woodcut/engraving style poster advocating "3 hours for work, 6 hours for rest, 15 hours for posting"

come on, have you never giggled at this image

if it wasn't for image sharing then yeah I'd say everyone can just go use IRC! but...

I think they do have the "program where people type at each other *and also send images that no participant involved in the conversation needs to host*" market on lock unforch. public image hosting is a hassle especially but not exclusively for porn

the training data is probably padded out with still images + zoom/pan. video is expensive unless you're Google

(my second para is total speculation, may not be relevant, I think the quoted line is the answer)

from the NPR story: "Cobb put a pause on her order until Dec. 11 to allow the Trump administration time to appeal."

also, part of the case that ruling was in was about the existing guards having been deputized to perform law enforcement functions. not sure the new 500 would be so deputized

tbc this does still seem catastrophically bad, I think a literal reading makes it illegal to write any software that doesn't check the user's age through this as-yet-nonexistent system starting in uh 2.5 months? presumably not enforceable as written but who the fuck knows

the bill is pretty readable here: leginfo.legislature.ca.gov/faces/billTe... and this article from the Verge also confirms that there's nothing about ID checks www.theverge.com/news/798871/...

I don't know, I can't find anything suggesting "confirmation" in the bill text. the article seems to make several mistakes including conflating this bill (AB-1043) with several separate bills signed around the same time, e.g. AB-621 (about deepfake porn) and SB-243 (about chatbots)

afaict this isn't really age "verification" in any sense, just requires entering a date in a birthdate field? even if it gets extended later it establishes one of the least bad possible frameworks for that to happen in: age checking ondevice, only a minimal "age bracket signal" is exposed to apps

the other point that struck me was about HSS being less "interpretable" from the donor side; it's harder to tell a simple story of how it will help. which is definitely a limitation of private charity albeit one that EA is intended to be less susceptible to...

Health Systems Strengthening — EA Forum Editorial note This report is a “shallow” investigation, as described here, and was commissioned by Open Philanthropy and originally produced by Reth…

found this review from an EA group of HSS interventions—they are advocating for more focus on HSS, bt some concerns they're responding to are indeed about lack of measurability forum.effectivealtruism.org/posts/c7b4Dm...

not an expert at all but I would imagine in a lot of cases healthcare infrastructure investment is a big upfront cost with a highly variable and hard to assess impact. curious what you're thinking of specifically though!

(one of the reasons longtermism is silly imo!)

I think the main reason a lot of EA charities focus on more immediate things like nets or cash transfers is that it's very hard to accurately estimate the benefit of things like building a hospital, which depends a lot on the circumstances and will have its biggest impacts many years in the future.

I think EAs usually do talk about QALYs. Givewell doesn't list QALY-per-dollar for their charities but they do cost-effectiveness studies e.g. for AMF docs.google.com/spreadsheets... the units there are an internal "unit of value" but link to more discussion of how they weight that