tammy's blog about
AI alignment, utopia, anthropics, and more;
hi! i'm tammy, and this is my blog. i've been running it for about three years now, and i thought i'd give a retrospective on how i got here and what i've been up to, especially in 2022.
in my early teens, i was very much into lego, but i also got into the belief that AI was going to be a very important, world-changing thing. and so, at the time, i described myself as wanting to be one of two things when i'd grow up: either lego set designer, or developer of strong AI — which used to be the term for general/highly-capable AI.
i must've read something, somewhere, about how if AI could improve itself, then it could improve how it was improving itself, and the whole thing would rapidly cause something like intelligence explosion — and the general idea of technological progress itself improving technological progress was the basis for the form of singularitarianism i've believed in to this day.
in my mid-teens, i started fantasizing of becoming important to the world — of being the person, or part of the group, who would save or take over the world, or something like that. but while my interest in various computer science topics kept growing, my focus with regards to world change started shifting towards politics.
around 2010 i discovered minecraft, and set upon the quest of building a game which would be like minecraft, but more like what i wish it was. this remained my overall main project, on which i'd keep working for about a decade; but i'd also work on a variety of projects generally intending to be better replacements to modern computing — better programming languages, better operating systems, better internet infrastructure, and the like.
university was okay. in 2014 i got my bachelor's in computer science, though most of what i learned about the field i learned on my own rather than from any institution. i tried getting a regular dev job as a backend developer, didn't like it, and decided i wanted my programming to be solely about things i actually cared about.
i ended up spending the latter half of that decade working on my game, while living on welfare and with the help of a few part-time jobs here and there, but otherwise mostly on my own at home, living the starving artist life. it's not that that lifestyle was particularly appealing to me; the lack of real-life socialization did feel increasingly bad. but i was otherwise feeling okay, and afraid of trying things. i'd never gotten around to emotionally realizing that i can go to places and do things — i'd mostly just been going along with whatever path of least resistance was available to me.
and i did have fun! making my game, working on my conlang, fiddling with programming language theory, and the like. i didn't output much in terms of finished things, but i did learn a lot of things along the way, and i got to do software engineering like i wanted to, which to this day i take great pleasure in.
and, while they all lived far away, i did make a bunch of friends. in 2017 i started using discord, and have made there most of the best friends i have today. my ideas about AI were that it would destroy everything and then do everything we'd care about — science, art, philosophy — better than us, and one of those friends shortly introduced me to the rationalist community and in particular to meditations on moloch and the orthogonality thesis. those realizations struck me, i read the sequences, and became convinced not just that AI doom was going to be bad, but that we could actually do something about it, and some people were working on that.
nevertheless, those ideas only stuck in the back of my mind while i kept doing the things i enjoyed doing. because of an interest which learning of AI alignment had sparked in me, i started thinking about values and utopia, and eventually made this blog in its present form in early 2019, with my first post. but it was only when github copilot came out, that i started panicking and refocusing my life significantly towards AI alignment, which i now believe to be the most important thing by far. this was part of a set of changes in my life which would culminate in 2022.
2022 was a year of enormous change in my life, more than any previous year. i went from comfily working at home on my blog and my game with no end in sight, to getting some money, traveling around a bunch, visiting the internet friends i've made on discord over the years, and eventually participating in the Refine research incubator which got me much more seriously and legibly involved in the AI alignment community.
in 2022, i started to actually do things, and by doing things i got a sense that i can do things. i'm not sure how i'd explain this to someone who hasn't gone through that, but, you know how going on a trip is a thing you can decide to do and then actually do? that is something that's historically been emotionally very alien to me, and to an extent still is. i feel scared about doing things in general, because i'm so used to the feeling that i can't or that something will go wrong or that that's just not what i am to do. but i keep trying things, and they keep not going wrong, and i'm okay so far.
2022 is also when, after many years of increased consideration, i moved significantly forward on my gender transition. this is a change that might be considered very important in one's life; but for me, while it was profoundly relieving, it was also one of the less abrupt changes i underwent this year. it's not like i'd been particularly pursuing a male-presenting social life; my real-life social life was almost non-existent before 2022, and on the internet i'd not been presenting in a particularly gendered manner anyways.
and 2022 is when my refocus into AI alignment made me aware of how much trouble i have emotionally accepting consequentialism and working on something that is actually important to the world. i've mentioned before that i've regretted changes i've undergone, but in fact working in AI alignment — as strange as it is of a situation to observe being in — is not unlike that wish i had, as an early teen, of being important to the world. and i have a strong personal culture of respecting my timeselves, especially the values i had as a kid; there's something that feels specially important about those.
i haven't fully reconciled how i feel about being someone who should dedicate my life as much as possible to alignment, even though i was pretty happy the non-consequentialist taking-it-easy person i was before, but i've made progress and overcome some less pleasant episodes of that quest already, and i hope i get better at doing what must be done.
it's the very end of 2022, and what a year it's been! i have some guesses as to what the next year, and however many years we have left after that, will hold for me: more alignment research, more traveling, and more meeting new and interesting people. but apart from those broad strokes, there's still a lot of unknowns and details to be filled in, and i hope they work out.
thanks to all the people who have supported me along the way; parents and friends of old and of new. i know that the person i've come to be, which i'm very happy being, has been very profoundly shaped by some, who know who they are. if we build this utopia that is better than anyone can yet concieve, then the help i'll have contributed towards that can in significant part be credited to them; and if we all soon die forever, which we probly do, then the real utility will really have been the friends we'll have made along the way.
unless otherwise specified on individual pages, all posts on this website are licensed under the CC_-1 license.
unless explicitely mentioned, all content on this site was created by me; not by others nor AI.