Erin Kissane

If you’re inside China there’s a more than decent chance that there are FAR fewer “radical Islamists” in reality than your (again, state-influenced) information streams have you believing exist inside unreality, never mind far more innocent people being killed by your government than you’d like to know. Read https://erinkissane.com/ (if you even can without a state-actor standing as a MITM) and replace Myanmar with China and you might start to see what I’m getting at. A lot of countries on this planet, and especially its current super powers, need a minority to demonize… you seem to be under the impression (apparently from inside the Great Firewall) that China is doing xenophobia and racism _better_? EDIT: by the way, my skip-the-Nobel-please first pass idea would be to talk to them and find out why their grievances have driven them to actions that the state wants / needs to [opp|sup]ress.

I read about that here: https://erinkissane.com/meta-in-myanmar-part-i-the-setup This remains one of the best things I've found on HN.

1. As much as I dislike the thought of law mandated overzealous moderation, you're probably right. On the other hand, if doing this will let us avoid another genocide [1] maybe it's worth it. 2. We're concerned about privacy, but ad supported businesses are as popular as everywhere else (just regulated a bit more strictly). There's probably more fear of processing too much personal data and being sued, but it's good in my opinion as a customer. 3. Also true. But, assuming you're American, isn't the selection of Europeans your know biased? Anecdotally, most of the interesting Europeans I know are not American :). [1] https://erinkissane.com/meta-in-myanmar-part-i-the-setup

In support of 3, 5, and 7 being the real drivers of the category listed separately just under them: This is a map of Mastodon instances. https://www.comeetie.fr/galerie/mapstodon/ I think it gives a good sense of proximity. From the place in this clustering where I spend my time (in the jorts.horse space, if that gets enough across), it is almost unimaginable that someone could get an experience reminiscent of LinkedIn... but that's just one spot on the map, right? I think my conclusions from this kind of thing are pretty different from Erin's in ways that could be anticipated from the framing in her earlier post: https://erinkissane.com/blue-skies-over-mastodon IMO that piece is worth reading first for context.

I'm happy it exists too. By far my favorite discussion about it was Oxide's last podcast: https://oxide-and-friends.transistor.fm/episodes/blue-skies-over-masto… Emily Kisane was on the ep, following their piece that blew up https://erinkissane.com/blue-skies-over-mastodon > In addition to Bryan Cantrill and Adam Leventhal, we were joined by special guest Erin Kissane and long-time acquaintances of the show Tim Bray and Steve Klabnik.

I imagine there will be the same problems as with Facebook and other large websites, that used their power to promote genocide. If you're in the mood for some horror stories: https://erinkissane.com/meta-in-myanmar-full-series When LLM are suddenly everywhere, who's making sure that they are not causing harm? I got the above link from Dan Luu (https://danluu.com/diseconomies-scale/) and if his text there is anything to go by, the large companies producing LLMs will have very little interest in making sure their products are not causing harm.

Reminds me of Meta mobile data partnerships in Myanmar. Same arguments and similar playbook. That did not go well for people in Myanmar. https://erinkissane.com/meta-in-myanmar-full-series

I don't think they should be intimidated, I'm just saying they are: https://erinkissane.com/mastodon-is-easy-and-fun-except-when-it-isnt But I wouldn't say it's a fully inconsequential choice, either. Your instance and how it's situated in the wider fediverse does matter to some degree. Not having a full view of the network, missing meaningful replies, at the mercy of some admin #fediblock drama, etc. Or maybe you finally find a good instance but your admin dislikes Meta so you're defederated from threads.net now. That's a bummer.

More on the same for those out of the loop: https://erinkissane.com/meta-in-myanmar-part-i-the-setup (https://news.ycombinator.com/item?id=37709284)

Any social network Meta produces is not merely more terrible than you imagine, but will eventually be revealed to be more terrible than you can imagine. For example: https://erinkissane.com/meta-in-myanmar-full-series

I think to start understanding the problem they're trying to address, you can to read this series on Facebook's role in the Myanmar genocide. https://erinkissane.com/meta-in-myanmar-full-series A LLM could easily find itself in the same situation without their 'alignment' safety system. It's hard to get a grasp on this because many of us can't imagine what it's like to be borderline literate, or from a totally different culture, etc.

> "Meta’s Yann LeCun argues against premature AI regulation". Let AI moderation commit genocide once, shame on Facebook. Let AI moderation continue to go unregulated, and... https://erinkissane.com/meta-in-myanmar-part-i-the-setup

A fitting rebuttal to Andreesen's manifesto is Erin Kissane's recent series about how Meta enabled a genocide in Myanmar. "Trust and safety" as enemies, indeed. https://erinkissane.com/meta-in-myanmar-full-series

>I wouldn't run any kind of publishing system for anons myself. It's potentially valuable for an actual social group though. That's pretty much how it works on the federated Internet. There are large open-access services run by communities with sufficient moderation capacity (to not get themselves nuked, anyway.) Turns out many "impossibilities" are trivial when you're not trying to abuse 1 billion active users at the same time through the power of their own (distr)actions - but instead you are simply trying to run a board for messages. And then there plenty of private servers, where publishing either is by invite, or does not have outsized reach in the first place. Those also defederate each other a lot, and many don't show you stuff from the big publics at all. There've been "bad people out there" always (or at least that's what the "good people in there" have been broadcasting, for about as long as I remember). The design/engineering problem here is how to figure out and deploy a relational dynamic that keeps hostiles at a safe distance. The practical problem stems from a technicality of how federation currently works: to display content from other services to your users, you have to mirror it on your storage. This mode of federating hazardous data is a real problem, and also it's exactly what some cheap-ass subcontractor of current-gen social media incumbents would be doing if said incumbents had the amount of good sense that they've demonstrated having (see e.g.…

OK, I spent ~an hour reading parts I [1], II [2], and III (TFA). My perception going in was that comment moderation is very hard, and essentially impossible at scale, and that Facebook/Meta cannot credibly be accused of literally causing genocide. Here is how my opinion shifted some after reading this series: The articles present a lot of evidence to support the notion that Facebook was well aware of clear-cut incitement to violence years before its escalation into large-scale genocide in late 2016. It isn't spelled out explicitly, but I think there are two things that make Facebook's initial apparent non-response easier to understand as incompetence rather than evil: * In the aftermath of the Arab spring, technology and social networks enjoyed a honeymoon in public perception that is hard to remember over a decade later. People, including (especially?) people in tech, could genuinely believe that making it easier for people to talk to each other was basically good, full stop. I'd suggest that this honeymoon only really ended after the 2016 US presidential election, which caused a lot of soul searching about information and echo chambers, and got even worse after the Cambridge Analytica scandal [3, this is a graph image from NBC news] * As part I of the article notes, "[t]he information landscape in Myanmar is so unstable that accounts of any given incident conflict, often in major ways". This dynamic seems to appear in a lot of places where well-off westerners interact…