update
mostly may/june
read more than usual, partially because i went home (texas) for a week and there wasn’t much else to do there. very mild spoilers ahead
sirens of titan was strange. the first 80% of the book is poorly written in my opinion and essentially consists of characters doing very nonsensical and improbable things so that at the end vonnegut can say Actually There Was A Reason For All The Random Things. that being said the ending is very beautiful so overall i still thought it was worth it. my advice is to get through the first 80% as quickly as possible because the details are quite arbitrary and really do not matter
finally read the three-body trilogy after putting it off for ~5 years! the three-body problem was cool because of the sheer number of ideas packed into it - drought cycles caused by a three-star system, an army using flags as nand gates to emulate a computer, a weird application of the one-electron theory, and so on - how does one person come up with all this? the dark forest was my favorite, mostly because of the ending - i have a dream that one day brilliant sunlight will illuminate the dark forest reads like an ordinary sentence but the payoff after getting to that point in the book is incredible. death’s end was a bit too bleak for me to enjoy. there was a comment i liked on reddit: it's hard to move on from this series. it changes the way you look at the sky
tomorrow and tomorrow and tomorrow was very good. i don’t really play computer games and this is the first time someone has been able to help me appreciate games as storytelling forms / works of art / vehicles for self-expression. i usually hate time skips and alternating narrators but both are done well here. the prose is also unusual in a way that i can’t describe - if i had to describe it i would say it’s very sharp? but that’s not quite right (every once in a while i come across a book where i can recognize the prose is good but am not a skilled enough reader/writer to pinpoint what is good about it, which frustrates me greatly. the most recent instance prior to this was when i read circe in 2022)
read goodnight punpun because multiple friends suggested it over the last few years and god it was awful. first time in my life that friend recs failed so spectacularly. pretty much every character in this series is disgusting and weirdly sex-obsessed with the exception of two minor characters that only appear ~50 chapters in. i understand the main characters are victims and the author is trying to make a point but seriously you didn’t need to make everyone so unlikeable, at that point you’re just alienating readers
why fish don’t exist was decent and quick. i didn’t really like the narrator but the content (an exploration of the first president of stanford and of fish taxonomy) was good
turtles all the way down was successful in doing what it set out to do, ie. portray what it’s like to live with ocd / anxiety / intrusive thoughts. unfortunately it was so successful in doing this that i had to skip half of the book - the thought spirals were stressing me out too much and i was unable to keep reading them. i am especially sensitive to this kind of thing though and i imagine a less sensitive reader would find it quite interesting
not books: really liked tell your friends you love them (friend art blog) and the port huron statement (a really beautiful manifesto from the 60s; i read it a few years ago and forgot about it until recently)
i guess i still do research at imbue or something
moved offices from castro to embarcadero so my commute is much shorter now. i enjoy being able to see the ocean every day!
i spent some free time learning random matrix theory, mostly to figure out where the semicircular and circular laws come from. i couldn’t decide whether to read anderson (an introduction to random matrices) or tao (topics in random matrix theory) so i ended up reading both at the same time, which turned out to be a good decision - anderson likes to give the shortest proof possible so you get very direct and poorly motivated arguments, while tao likes to take hundreds of pages to build context, and it’s quite helpful to get both views at once
finally started using cursor and i love it
we released some work which i’m quite happy about! i helped write much of the infrastructure blog post, which discusses all the steps that went into setting up a large h100 cluster and training 70b-sized models on it, along with open-sourcing some code
some road trips
went on a surprise trip from 12-4am one night to see auroras. it was my first time renting a car and driving in california! driving on the golden gate bridge while it’s foggy is cool because you can barely see anything and then giant bridge pillars appear out of nowhere. we were unable to see auroras but overall i was surprised by how well everything turned out - i started planning at ~7pm and by midnight 6 people showed up at the zipcar location
went to rapid city, south dakota with some friends from school! we drove through colorado / nebraska / wyoming on the way. the highlight was a hike in the badlands, which included a flimsily-supported 50-foot ladder:
i am quite scared of heights so going down was a struggle. i started hyperventilating and gripped the ropes so hard my shoulders were sore for days afterwards and it didn’t help that i was holding up a long line of people waiting to use the ladder
i have felt quite underprepared in all my recent travel-related endeavors, but nothing terrible has happened so far and i have been getting better at logistics so i think things are going well? i am often reminded of that sentence petey wrote - you don’t become capable of doing hard things before doing them; you become capable of doing hard things by doing them. not that travel logistics or road trip driving or climbing rope ladders are particularly hard tasks in the grand scheme of things, but these things don’t come easily to me either
