MudMan

Pi and Requiem for a Dream as a one-two punch.

“Being smart is such a curse I’d rather get a lobotomy” is boring, self-serving and trite, but the Reefer Madness-level “drugs bad” thing in Requiem is unbearable and requires every character to be entirely nonsensical.

Trainspotting was four years old by that point. How critics let Aronofsky get away with it is beyond me. To this day the closest I’ve been from walking out of a movie theatre. The only reason I stuck around was it was in a festival and I was there with other people.

May I suggest this is news in exactly one household around the time of Christmas dinner and perhaps not… you know, the World? I’d suggest that to the BBC, too.

True. Lies are free, information is paywalled.

The takeaway can’t be to stick with the lies, though.

I mean, that’s great and you’re well within your rights, but that’s not what people generally say when they express outrage about AI scraping. People straight up call it theft very often and seem to consider using online content for training is the equivalent of copying or distributing it.

Which stands out to me because that was not what happened when the EU decided that Google News was effectively piracy after a whole bunch of news outlets complained. The consensus there seemed to be that it was a bummer to lose the service despite all the scraping.

I just don’t think that’s a reasonable view, and it’s certainly a marginal one in the community. Nobody is out there claiming that the core feature of Fedi apps is self-hosting a tiny social network for your friends, disconnected from every other piece. The selling point is supposed to be that your tiny, self-hosted instance is still connected to this distributed, crowdsourced larger network.

Building a social network sure is hard and requires a building a lot of software, but unlike other pieces of software, social networks carry a LOT of additional costs to run at scale and make no sense to run without the scale. You can host Jellyfin for your small group of friends. Maybe a chat server or a list service, not a forum or a link aggregator.

In any case, even if you are an outlier and see that as a valid use case, that’s definitely not a majority view, and the Fedi community has both ambitions to get larger and an expectation that this will be done with effective moderation baked into the service. You and I agree on the existence of that problem, we just disagree on the resulting state after it surface.

You’re still thinking about it as an asymmetrical problem. Taking one portion that has a problem and isolating that from the rest. I’m saying if every part has the same problem that doesn’t solve it AND it means the entire network is no longer interoperable, which was the entire point from the start.

What you’re ultimately saying is that you can have a small interoperable network or a large centralized network, but not both. Which, if you’re right, begs the question of why try to decentralize and federate in the first place if you don’t have a solution to secure that arrangement.

And, to be clear, even in that scenario now you have an isolated, self-run social network that has exactly the same moderation issues and running costs as Reddit or any other alternative.

I am old enough to remember that IRC had more in common with 4chan than modern social media and that moderation of atomized, non-interoperable forums was either just as bad or handled at much smaller scales by people with commercial interests.

You care about how many bad actors there are if they are enough to be in every instance. Again, you’re presuming that bad actors will choose a specific instance to populate. You can’t defederate from every instance that allows people to sign up or you end up with a group chat instead of a social network.

That’s the Fedi-wide problem to solve if it ever gets truly popular. If I put together a bot farm or a sweatshop tomorrow that targets every instance of every Fedi app with multiple spam signups per minute how would you stop that? Especially if I’m not immediately posting spam and instead generating bad content slowly over time.

What if instead of one person doing it it was thousands? How high are the garden walls at that point? Is there anybody left inside them?

“There are tools to do that” is a bold assertion, but nobody has been able to explain to me what those tools are or how they’d work at scale. I’m all ears. Even if I don’t think it’ll be needed I’d love to know what the plan is, if there is one.

Agreed, for sure. If anything, decentralization makes those things harder, I’d say. And also agreed that there are benefits to decentralization along the lines you mention. Those two things can be true at the same time.

I think it’d be cool to figure out what the toolset to handle those issues is before they become a problem. Or, honestly, just because figure it out would be a meaningful challenge and may move the sorry state of social media in the right direction just in general. That said, there is a LOT of overcomplacent assumptions, at least in the userbase, regarding decentralization being a magic bullet. I think the development side is a lot less… I don’t want to say naive, but a lot more realistic about the challenges, in any case.

No, see, you’re assuming that this is a problem for one instance. Which makes sense because there’s nobody here and not much incentive to target people who are.

If you’re the size of a Twitter (and that’s a couple hundred million accounts) or a Facebook (about ten times that), then there are more than enough people to be targeted by more than enough bad actors to swamp EVERY instance with more spam sign-ins than Beehaw ever had, legitimate or not.

And you have nothing to stop bad actors from spinning up entire instances, which you then have to moderate individually, too.

You can’t defederate from every instance that gets hostile accounts because the logical thing if you’re a malicious actor is to automate signups to ALL available instances. Spam is spam is spam. You do it at scale. And you can’t shut down all signups on all instances if you want to provide the service at scale.

There is no systemic solution to malicious use. If there was, commercial social media would have deployed it to save money, at least when they were still holding to the pretense that they moderate things to meet regulations. Moderation is hard and expensive, and there are no meaningful federation-wide tools to manage it in place. I don’t even know if there can be. The idea that defederation and closing signups will be enough at scale is clearly not accurate. I don’t even think most of the big players in making federation apps would disagree with that. I think the hope is the tools will grow as the needs for them do. I’m not super sure of how well that will go, but I’m also not sure things will get big enough for that to matter at any point soon.

Yep. People around here love to attribute some magic powers to decentralization it definitely does not have. The assumption that crappy behavior is somehow localized to a specific instance is bizarre, nothing is keeping people from spamming accounts on instances with free signups. If anything, the decentralization makes it significantly harder to scale up moderation, on top of all the added costs of hosting volunteer social media servers.

That said, I’m not concerned at this point. There is nowhere near enough growth happening to make this be a problem for a long time. Masto worried about it legitimately for like twenty minutes back in some of the first few exodus incidents, before all the normies got alienated and landed on Bluesky.

Don’t get me wrong, I like it here, it feels all retro and kinda like 90s forums, but “what if it gets so popular it’s swamped with bad actors” is VERY low in my list of priorities. We have like two spammers and they’ve become local mascots. Mass malicious engagement is NOT the concern at the moment.

I’ll make a complementary argument below in a sec, but “enforcing driving traffic” seems like a feature, not a bug.

For how testy people get about crawling for copyrigted stuff for things like AI, everybody seems super chill about search engines and aggregators ripping off content at industrial scales with zero repercussions.

You’re going to pretend we’re not on that timeline right now? I see your play, time traveller.

Not what he’s talking about. He’s still lying and whining, but not saying what you (or the headline) imply he’s saying.

He’s saying that the pre-existing tariffs on out of quota dairy products are “cheating US farmers”. Which is not true. The body of the article explains this correctly and in good detail, but the headline sucks and nobody ever reads past the headline because we all have brain rot as a species.

I wonder if a good Fedi alternative to Reddit would do something like force the link to be previewed in full or opened before getting to respond to the aggregation. Or maybe all social media was a mistake and none of it should exist, I don’t know.

And let me be clear, I’m not attacking you here, this is a sytemic issue. Every human is subject to these patterns. Blame our collective wetware.

If I’m reading this right this still required a manual clickthrough (seemingly forced through a fake video player) and running an executable, right? The description is simultaneously very detailed and fuzzy on the social engineering portion.

Analysis of the redirector chain determined the attack likely originated from illegal streaming websites where users can watch pirated videos. The streaming websites embedded malvertising redirectors within movie frames to generate pay-per-view or pay-per-click revenue from malvertising platforms. These redirectors subsequently routed traffic through one or two additional malicious redirectors, ultimately leading to another website, such as a malware or tech support scam website, which then redirected to GitHub.

Not to say you don’t want an adblocker for security reasons, but still, the implication in the reporting is “have an ad pop up, get infected”, when it was more “click on the “watch PopularseriesS02e04” prompt, fail multiple times due to it being an obvious scam, get prompted to download some files, install said files, get infected”.

Did you guys get a lot of viruses in the 00s? I remember bumping into it more back in the good old disk days, where we got infected with skin-to-floppy contact, as the lord intended.

Is this a regional thing? This isn’t on by default on my accounts.

I keep having to repeat this, but the conversation does keep going on a loop: LLMs aren’t entirely useless and they’re not search engines. You shouldn’t ask it any questions you don’t already know the answer to (or have the tools to verify, at least).