November 22nd, 2023

edited 2023-11-22 · open in Roam

wondering if it makes sense to think of AI xrisk etc in terms of magic moreso than in terms of logic. not that logic isn't also very very relevant—all the tautologies & left-hemisphere stuff & math & so on is vital. but rather than being like "here's an argument about why this should terrify you" feeling something more about vast forces and rising to create a collaborative force that can meet it
vague thought that it wouldn't be wrong to call me a queer straight cis man
#2tweet and then we tweak the representation to look better, by cleaning it up
- - when you clean your room, you generally don't use a checklist—you do it by looking around and seeing the difference between what is and how you'd like it to be. when taking care of some sort of project, especially digitally, it can be hard to see the state of things, and to keep the vision in mind. with a piece of writing or a small coding task, you can sometimes look at it as a whole, but in many other cases we end up looking not at the situation itself but at a representation of it.
  - this has a few issues:
    1. the representation is kind of disconnected from the actual situation—is our representation apt?
    2. we can end up focusing on improving the representation, which may make it more attractive to use but is in some sense
  - and then we tweak the representation to look better, by cleaning it up
  - ---
  - likewise if you have to pack everything (eg when leaving a hotel room). but if you're packing to travel, you may use a checklist to synthesize a bunch of future possibilities into a current plan of what to pack
  - representations
#2tweet new linguistics theory:
one reason escher sentences (eg "more people have been to Russia than I have") are parsed as grammatical: comparative deletion is mandatory in English (eg one doesn't say "Doris looks more satisfied than Doreen looks satisfied."—the last two words are omitted) and so people doesn't expect to feel a senseful clunk when adding back the implied "been to Russia" at the end of the sentence.
- - new linguistics theory:
    one reason escher sentences (eg "more people have been to Russia than I have") are parsed as grammatical: comparative deletion is mandatory in English (eg one doesn't say "Doris looks more satisfied than Doreen looks satisfied."—the last two words are omitted) and so people doesn't expect to feel a senseful clunk when adding back the implied "been to Russia" at the end of the sentence.
wtf this crypto know-your-customer review has a finite list of allowable industries and job titles. always kinda scary when mandatory fields on potentially-illegal-to-lie-about-it forms have limited legible options and refuse to allow true legal facts to be represented
#2tweet how much would it cost to make the moon have rings like saturn? uhhhh asking for a friend
- - how much would it cost to make the moon have rings like saturn? uhhhh asking for a friend
#2tweet the climate activists who protested burning man this year might want to take note of the fact that festivals like these have been shown to propagate increased empathy out into culture at large

news.yale.edu/2022/05/27/transformative-effects-mass-gatherings-burnin…
- - the climate activists who protested burning man this year might want to take note of the fact that festivals like these have been shown to propagate increased empathy out into culture at large
    
    news.yale.edu/2022/05/27/transformative-effects-mass-gatherings-burnin…