hear. Without economic pressure, BS jobs are more likely to continue.

Supporting Graeber

But reading Graeber is still inspirational despite my disagreements. Specifically it makes me excited about:

• A world where people spend more time doing care jobs, and helping people directly, because most technical jobs are replace by robots
• A world where there is a lot innovation around helping people. Large-scale “investment in people”, in terms of labour-hours. Personalised learning customised to children which takes into account neurodiversity.

And more worried (in the shorter term) about a world where there is:

• An increasing number of “computer-minder” jobs, such as truck drivers who just monitor their trucks, or Mechanical Turk-esque jobs where people manually add labels to data.

The entire article:

• How will we feel about it 10 minutes from now?
• How about 10 months from now?
• How about 10 years from now?

2. Unz on Race/IQ – The Rural/Urban Divide

The importance of urbanisation of the increase of IQ, which seems to counter the argument that genetics is important for IQ. At least, to me, it suggests that the most IQ gains to be made are improving standard of living of those in rural/poorer areas. But are there still genetic effects looking at urbanised people?

Also:

[The IQ of East Asian populations], seem almost entirely unaffected by even the most massive trends of urbanization. This has led me to suggest that for some unknown, possibly biological reason, East Asians can achieve nearly their full IQ potential without requiring the same beneficial “cognitive development environment” which white Europeans seem to require.

3. The Hour I First Believed – SSC

Only read this if you’re interested in stuff like acausal trade, multiverses, and superintelligences.

4. On the Phenomenon of Bullshit Jobs: A Work Rant

I suspect many have read this

A Return Of That Classic Section: What I Have Been Reading/Watching/Hearing

• Selfish Reasons To Have Kids (book) – Caplan (i) argues using twin studies that most parenting has little long-term impact (as long as you’re not a monster and aren’t poor); (ii) having children is generally good with the world; (iii) says you should think about how many kids you would want to have when you are 60, 70 or 80, not just when you are 30.
• The Good Place (TV show) – Funny TV show created by the creators of Parks and Recreation and The Office
• Very Bad Wizards (podcast) – Stimulating podcast by two smart people talking about whatever they want to talk about, sometimes with guests. Previous cool podcasts include ones about Effective Altruism and Why Are Buddhists So Afraid of Death

https://news.harvard.edu/gazette/story/2018/04/most-transformative-meds-originate-in-curiosity-driven-science-evidence-says/

2. Why I’m Not a Negative Utilitarian

http://www.amirrorclear.net/academic/ideas/negative-utilitarianism/

3. Why collabrative truth-seeking is beautiful

http://slatestarcodex.com/2017/03/24/guided-by-the-beauty-of-our-weapons/

4.

Where and how to have the most impact in animal welfare. Promising: corporate outreach, research and creating animal product substitutes. Less so: consumer outreach.

http://effective-altruism.com/ea/1s9/how_to_make_an_impact_in_animal_rights_a_survey

5. Bayesian Conspiracy podcast, here’s a suggestion: 61 – Biohacking 101

http://podplayer.net/?id=51307146

6.

Poll results can be suspect because people a) don’t pay attention, b) may be trying to mess with your data

http://slatestarcodex.com/2013/04/12/noisy-poll-results-and-reptilian-muslim-climatologists-from-mars/