knowing what to do

i watched the demis hassabis documentary over the holidays and it greatly exceeded my expectations - i learned a lot despite already being familiar with deepmind’s work (whereas, for instance, i found the alphago documentary somewhat boring). the documentary portrays demis as someone who’s been obsessed with understanding the nature of intelligence since a young age, and does a great job of connecting all his projects to this thread

some people, after consuming narratives about individuals who have been dedicated to one thing for their entire life, will panic about not having found The Thing in their own life and then begin searching frantically for it. i think it’s a natural reaction to have, especially if you’re younger / less secure / prone to comparison, though unfortunately i also think it’s misguided. every time i followed the urge to Figure Out My Life As Quickly As Possible i realized a “truth” about what to do that ended up being completely wrong in hindsight (eg. working in crypto, doing psychology research, building a large collection of webapps that i haven’t touched in the years since). it’s very easy to fabricate epiphanies under time pressure

one thing i’ve learned from the last few years is that i should stop pretending to know what my future preferences will be. if you asked me back in 2022 i would’ve told you that {safety / alignment, sci-fi, hci, fashion, dance, games} were unserious subjects and now i think about all of them regularly. meanwhile many of the topics i thought were really important, like {performance engineering, compbio, zero-knowledge proofs, teaching / education}, are no longer part of my life. some of these changes come from trying new things, some of it is the result of becoming less mimetic, and a lot of it is downstream of ai progress changing the landscape of what makes sense to work on

---

recently a friend asked me how i knew i wanted to invest a lot of time into writing. i think the answer is that i never did? i spent five years (~2014-2019) posting on a math website because i needed somewhere to document all the math i was doing, and along the way my math blog turned into a feelings blog because i didn’t have any friends to talk to at school. i racked up a few hundred thousand views despite being a shitty writer because people on that site liked that i was good at MATHEMATICS. then i went to college and stopped doing math, but by then i’d gotten so many reps that internet writing came fairly naturally. now it is 2026 and here we are on substack dot com

i think the only reliable method for figuring out what to do is to spend a few years trying everything and seeing what sticks. interests are emergent rather than discovered. many of my friends are reaching the age where inspiring media celebrities (eg. billie eilish / alysa liu / wembanyama / etc) are now younger than them, which is annoying because it gives the impression that if you haven’t figured out what you like by now then you never will. i think that is really silly. i’m reading a new sci-fi story every day and when i look at the author intros a lot of them are part-time / have a day job in tech or consulting or some other unrelated field / didn’t publish anything good until they were old. there’s an excerpt from the motern method i like which goes something like:

You’re 45 years old. You’ve always wanted to make music but you’ve never done anything about it and you don’t know where to start.
Great news! Life is long.
You could spend several years learning an instrument.
By the time you’re 48, you’ll be good enough that you can start writing and recording songs.
By 50, you might feel ready to release your first album.
You’re only 50, and by now you’re good enough to put out a new album every year.

---

the past few months have felt unusual for me because i did quarterly (life) planning for the first time. basically i listed all the major projects i wanted to do between january-march 2026 as well as how long i thought each one would take, and then i made a schedule for when to start and finish each one (“project” here can mean any number of things, eg. interp research / going to taiwan / organizing birthday celebrations / building some writing tools with no additional details figured out at the time of planning). it’s strange because for the first time since leaving school i have a clear idea of what i will be doing, months into the future

(obviously quarterly planning only accounts for things that i am trying to forecast. for instance, i still have no idea what my next hangout or blog post will be)

i thought having a plan would feel different. i thought maybe i would no longer experience late-night anxiety, or that i would be more certain about what i wanted and how to get it, or that for once in my life i would stop second-guessing my own decisions

but instead, what i discovered is that all those feelings still remain. the only thing that’s changed is that in addition to those feelings i also have a plan, and instead of spiraling whenever the feelings surface i can listen to the plan i already have and direct all those feelings towards making the next plan better. there might be nothing in the universe that can prevent my hyperactive brain from crashing out, but that doesn’t mean i need to listen

---

i used to find not knowing what to do unbearable. this resulted in me being especially moody and annoying to be around from 2021-2024. but then at some point i realized the uncertainty itself was not the cause of my misery; instead the cause was a self-imposed expectation that i should have figured everything out by now. it’s similar to how being single is totally okay, but some single people make themselves miserable by expecting to have a partner and then constantly berating themselves for not having one

the expectation of having already figured everything out seems to be quite common among my friends. i’d guess it’s largely caused by: comparisons to other people, general impatience and anxiety, lack of appreciation for the journey, being accustomed to solving closed-scope problems that already have well-defined answers.

i don’t know how people typically learn to let go of this expectation. in my case the most helpful thing was doing very open-ended research where uncertainty is the default state to be in. and once i did let go i was able to explore much more productively, probably because i was no longer implicitly judging / pressuring myself