Objects in JavaScript, just as in many other programming languages, can be compared to objects in real life.

@DakkiReads and one important point to note here is that typescript is literally a microsoft project. so when lots of Very Special Boys make a point to say "your language is archaic if it doesn't support typescript, here's my half-assed sketch" and they all mysteriously shut up when the maintainer says "here are two ways we could do this", i never know how much of it is a literal influence campaign from a monopolistic competitor vs just incredible obtuseness

@DakkiReads literally what are you doing caping for forks of my fucking language bro he links to a github repo that describes itself as "coffeescript rewritten under proper compiler design principles" what you mean utterly fucking inscrutable like usual?

@DakkiReads one of the fucking losers who won't iterate upon their pet proposal to make it backwards-compatible is an incredibly pretentious MIT prof whose profile i landed upon the other day when trying to find a single prof there i would want for my advisor

@DakkiReads the repeated FOMO propaganda that "coffeescript is losing out without types bro" is absurd when the maintainer specifically said (1) we don't have enough contributors to do a major release (2) so it needs to be backwards compatible and that's TOO FUCKING HARD FOR THE GENIUSES IN THE THREAD

@DakkiReads i was trying very hard to analogize "the spirit of coffeescript" with "maintaining backwards compatible syntax" and trying to use both of those as an epistemic basis for "coffeescript is a leader not a follower". i don't know too many frontend devs but i was very much trying to appeal to any of their artistic spirit in the audience of potential coffeescript contributors (the compiler is extremely small and thoughtfully architected)

@aud "the nickelback of pop punk" is so fucking funny

they said follow this link if you have ideas about type syntax for coffeescript (https://github.com/jashkenas/coffeescript/issues/5307#issuecomment-2450591744), but it turns out most people who commented do not have any ideas about that and had more of a comment than a question. this inexorably set in motion a series of events, starting with the first time i have ever invoked the xz backdoor as an analogy, and culminating in:

What does a type taste like? When you pick up a function, how much does it weigh?

twitter was too late to unionize—or put another way, its management skillfully avoided creating too much pro-union sentiment, so it was left unprepared for the takeover. so outlets like the OSPO can also be used to placate and stimy organized labor. but i think they can also form the seeds of it too

the open source program office (OSPO) at twitter showed me how to frame arguments about societal organization within the narrow terms of bureaucratic efficiency—and succeed in making change

my paper comparing distance metrics for cell signalling took two years to be published (no i'm never going to shut up about this), because the scientific establishment only understands progress in terms of the accepted canon of artificial benchmarks. new ways to measure and compare become new axes of hierarchy—and that sets the existing hierarchies off balance. how presumptuous!

i think figuring out what and how to measure is the very purpose of science—using domain expertise to identify what to quantify, so you can employ all these powerful numerical methods we have to solve problems. to my dismay i've come to believe most scientists instead believe that wielding numerical methods imbues them with a sort of intrinsic righteousness that others are unable to grasp

that was 2020. at this point "open source" "AI" has been whittled down to meaninglessness. and as i've expanded my thinking outside of free software's narrowly technical focus, i see that software like LLMs serves the purposes of accumulation via austerity and deskilling. they are extremely skillfully and carefully designed to achieve this, because "democratizing" their access requires fundamentally restructuring society around them. i gasped

a few weeks later, comms announced it would be made open source, and later that year the twitter META ethical AI team hosted a groundbreaking "AI" safety hackathon at DEF CON

nobody ever thanked me for it but when people found the twitter autocropper (which used statistical methods) was racist (focused on white people and not Black people) and misogynist (highlighting cleavage), i was incredibly, unbelievably annoying about open sourcing it in the internal slack, so users could trust what we were telling them about how the software worked (and so i as an employee could trust what they were telling me and others about how it worked and why it was used)

it was rocky at times. it chipped away at all of us. but nothing lasts forever and we damn well made the most of those electric months and years together

i learned so much from remy decausemaker and alan velasco. alan was an intern but he regularly did my job better than me and i am still learning from his ideas and implementations. remy showed me that love is more than just an academic concern in building lasting infrastructure.

in case you want to know why i am like this. this is what radicalized me. heat light love. twitter oss showed me you can change the way people think and work for the better https://opensource.twitter.dev/

@aud @lina @ireneista i'm very interested in distributed text search i will keep an eye on your work

@aud @lina cc @ireneista who i'm paraphrasing