all of the issues listed are closed so any recent version is fine.

also, you probably don’t need to deploy this unless you have a problem with bots.

there was a series of articles during his first term where translators in (i think) india complained that it made them look bad, because they worked in one of those modal languages where politicians use a specific modus as a matter of course. since his speech is so erratic it was basically impossible to represent in that “political mode” and it made the translation look wrong, they stopped publishing what he said and just produced summaries edited to look correct.

i think about that a lot, especially in combination with polled populations in some countries being more in favour of don than others.

sidenote, i tried to find one of the articles i remember but that’s all been overshadowed by an event from february in which he used a translator app on stage to try to understand what an indian reporter asked him in english.

Edit: found one: https://www.theguardian.com/us-news/2017/jun/06/trump-translation-interpreters
i think i mixed things up a bit, the indian translators were indeed producing summaries because he was hard to translate but it was the japanese translators that had problems with modal speech. makes sense really.

being able to deny involvement is sorta the point of stochastic terror

the name is widespread enough to have a wikipedia article

it’s a cartel

ddg doesn’t run it’s own llm, they’re just a frontend to chatgpt that (allegedly) strips out all the tracking.

a thought from reading the other reply, i think we can safely say that the numbers of abusers are around there, but i don’t think we can ever find out how many of them are doing it only due to getting off on the power imbalance rather than on the actual person they are abusing. it could be that it’s situational, like suicides. it’s not something i want to think about too much.

i also read that somewhere, that it’s not a question of attraction but a question of power. it’s still abuse though.

that’s insane stats. horrible.

…if 51 million us citizens are pedophiles i think psychology and anthropology has a lot of work to do

text was added manually. if it was all one pass there would be space for the text in the image. image models are sticklers for consistency.

like one of those italian airplanes from the 1920s!

tall ships are a hate crime to dwarves

what, the same one? do you sell it to the same guy every time and then buy it back?

well i- let me stop y- before that, can you first define “didactic”, “supposition”, “metaphysical”, “ass”, “one”, “ask”, “the”, “is” and “for”?

i thought he said he didn’t draw pictures

it’s her bedtime pancake, don’t judge

the muted palette, the block letters with rounded corners, the slightly inconsistent spacing and kerning, the way each line is angled just slightly differently, the way the background sort of hugs the text like it was especially made for this even though it is so generic.

not really what i was going for

one of python’s strong suits is it’s foreign function interface. it’s very easy to call code in other languages from python, which can speed execution up significantly