I have actually worked in this area. I like a lot of Yishan's other writing but I find this thread mostly a jumbled mess without much insight. Here are a couple assorted points:
>In fact, once again, I challenge you to think about it this way: could you make your content moderation decisions even if you didnʻt understand the language they were being spoken in?
I'm not sure what the big point is here but there are a couple parts to how this works in the real world:
1) Some types of content removal do not need you to understand the language: visual content (images/videos), legal takedowns (DMCA).
2) Big social platforms contract with people around the world in order to get coverage of various popular languages.
3) You can use Google Translate (or other machine translation) to review content in some languages that nobody working in content moderation understands.
But some content that violates the site's policies can easily slip through the cracks if it's in the right less-spoken language. That's just a cost of doing business. The fact that the language is less popular will limit the potential harm but it's certainly not perfect.
>Hereʻs the answer everyone knows: there IS no principled reason for banning spam. We ban spam for purely outcome-based reasons:
>
>It affects the quality of experience for users we care about, and users having a good time on the platform makes it successful.
Well, that's the same principle that underlies all content moderation: "allowing this content is more harmful to the platform than banning it". You can go into all the different reasons why it might be harmful but that's the basic idea and it's not unprincipled at all. And not all spam is banned from all platforms--it could just have its distribution killed or even be left totally alone, depending on the specific cost/benefit analysis at play.
You can apply the same reasoning to every other moderation decision or policy.
The main thrust of the thread seems to be that content moderation is broadly intended to ban negative behavior (abusive language and so on) rather than to censor particular political topics. To that I say, yeah, of course.
FWIW I do think that the big platforms have taken a totally wrong turn in the last few years by expanding into trying to fight "disinformation" and that's led to some specific policies that are easily seen as political (eg policies about election fraud claims or covid denialism). If we're just talking about staying out of this business then sure, give it a go. High-level blabbering about "muh censorship!!!" without discussion of specific policies, is what you get from people like Musk or Sacks, though, and that's best met with an eye roll.
Well, that's because Quora has distribution mechanics like the home feed and email digests that can amplify content by popular contributors. It's not like Reddit or HN or StackExchange where answers/comments compete directly against each other on the question page. I don't think the situation is analogous.
>An actual hiring freeze sends a very strong negative signal to employees and investors. To employees it says "the ship is sinking." Layoffs are right around the corner and that is absolute poison to morale.
I don't agree with this analysis. I'd be shocked if Google did any layoffs (though it's always possible) instead of just freezing hiring and letting natural attrition happen for a while. As for the signal to employees, there are so many employees already that I imagine many would be agree that the company is too big. It doesn't mean that the ship is sinking, just that they don't need an infinite number of employees. And Google's products are so dominant that you can't really believe the business is failing.
Now, FB on the other hand... I'd feel a lot more antsy, yeah. Their business doesn't feel nearly so stable.
It's certainly stable for now but it doesn't feel as stable as Google's, at least to me. The main FB product is in a slow decline and IG, though doing well, is under constant threat from the latest social media app du jour. The whole VR push could work out but right now it feels like a long shot and if I were working there I wouldn't be happy betting the farm on it. I mean the very fact that they are anticipating such a drastic pivot to their core business in the future should surely be a cause for some concern.
That's one of the things, though. If you are in a dire situation, like companies nationalized during wartime, you can't work or function. Certain things like advertising your services are clearly out the window.
In WWII, some companies who were turned to serving the war machine continued to advertise the things they couldn't make or sell, anyway. When things came back those folks were on top of the world. If the big FAANG companies can ride out business cycles without acting like they've become vulnerable, they're likely to further cement their positions.
It definitely is a long term bet that could end up costing a lot to Meta, but on the other hand if that bet materializes... At the very least we're getting really cool VR development!
Right, he doesn't actually dominate Geoguessr as a game, there are hundreds or maybe even thousands of people who could beat him in the competitive mode or in eg some no moving competition. Personally I'm in the top few hundred in the world and I do think I could beat him handily--I just have memorized enough bullshit about utility poles and the Google car in different countries and so on that he hasn't.
But he does dominate Geoguessr social media content (along with rainbolt who has really blown up on TikTok). And I think Tom is pretty widely liked in the community. He's honest about his skills, his videos are fun, and he brought a bunch of today's top players into the game in the first place.
>Something to keep in mind is that they aren't playing random points on the map, they're playing fixed maps with say 20,000 locations in them. They play them to the point that they recognize many of them.
This is only partly true. It's true of ADW, the most popular community map, though it has over 50k locations. But the top players have played this map so much that some people do have some locations memorized (which is insane). Then there are a few countries or territories where the Street View coverage is so limited that you can basically memorize it, like Madagascar and Greenland.
There are enough locations & maps out there that the game isn't really about memorizing exact locations, though. But if you play a lot you do get good at "vibes" which is essentially pattern matching against other times you've played.
>In fact, once again, I challenge you to think about it this way: could you make your content moderation decisions even if you didnʻt understand the language they were being spoken in?
I'm not sure what the big point is here but there are a couple parts to how this works in the real world:
1) Some types of content removal do not need you to understand the language: visual content (images/videos), legal takedowns (DMCA).
2) Big social platforms contract with people around the world in order to get coverage of various popular languages.
3) You can use Google Translate (or other machine translation) to review content in some languages that nobody working in content moderation understands.
But some content that violates the site's policies can easily slip through the cracks if it's in the right less-spoken language. That's just a cost of doing business. The fact that the language is less popular will limit the potential harm but it's certainly not perfect.
>Hereʻs the answer everyone knows: there IS no principled reason for banning spam. We ban spam for purely outcome-based reasons: > >It affects the quality of experience for users we care about, and users having a good time on the platform makes it successful.
Well, that's the same principle that underlies all content moderation: "allowing this content is more harmful to the platform than banning it". You can go into all the different reasons why it might be harmful but that's the basic idea and it's not unprincipled at all. And not all spam is banned from all platforms--it could just have its distribution killed or even be left totally alone, depending on the specific cost/benefit analysis at play.
You can apply the same reasoning to every other moderation decision or policy.
The main thrust of the thread seems to be that content moderation is broadly intended to ban negative behavior (abusive language and so on) rather than to censor particular political topics. To that I say, yeah, of course.
FWIW I do think that the big platforms have taken a totally wrong turn in the last few years by expanding into trying to fight "disinformation" and that's led to some specific policies that are easily seen as political (eg policies about election fraud claims or covid denialism). If we're just talking about staying out of this business then sure, give it a go. High-level blabbering about "muh censorship!!!" without discussion of specific policies, is what you get from people like Musk or Sacks, though, and that's best met with an eye roll.