a non-exhaustive list


  • convert more of my model-based policies into equivalent model-free policies/heuristics to reduce cognitive effort
  • factor in the value of information gain to more decisions
  • be more experimental
  • explicitly write out the current meta questions i’m considering
  • recognize physiological causes (I used to be someone who only modeled myself psychologically and am now discovering the power of also incorporating in physiological data)
  • when should I resort to generating rewards for myself? is this is a sustainable long-term strategy or does it just result in fatigue and burnout?
  • stop forcing yourself to do things. try a bottom up approach for a while and just act heuristically based on what interests you.
  • do a thorough post hoc analysis of my motivation trends to figure out what experiments to run in future
  • NEED to do better experiments on myself or i’ll never improve. this means better experimental design, collection of data, analysis of data, …
  • keep in mind {X tragedy} as a baseline to life & be grateful
  • when you lose purpose remember that purpose is currently figuring out what purpose is
  • eventually i want to speak up about things i can’t speak up about now. it is annoying when you have to withstand short-term BS to get the power to change the BS in the long-term, but sometimes that’s just how it is. 🤷‍♀️
  • disappear from the world once in a while
  • I frequently ignore / block out entire areas on my mental todo list (adminstrative tasks, emails, boring stuff, ..). fix this / work into process
  • do a post-hoc analysis of my habit formation


  • there are things i overemphasize and underemphasize in people. calibrate.
  • talk with and see your family more
  • you get the most out of depth in friendships, not breadth
  • host things at our apt #community
  • figure out how I want to balance {social}-{work}-{alone time reading/thinking}
  • figure out how to do social-work context switching. either reduce the amount of context switches or get better at the context switch.
  • anything related to dating ???
  • revise my models of the value of exploration based on empirical evidence and adjust exploration-exploitation accordingly
  • spend more time with roommate, cfc, ex-roommate
  • don’t care about the irrational haters
  • too many people in SF are good at faking intellectual depth, don’t fall for this
  • in some ways i’m socially overcommitted. sigh, don’t currently know what to do about this.
  • just be yourself more.
  • conflict is not as bad as you think it is
  • listen intently. listen the way you would want others to listen to you.


  • theory
  • read more papers
  • recalling information on a regular basis to check that I’ve learned it. surprising how often I mess up concepts I think I’ve learned.
  • figure out a better note-taking system for papers
  • reduce activation energy to run experiments and write code
  • philosophy of science, causality, explanations
  • figure out what sort of working environment works best
  • take classes outside your field
  • invest time early to learn the fundamentals (more stats & opt)
  • summarize my research at different lengths (2 min, 5 min, 10 min) for different audiences
  • i think of everything in story-format, apply this to research too. what is the context? why do we care?
  • always ask yourself what the weakness of this is
  • talk about my research more
  • don’t be afraid to ask others to review drafts
  • read books more (2 causality books, deep learning book, some stats books)
  • often much more productive to write stuff down and restructure organization of information in your mind
  • probably should spend less time deliberating what path to take and just take one

health / being an adult

  • make a freaking budget and actually cook for yourself
  • get a more ergonomic setup
  • take care of your eyes
  • do routine checkups
  • FIX YOUR CIRCADIAN RHYTHM AND SLEEP BETTER (i lose so much productivity to sleep)
  • try fasting
  • my god why don’t i have more data on myself. do some research / biohacking binge.


  • generally, figure out what bullets i’m currently biting and what bullets I would be biting by switching
  • read about arrhenius’s paradoxes


  • find some interesting good streams of information to subscribe to
  • facebook is such a useless app, limit to using 2 times a day