@inthehands@mattly the tool for that is shame. People should be ashamed to stand at the microphone in their city council meeting and insist that the earth is flat. But shame is learned, and we're letting the people who think the earth is flat do all the teaching of it
@mattly Extremely relatable, and not a problem at all
I'm hoping that in another couple of weeks I'll be able to have a first test instance online, and that should make it so much easier to share and discuss and contribute to the whole project
@inthehands I linked to an article from tidelift in that post, as part of a discussion that money is at best part of the solution, and would just exacerbate the problems in isolation.
That link now redirects to sonar source, because they bought tidelift. I think rereading that section with that context really accentuates the point
@mattly I bought myself a home server earlier this year, on the premise that laptops might just become unavailable in the US in the near future, but hoping that we can sustain a 2nd hand market for server hardware for some time
Just once, I would like to be wrong about this kind of thing
@mattly It's more about the star wars at the moment. I'd likely go back and finish squadrons, and then maybe play battlefront 2. I skipped it originally because I thought it was just a multiplayer shooter. But they reused the name for a game with an actual story, and it turns out to be really impactful
@inthehands I'm also not claiming that there are no positive uses, except in the narrow case of the actual products that come to mind when we talk about AI. Large language models definitely have some positive uses. Likely many uses that would more than justify their creation, if it were not so extractive and resource intensive, which is very possible. ChatGPT, the privately controlled plausible fiction generator, does not.
@inthehands honestly, I struggle to find any other uses for the things we're calling AI.
People are very quick to defend the component elements of the technology, for some reason. But as it exists in the real human world with real social and political dynamics, that's what AI does.
@inthehands@KimPerales that very much depends on what the purpose of AI is. If it's to codify, reproduce, and shape what things are knowable and expressable, and to put that power in the hands of specific people, then this seems right in line with that.
@dalias@hipsterelectron 3rd party servers don't constantly re-iterate reply collections every time they encounter a post that claims to be a part of one. Because it would take forever, and it would produce an enormous traffic amplification attack on every participant in the network.
@dalias@hipsterelectron the de facto definition of a reply is any post that sets a value for the inReplyTo field, and the set of instances that can see it is any instance that received it, plus the few that might later request it.
Whether it gets added to a replies collection is basically never a factor.
It's also relevant that there is no way to query a collection. All you can do is iterate it. So the only way to confirm an objects presence or absence in a collection is to iterate the whole thing. Which contributes to the wild west authorization.