February 08, 2025
Not a bullet journal in the Pinterest way, just a journal in bullet points
(2.18.25) I notice that most “intellectual” conversations I have lean much more towards the other party’s area of expertise?
(2.16.25) Greek life is popular because it’s a safety net: it confers some default amount of status, it offers a default social life, and it offers clear boundaries for what relationships one should or shouldn’t put time/effort into
(2.12.25) one of the most jarring realizations I’ve had is that even purported experts, with stellar careers, are often just fundamentally wrong about something. it’s weird to re-confront my relationship with authority, especially intellectual authority
(2.11.25) I would love to see completely gamified technical education – progressive levels, instant feedback, for problems of varying time/effort scales – already learning against model-designed curricula sort of gets there
(2.8.25) when I was insanely sick for the last ~week I hit upon some weird pit of despair: why do we put all this on ourselves, have so much agency, embrace so much responsibility for our real lives, when sometimes it’s all we can do to just survive/feel ok? now that I’m less sick I’m getting my motivation back in bits and pieces haha
(2.8.25) why is it so much easier to get music off tiktok clips? (a) repeated, semi-forced exposure (b) increased chance of hearing the catchiest parts of the song (or at least not the intro/first verse) (c) most importantly, the vibe has been matched for you. really good music evokes some sort of feeling or urge: to roll the windows down and speed on a freeway, to be in love, to dance in a certain way. a user that posts a video to a certain song has chosen the visuals to match whatever the song evokes; this is usually faster and lower effort for the viewer than becoming attuned to that feeling by themselves; videos are 15-second vibe capsules; this is probably really biased towards the kinds of tiktoks I see
(2.8.25) I’m not sure that we can have safe superintelligence without some amount of pure blind faith; all my models of the bridge to the far future involve some leap
(2.3.25) how much of “realism” can be attributed to the camera / how does contemporary technology + media shape our internal perceptions of the world
(2.2.25) when science + philosophy is solved by models, I’d like to know if it was necessary, if we were hitting some upper limit to human abstractions and collaboration
(1.28.25) few things feel worse than not being taken seriously
(1.28.25) it seems like every startup must have a little bit of fraud in the early rallying stages, until the phase transition happens at some critical momentum + scale
(1.27.25) it feels really good to have top-line direction in my life; I think it makes me a more motivated and more disciplined person across the board. usually I model my behavior as some interaction between my “true self” (i.e., Christine who actually makes all the decisions second-to-second) and my “higher self” (Christine who knows the kind of person she’d like to be and is actively trying to steer herself, work through her emotions, etc.)
(1.26.25) what ideas in the American psyche are just luxury beliefs/memes of long-term hegemonic stability?
(1.24.25) I would benefit from more orchestration experience – organizing something in the real world – I miss hosting high school house parties
(1.21.25) the world neds more experiential art, more gimmicky surprises
(1.19.25) for a brief 12 hours we thought TikTok was gone, and I felt genuinely mournful. in some ways I was raised on the musical.ly/TikTok/early YouTube culture, and it all feels very deeply intertwined with my teenage/high school years. mass Internet culture has its evils but it’s also a unique and deeply beautiful experience to share ideas, trends, experiences with millions of people.
(1.17.25) feeling the AGI, also feeling the techno-religion. somehow
(1.16.25) are you more of a “ideas are cheap, execution is expensive” person or an “anti code monkey” person
(1.15.25) agency is one of those human traits where the distribution is better visualized on a logarithmic scale. or at least, the effects accumulate exponentially –> massive apparent dynamic range
(1.14.25) the main reason to be hotter is to have more societal tolerance for your weirdness
(1.9.25) literally everything can be boiled down to just being attuned to what you want. curiosity is extremely multi-scale; on the nano end it is just noticing what you like/dislike, and interrogating the emotions provoked by stimuli. intuitive eating! people are jealous/status-seeking because they have no internal sense of self.
(1.9.25) different scientific/engineering disciplines have vastly different intellectual cultures; one main dimension is “first principles/autodidacts rule” vs. “defer to authority”
(1.8.25) notes from FashionX: (1) intentionally create the improbable (2) cities and spaces balance transmission and transgression to various degrees; think SF vs. Paris
(1.7.25) being trapped on a bad/boring large group project is the worst way to (a) kill passion and (b) lose self-confidence
(1.6.25) isn’t it amazing that every social network has its own emergent, niche social norms?
(1.5.25) pictographic languages are much harder for new learners to bootstrap (think phonetics in Chinese vs. English); the compositional space of symbols with literal meanings is only so large. maybe pictographic languages must evolve ~slowly at first, s.t. symbols are sedimented before the language is extended to higher levels of abstraction
(1.3.25) two career thoughts. (1) it’s hard to force a career that makes use of your comparative advantage; not every “intersection” is fruitful. (2) choose work you would do at every stage of the prestige/compensation/”coolness” ladder, not work where only the best outcomes feel desirable
(1.2.25) the best way to balance rationality with empathy/”normalcy” is to approach everything with humility. be deeply curious about but not above anything; irrationality is a neutral behavior
(1.1.25) happy new year! my main intentions for the year are (a) authenticity (b) epistemics (c) creation
(12.31.24) is there any real truth to taste? how would we actually test “primitives of taste”? this is related to convergent vs. divergent experiences of experts (see lower). or, is it just authenticity and alignment with the zeitgeist (i.e. the brat album)?
(12.31.24) someone should write out the anthropology of progress and the way technological growth has been felt culturally
(12.31.24) people act the way they act because of how the world feels to them; on one hand we may never understand subjective perception and on the other hand that is all there is
(12.29.24) the path to any good agent/labor-replacing product passes through the “exoskeleton” phase (IDE, platform, etc.); the action traces & proprietary knowledge are right there
(12.29.24) self-expression is our most basal form of freedom. also, specificity is deeply tied to authenticity
(12.29.24) losing track of how many “you should think about AGI/ASI” conversations I’ve had this winter break
(12.26.24) in general, a fulfilling and spiritually healthy life involves the growth and then relinquishment of attachments, the transition from active participant to content but passive appreciator. we undergo education, build careers, and nurture families, until eventually our prime years for earthly rewards are beyond us; then we relinquish attachment to a more selfish world and externalize our sources of pleasure, finding fulfillment in enlightenment, simplicity, or even observation of the world. for example, Hindu ashramas break life up into four stages: Brahmacharya (learning), Grihastha (wealth accumulation, expertise, family creation), Vanaprastha (retiring), and Sannyasa (focus on enlightenment). one of the weirder things about AGI/ASI is that it will force ~generations (including mine) to (family life aside) skip rather quickly to the end, without full careers or much opportunity to be active players in the world. in a superintelligence universe humans will not be active shapers of fate; it seems like healthy,non-hot-wired-brain happiness must come from enlightenment, detachment, and enjoyment of the universe itself.
(12.26.24) I am trying to (a) be more cognizant of my anxieties and subconscious claustrophia and (b) actively lean into them when I think those fears are irrational. More or less exposure therapy. Today it was piercing my ears.
(12.25.24) what childhood story do your parents always tell? My parents talk about when I allegedly wrote down “Ode to Joy” from ear after hearing it on a Kid’s Meal toy
(12.25.24) it would be unfortunate but poetic if the technology humans create eventually ends up destroying us. The optimistic view: technology is the driver of human progress, the curer of diseases and the bringer of abundance. The pessimistic view: technology is inevitable, the local minimum solution to present problems, and a necessity for survival, but every additional innovation is also a greater prison. A hedonic treadmill and a red queen race againsts ourselves that we can never backtrack on.
(12.25.24) almost everything is more fun if it has a fast feedback loop; this is how I get into “flow”
(12.25.24) prestige chasing can never end because the motivation is access to spaces you wouldn’t normally get, thus the end goal becomes unsatisfactory as soon as it is achieved
(12.22.24) I wish I could go back to when I was 15 and fine-tuning GPT-2 to make Rupi Kaur parody poetry. I wonder why I didn’t see all of this coming back then?
(12.22.24) is economic stability and growth a precondition to social stability? is bureaucracy and over-governance both a cause and an effect of stagnation? bikes and surfboards are much more stable when they have momentum
(12.22.24) if you are communicating something that may not be received well, you really should care about your image. @Yudkowsky, OpenAI Santa aesthetic, EAs in general
(12.22.24) are we transitioning from an “inter-generational” to “intra-generational” means of cultural normalization and teaching? as in, are people today “learning” more from their peers (Internet culture, etc.) than their elders?
(12.22.24) as an exercise, we should all concretely write down (a) our ideal AI future and (b) how we would govern if we were a superintelligence (lol)
(12.22.24) (assuming slow-ish takeoff, which IMO is the only scenario we should even be really thinking about) someone should be the cultural anthropologist of the AI transition
(12.21.24) I both feel like there’s a hole at the bottom of my epistemics, and also that I basically know ~nothing. I wish I could ingest and process all this information faster!
(12.21.24) even if I know that because I’m in position X, there will probably be people ~similar to me also in position X, I always fear adverse selection. related: if you’re not already excellent, how do you find a {chill run club, casual classical orchestra, recreational dance group} that is still “good enough” to be meaningful?
(12.21.24) absolute things that bring me joy: learning, solving puzzles, being awed by the Earth, cute human cultures, feeling understood, dancing
(12.21.24) startups should either be creating the footholds for future AIs (agents, domain expertise, etc.) or imagining the primitives of our post-scarcity human future (interfaces, joyful technology)
(12.20.24) most of the world operates deeply irrationally, hence there is always a local optimum to be found by applying first principles thinking and rationality. does that really mean we should take those ideas to their logical conclusion?
(12.19.24) new ideas feel like new eigenvectors, or partitions, or pivot points, for an existing jumble of concepts. you pull up the fabric at some junction and the pieces fall into place.
(12.19.24) are there domains with “convergent” vs. “divergent” expertise? i.e., given X number of hours of non-normative training/attunement, are you more or less likely to share opinions with your peers? my guess: fields with “ground truth”, i.e. science, must be convergent to some extent; fields that are more deeply about idiosyncrasy and taste, such as music or painting, are probably more divergent