Elon: Let Me Help You Speed Run the Content Moderation Learning Curve • Hacker News Board

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#321

Post by tzs » Sat, Dec 03, 2022, 1:43 AM UTC

bjt2n3904 wrote: post #137

Yeah, this guide is pretty ridiculous. So many of these "problems" came from bad legislation and a culture that enabled it. "Hi boss, we're getting millions of DMCA takedown requests." The solution? Fix the brokenness of the DMCA. "Hey boss, we're getting a lot of notifications about misleading info!" Ok, tell the users to not believe everything they read on the Internet. "Hey boss, we're getting lots of reports abou…

Dealing with a large number of DMCA takedown requests should not be much of a problem. They just need to:

1. Provide a form for copyright owners to use to request a takedown of material they claim is infringing. When the form is submitted verify that all the required fields have been filled in.

2. Take down the content claimed by the form submitted, and notify the user who posted the content. The provider does not need to try to figure out if the material claimed really infringes the claimant's copyright.

3. Provide a form for the user to submit a counterclaim. If the user does not submit a counterclaim, there is nothing more to do.

4. If the user submits a counterclaim, restore the taken down content, and notify the party that requested the takedown. The provider does not have to try to evaluate the merits of the counterclaim, such as whether or not the use was fair use.

If the claimant still wants the material taken down they have to take it up with the user, such as by suing them. The provider, having followed the DMCA procedure given above, is protected by the DMCA safe harbor.

It only gets hard when a site wants to use some system of their own, like YouTube does, instead of DMCA. Due to the nature of user created content on YouTube and how YouTube is monetized it is better for them if something that infringes gets licensed and stays up, so they have deals with major content distributors to encourage that.

For a site like Twitter, where the user content is small text posts instead of longer videos that can have ads within them, I don't think they need more than the DMCA procedure.

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#322

Post by slibhb » Sat, Dec 03, 2022, 1:47 AM UTC

tstrimple wrote: post #203

slibhb wrote: post #69

People want an algorithm for how to moderate speech. But there's no such thing. You just need people with good judgement and it's hard to define what good judgement is. HN has it. Previous twitter did not have good judgement. It had a left-bias and banned people for saying things like "there are two genders" and "men aren't women tho". It also banned various accounts for jokes (prominently Babylon Bee, a right-wing T…

> It had a left-bias I don't know what it will take for this lie to die. It has absolutely zero supporting evidence. The data suggests just the opposite. https://www.theguardian.com/media/2021/feb/01/facebook-youtu... > "There is no evidence to support the claim that the major social media companies are suppressing, censoring or otherwise discriminating against conservatives on their platforms,” Barrett said. “In fac…

"Left-bias" is putting it mildly. It's becoming increasingly clear that twitter censored certain people/content at the behest of the Biden administration.

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#323

Post by DFHippie » Sat, Dec 03, 2022, 1:52 AM UTC

encryptluks2 wrote: post #211

themitigating wrote: post #96

Earlier quoted context omitted.

I'm referring to pre musk ban. The exact tweet was “I’m a bit sleepy tonight but when I wake up I’m going death con 3 On JEWISH PEOPLE,” Someone was proposing incitement to violence has a threahold etc. This to me is clearly incitement to violence. Some have claimed he didn't know what "death con" (defcon) means to which I say what could he have possibly meant. It's on Jewish People.

Defcon levels are about preparing for defense not attack. I think your bias is showing.

When the government speaks of defcon levels, they don't speak of going "defcon X on somebody". This template is borrowed from expression like "going medieval on your ass", which is clearly not about preparing for defense. What would it mean to prepare for defense on somebody?

When you combine that with him using "death con" instead of "defcon" and the tenor of his comments both before and since (posting a swastika?) it is a little hard to imagine he meant he was preparing to defend himself from an attack rather than going on attack. "I'm a little sleepy, but after I wake up I'm going to prepare for an attach By JEWISH PEOPLE." Is this seriously what you think he was saying?

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#324

Post by DFHippie » Sat, Dec 03, 2022, 2:07 AM UTC

shadowgovt wrote: post #295

squarefoot wrote: post #144

Earlier quoted context omitted.

> “But they did good things too,” Ye said. That is the same exact boilerplate response we get over here by extreme right wingers (assuming non extreme right wingers do exist, which over here is not the case) when talking about the atrocities committed by the Mussolini regime: "he did good things too". Yeah sure, like conscience is something that can be recharged like a cellphone battery: "Ok, I've helped 99 old ladie…

IIUC, this example is specifically why Utilitarianism breaks down as a consistent philosophy.

Utilitarianism doesn't say that you can store up moral chits and then cash them in to do some evil. It says an act should be judged by its net consequences. Basically, it says you can do some evil if, of the choice available to you in some context, it results in the most good (or the least evil, if no choice results in good). The evil in other words is inextricable from the good. It isn't just a lark you can engage in after you banked some good. Jabbing someone with a needle to administer a vaccine is allowed. Jabbing them with a needle after you gave them a cookie is not.

So no, whatever you think of utilitarianism, this is not its problem.

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#325

Post by robofanatic » Sat, Dec 03, 2022, 2:33 AM UTC

chasd00 wrote: post #246

robofanatic wrote: post #163

Earlier quoted context omitted.

I can't see your email and you can't see mine, but I may see your tweets even though I don't know who you are.

> but I may see your tweets even though I don't know who you are. That sounds like a bug not a feature. If Twitter collapses i'll dance in the streets.

Twitter is a virtual public square, so I'd say thats a feature.

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#326

Post by arvigeus » Sat, Dec 03, 2022, 2:44 AM UTC

I had some ideas about how one can run a free speech social media without actual or shadow banning and without letting “problematic” accounts disturbing others. I think my ideas completely died around level 13. I mean I can make them work to fit the extra requirements, but it will be just a game of whack-a-mole where some other powerful entity dictates the rules for you. Welcome to hell…

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#327

Post by willis936 » Sat, Dec 03, 2022, 2:45 AM UTC

anigbrowl wrote: post #291

willis936 wrote: post #250

Earlier quoted context omitted.

> There's always a limit to speech, and it's not easy to find It is. Mid 2000s 4chan and more recently 8kun found it immediately. When you remove limiters people slap the rail immediately. What value does it add? No one is learning anything when posting anonymously without moderation, they're just lulling themselves into an easily manipulated state.

It'd be good for John Gabriel's Greater Internet Fuckwad theory to make a comeback https://www.penny-arcade.com/comic/2004/03/19/green-blackboa...

I remember Splinter Cell multiplayer. Great map design and assymetrical loadout balance. I can't think of another game that even attempted to balance 1st person and 3rd person gameplay as hard requirements for opposing teams.

My favorite memory is playing carnival music through my mic while spinning a grabbed merc around. That and everything about aquarium.

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#328

Post by guiambros » Sat, Dec 03, 2022, 2:52 AM UTC

cma wrote: post #237

nightski wrote: post #151

Earlier quoted context omitted.

People aren't as stupid as HN wants to think they are. Twitter will be judged based on how it actually evolves, not Musk's brand. Time will tell. Let's not pretend the site wasn't an absolute disaster before the acquisition though. If I clicked on even a single trending topic about a game or news item I'd often be presented with unfiltered hardcore porn. Then there would be topics of innocent things like landscapes t…

> Let's not pretend the site wasn't an absolute disaster before the acquisition though. If I clicked on even a single trending topic about a game or news item I'd often be presented with unfiltered hardcore porn. Settings | Privacy and Safety | Display media that may contain sensitive content Did you enable that maybe?

Anecdata: mine has always been enabled, and I've been a user since 2007, yet I never encountered any kind of porn (except from some random replies or follow-spams from obvious bikini girls).

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#329

Post by SECProto » Sat, Dec 03, 2022, 4:15 AM UTC

zimpenfish wrote: post #297

sidibe wrote: post #275

Earlier quoted context omitted.

As time goes on, it turns out that pre-Elon Twitter is becoming more and more of a disaster. Now it turns out it was full of porn, and especially child porn which previous leaders didn't care about. Also instead of a company struggling to be consistently profitable it was actually a complete dumpster fire that was doomed to fail in the near future if not for the timely intervention

> Now it turns out it was full of porn, Well, it never had a ban on porn as such, did it? Although I'd venture "full of porn" is an exaggeration. Do you have numbers? > and especially child porn which previous leaders didn't care about. Got a source for that? As best I know, they weren't considered a problem by the authorities. Unlike now where they have one (1) moderator (thanks to Musk's firings) for the entirety o…

The entire comment you responded to was meant to be read with a /s tag

Re: Elon: Let Me Help You Speed Run the Content Moderation Learning Curve

#330

Post by martindbp » Sat, Dec 03, 2022, 6:50 AM UTC

themitigating wrote: post #138

martindbp wrote: post #130

Earlier quoted context omitted.

This is unfortunately exactly what will happen. What is scary to me is how censorship has become completely accepted by the new generation, including software engineers here in their 20s. No true hacker (this is Hacker News after all) would be OK with corporation or government censorship. Slippery slope is real.

In the past and in some places today profane and obscene material was/is censored and no one bats an eye. Books are being removed from libraries and that has happened apparently for decades. This type of censorship has been completely accepted by all past generations.

We went from censoring things that everyone thinks is bad (CP, death threats), to now censoring opinions which 50% of Americans share, and probably 90% of the world in just a few years.