I used to use a .txt file like this for over 10 years that I edited with Vim and could also edit on my phone using Dropbox. And then I used the Mac Notes app and realized the search, sync, autosave experience, and other properties made Notes way more useful for me.
Similarly, for over 10 years I used to compile my CV with LaTeX using a fancy template I designed, until I realized the time it took to edit it was just not worth it. So I pasted it into a Google Docs file with Arial font and just text bullet points and no rich text features. My life became way easier and my CV looks way less annoying.
(This may be irrelevant to the original post but) I found that a lot of the optimizations I thought were good for me were stemming from an unnecessary attachment to a minimalist technical aesthetic.
If your Honda Civic had a problem, would you take it to a physics professor who teaches thermodynamics or your local shop that has fixed thousands of cars with the exact same issue?
I am grateful that exceptionally capable people exist in the world. I find them incredibly motivating because their successes keep elevating the lower bound of collective human potential and set a known standard for what’s possible. I find that ignoring the bar for success (at whatever level one strives for) sets you up psychologically to justify mediocre efforts, when in fact most humans have incredible potential. I find that envy is a separate issue that can be managed by thinking carefully about choosing the right role models.
Love the post. Just FYI, your post is not mobile-friendly. When scrolling down on iPhone it’s impossible not to accidentally shift the viewport away from the left margin making the left side hard to read
Knowing how to be alone and being social is not a rigid dichotomy. I think it’s clear what Tarkovsky means: that there is immense value in self-discovery. I’ve heard similar sentiments from other great artists.
Grothendieck famously attributed his mathematical confidence to a long period of solitude in his youth where he independently recreated some major results.
The late Harold Bloom used to say that reading is ultimately a solitary endeavor—that its purpose is to learn to have a conversation with yourself.
However I think for reflective people there is a “danger” on the side of being too entangled in this conversation. Socialization can start to seem banal and unambitious in comparison. But I’ve found that this just suggests that you haven’t found people who aren’t in the habit of talking with themselves.
I find this story apropos: There was a screening at the Yale Humanities Center of Tarkovsky’s “Ivan’s Childhood” that my roommate wanted to take me to when we were undergrads. I only reluctantly agreed—it was freezing outside and I had been cooped up in my room wrestling with a statistical mechanics problem set. Frankly I didn’t feel like we got along. But my decision to go became the beginning of one of the deepest friendships I continue to have and my endlessly enriching relationship with film and literature.
I love Standard ML. As an undergrad, I wrote a compiler in it (through Zhong Shao's course at Yale), and it remains the most fun I've had using a programming language, let alone a functional programming language. The syntax is so simple yet it is extremely expressive. Hindley-Milner types are great, and the pattern matching is slick. SML also strikes a logical balance between purity and practicality.
I bit my nails unconsciously for 15 years. Constantly bleeding and got to the point where my thumbnails were gone and I thought they wouldn’t ever grow back.
I got braces for my teeth, the biting immediately stopped, and my nails grew back to perfect condition in a few months. In dealing with the discomfort of braces, I developed a habit of pushing my tongue against my teeth which completely replaced my nail biting habit. (To be clear, the braces per se didn’t deter me from biting.)
I do this all the time now unconsciously. Fortunately it’s imperceptible and doesn’t do any harm.
The authors claim in the appendix that "linear models were estimated due to failure of the maximum likelihood algorithm to converge for logistic models."