WEBVTT - Part Two: Let's Look at the Facebook Papers 0:00:03.520 --> 0:00:09.559 Let's do it. Start the podcast podcast. All right, well 0:00:09.680 --> 0:00:12.119 have the let's have that be what starts the podcast. 0:00:12.160 --> 0:00:16.279 What we just said, let's start the podcast podcast. Let's 0:00:16.280 --> 0:00:19.720 start the podcast. Well, I'm Robert Evans. Yep, I'm so 0:00:19.880 --> 0:00:23.600 feeling I'm in. I never introduced myself. I'm Jamie. Who 0:00:23.640 --> 0:00:27.080 are you, Jamie Loft? Anything more we need to say? 0:00:27.120 --> 0:00:29.720 Are we done with the episode? Anderson's here? No, I 0:00:29.800 --> 0:00:34.400 think that. Yeah, yeah, Anderson is here. Uh sure, Well, 0:00:34.800 --> 0:00:39.360 you know what's happening in the world. Facebook is happening 0:00:39.360 --> 0:00:42.920 to the world, and it's not great. Jamie. It's not great. 0:00:42.960 --> 0:00:45.480 Sophie not not a fan of the Facebook. We left 0:00:45.520 --> 0:00:48.800 off having gone through some of the Facebook papers, particularly 0:00:48.880 --> 0:00:53.120 employees attacking their bosses after jan six, when it became 0:00:53.159 --> 0:00:55.800 clear that the company they were working for was completely 0:00:55.840 --> 0:01:01.680 morally indefensible. Attack they are. Yeah, they already knew. I 0:01:01.680 --> 0:01:05.200 wouldn't call it attacking either. I would call it they 0:01:05.280 --> 0:01:09.480 were pretty I mean, there's a guy like the quote there, 0:01:09.480 --> 0:01:12.280 there was the guy who's like history won't judge as kindly. Uh, 0:01:12.640 --> 0:01:16.520 the guy, and we didn't ban Trump into thus in fifteen. 0:01:16.600 --> 0:01:20.120 That's what caused the Capital riot. I mean facts or facts? 0:01:20.200 --> 0:01:23.440 Is that really attacking if you're just like, well, I 0:01:23.480 --> 0:01:28.639 think that, Yeah, I think stating facts can be an attack. Well, okay, 0:01:28.720 --> 0:01:31.759 put it on a T shirt. I mean for people 0:01:31.800 --> 0:01:35.240 like this, you know, Um yeah, I think I think 0:01:35.240 --> 0:01:37.320 stating facts can be an attack. And we ended Part 0:01:37.360 --> 0:01:40.880 one by sharing some of the blistering criticisms of Facebook 0:01:40.880 --> 0:01:43.319 employees against you know, management in the service itself. So 0:01:43.360 --> 0:01:45.800 as we start part two, it's only proper that we 0:01:45.880 --> 0:01:49.520 cover how Facebook responded to all of this internal criticism. 0:01:49.600 --> 0:01:51.920 As I stated last episode, Facebook is in the midst 0:01:51.920 --> 0:01:54.480 of a year's long drought of capable engineers and other 0:01:54.520 --> 0:01:57.080 technical employees. They are having a lot of trouble hiring 0:01:57.120 --> 0:01:58.640 all of the people that they need for all of 0:01:58.680 --> 0:02:01.680 the things they're trying to do. Um. So, one of 0:02:01.720 --> 0:02:04.000 the things is for a lot of these employees, when 0:02:04.000 --> 0:02:06.360 they say things that are deeply critical, they can't just 0:02:06.480 --> 0:02:11.160 dismiss the concerns of their employees outright because act like 0:02:12.000 --> 0:02:14.120 if they were to do that, these people would get 0:02:14.160 --> 0:02:16.400 angry and they need them, right. Facebook's not in the 0:02:16.400 --> 0:02:18.080 strongest position when it comes to the people who are 0:02:18.080 --> 0:02:20.720 good engineers. They have to walk a little bit of 0:02:20.720 --> 0:02:23.240 a tight rope. However, if they were to actually do 0:02:23.320 --> 0:02:26.320 anything about the actual meat of the concerns, it would 0:02:26.320 --> 0:02:29.280 reduce profitability and in some cases destroy Facebook as it 0:02:29.280 --> 0:02:32.200 currently exists. So they're not going to do anything, um, 0:02:32.320 --> 0:02:33.920 which is meant that they've had to get kind of 0:02:33.919 --> 0:02:36.960 creative with how they respond. So Mark and his fellow 0:02:37.160 --> 0:02:41.600 bosses pivoted and argued that the damning critique like, yeah, 0:02:41.680 --> 0:02:43.799 old Zucky z uck. So when when this all comes 0:02:43.800 --> 0:02:45.519 out and people are like, boy, it sure seems like 0:02:45.560 --> 0:02:47.400 all of your employees know that they're working for the 0:02:47.400 --> 0:02:51.960 fucking death star. Um, Zuckerberg and his like mouthpieces made 0:02:51.960 --> 0:02:54.200 a statement that like, all of these damning critiques from 0:02:54.200 --> 0:02:56.600 people inside the company were actually evidence of the very 0:02:56.680 --> 0:03:00.240 open culture inside Facebook, which encouraged workers to share their 0:03:00.240 --> 0:03:03.840 opinions with management. Um. That's exactly what a company spokesperson 0:03:03.919 --> 0:03:06.480 told The Atlantic when they asked about comments like history 0:03:06.480 --> 0:03:09.200 will not judge us kindly. The fact that they're saying 0:03:09.200 --> 0:03:11.560 will be damned by historians means that we really have 0:03:11.600 --> 0:03:16.320 a healthy office culture. Um. Hashtag dust star, proud, dust Star, 0:03:16.400 --> 0:03:19.480 proud everybody. Yeah. Yeah, it's like the fact that we 0:03:19.639 --> 0:03:22.839 removed a stigma of working for the devil, right, I mean, 0:03:23.120 --> 0:03:25.760 come on, the devil I would be proud to work 0:03:25.800 --> 0:03:28.240 for because he's done some cool stuff, Like have you 0:03:28.280 --> 0:03:32.240 ever been to Vegas? Nice town? Oh, I've been to Vegas. 0:03:32.280 --> 0:03:34.200 I've seen I saw that. I saw the Backstreet Boys 0:03:34.200 --> 0:03:37.480 in Vegas right before two of them were revealed to 0:03:37.520 --> 0:03:40.000 be in Q and On. So really caught the end 0:03:40.000 --> 0:03:43.120 of that, Locomo, Wow, I did not realize that a 0:03:43.280 --> 0:03:45.840 sizeable percentage of the Backstreet Boys had gotten into Q 0:03:46.000 --> 0:03:50.040 and On. That makes total sense. The Backstretoys, They're from Florida. 0:03:50.080 --> 0:03:53.400 They're ultimately five men from Florida. So what can you do? 0:03:53.600 --> 0:03:56.600 As the author of that article, The Atlantic article noted, 0:03:56.920 --> 0:04:00.000 this stance allows Facebook to claim transparency while ignoring the 0:04:00.080 --> 0:04:02.720 substance of the complaints and the implication of the complaints 0:04:02.760 --> 0:04:05.640 that many of Facebook's employees believe their company operates without 0:04:05.640 --> 0:04:08.640 a moral compass. All over America, people used Facebook to 0:04:08.760 --> 0:04:10.960 organize convoys to d C and to fill the buses 0:04:11.000 --> 0:04:13.960 they rented for their trips, and this was indeed done 0:04:14.240 --> 0:04:17.640 in groups like the Lebanon main truth Seekers, where Kyle 0:04:17.640 --> 0:04:20.360 fitz Simmons posted the following quote, this election was stolen, 0:04:20.400 --> 0:04:22.919 and we're being slow walked towards Chinese ownership by an 0:04:23.080 --> 0:04:25.520 establishment that is treasonous and all too willing to gaslight 0:04:25.520 --> 0:04:27.760 the public into believing the theft was somehow the will 0:04:27.800 --> 0:04:29.839 of the people. Would there be an interest locally in 0:04:29.920 --> 0:04:32.640 organizing a caravan to Washington, d C. For the electoral 0:04:32.680 --> 0:04:37.760 College vote count on January six? Um? Yeah, and Kyle 0:04:37.880 --> 0:04:40.680 recently played guilt not guilty to eight federal charges, including 0:04:40.680 --> 0:04:43.920 assault on a police officer. Mark Zuckerberg would argue that 0:04:43.960 --> 0:04:47.800 like Facebook didn't play a significant role in organizing January six, 0:04:47.880 --> 0:04:50.240 and couldn't have played a significant role in radicalizing this 0:04:50.320 --> 0:04:53.200 guy and many other people. But the reality is that 0:04:53.800 --> 0:04:56.880 for the people, like the people who managed part of 0:04:56.920 --> 0:04:59.720 what led Kyle fitz Simmons to go assault people on 0:05:00.040 --> 0:05:02.880 Nuary six, UM was the fact that he had been 0:05:02.960 --> 0:05:05.599 radicalized by a social network that for years made the 0:05:05.600 --> 0:05:09.960 conscious choice to amplify angry content and encourage anger because 0:05:09.960 --> 0:05:12.280 it kept people on the site more right, Like all 0:05:12.320 --> 0:05:14.240 of the anger that boiled up at January six that 0:05:14.320 --> 0:05:16.160 came from a number of places, but one of those 0:05:16.200 --> 0:05:20.919 places was social media. Because social media profited and specifically, 0:05:20.960 --> 0:05:25.160 Facebook knowingly profited from making people angry. That was the business, 0:05:25.360 --> 0:05:27.120 and of course it blew up in the real world. 0:05:27.279 --> 0:05:30.000 I have a question, um, just out of your own 0:05:30.040 --> 0:05:34.840 experience and observation, which is, how do you, um like, 0:05:34.880 --> 0:05:37.599 if you're doing a side by side case study of 0:05:37.640 --> 0:05:40.960 how Facebook responded to events like this versus how like 0:05:41.480 --> 0:05:47.400 YouTube slash Google responded to radicalization? Are there like significant differences? 0:05:47.400 --> 0:05:50.960 Did is there any did anyone do better or difference? 0:05:51.240 --> 0:05:56.960 Twitter has done better than probably most of them YouTube. 0:05:57.000 --> 0:05:59.159 I mean, and again I'm not saying this the Twitter 0:05:59.200 --> 0:06:01.880 has done well, um or that YouTube has done well. 0:06:02.040 --> 0:06:05.719 They've both done, particularly with coronavirus disinformation, a bit better 0:06:05.760 --> 0:06:08.440 than Facebook. Um, and they were they were better in 0:06:08.480 --> 0:06:11.599 general on not really YouTube as much, but like Twitter 0:06:11.720 --> 0:06:15.400 was definitely has taken has been the most responsible of 0:06:15.440 --> 0:06:18.799 the social networks around this stuff. Um. It did seem 0:06:18.800 --> 0:06:21.920 like for a while there the various networks were kind 0:06:21.920 --> 0:06:24.120 of like duking it out to see who could do 0:06:24.160 --> 0:06:27.080 the absolute worst and damage the lives and it's it 0:06:27.160 --> 0:06:30.640 seems like Facebook one. Yes, that I would say Facebook, 0:06:30.640 --> 0:06:34.919 but and again Twitter chose to do a lot of 0:06:34.960 --> 0:06:38.039 the same talxic things Facebook did, so did YouTube, and 0:06:38.040 --> 0:06:40.039 they did it all for profit. A number of the 0:06:40.040 --> 0:06:43.120 things we've criticized Facebook for, you can critique YouTube and Twitter. Four. 0:06:43.640 --> 0:06:46.920 I would argue Twitter certainly has done more and more 0:06:46.920 --> 0:06:51.200 effectively than Facebook, not enough that they're not being irresponsible, 0:06:51.200 --> 0:06:53.520 because I would argue that Twitter has actually been extremely 0:06:53.560 --> 0:06:57.839 irresponsible and knowingly so. Um. But I think Facebook, in 0:06:57.880 --> 0:07:01.159 my analysis, Facebook has been the worst. Although I'm not 0:07:01.200 --> 0:07:03.920 as I haven't gotten studied as much about like TikTok yet, 0:07:03.960 --> 0:07:07.800 so we'll see. But but my analysis, you've got to 0:07:07.880 --> 0:07:12.720 get on TikTok, pivot out of podcasting and into TikTok dances. Yeah, 0:07:12.800 --> 0:07:15.280 I mean, I'm it's it's not the dances that concerned 0:07:15.280 --> 0:07:19.480 me on TikTok. It's the minute long conspiracy theory videos 0:07:19.480 --> 0:07:21.920 that have convinced a number of people that the Kardashians 0:07:21.960 --> 0:07:24.200 are Armenian witches and had something to do with the 0:07:24.200 --> 0:07:27.040 collapse of the astro worlds or the deaths in the 0:07:27.080 --> 0:07:31.760 astro working. My my concern there is the dances that 0:07:31.840 --> 0:07:35.720 go over those conspiracy videos and really marry the worst 0:07:35.720 --> 0:07:40.040 of both worlds. Because I have seen dancing on there. 0:07:40.080 --> 0:07:43.800 I have seen conspiracy videos that involved dancing incredible and 0:07:43.920 --> 0:07:47.360 skincare routine. Have you ever seen a conspiracy video where 0:07:47.400 --> 0:07:50.400 someone's also doing their skincare routine? Because that is a 0:07:50.560 --> 0:07:58.280 thriving SOMETI I'm sure that's yeah. So well, I was like, 0:07:58.360 --> 0:08:03.960 that is just a thing that on many platforms. Change 0:08:04.280 --> 0:08:10.920 media companies are willfully bad at stopping radicalization because making 0:08:10.920 --> 0:08:14.280 people angry and frightened is good for all of their 0:08:14.280 --> 0:08:17.960 bottom lines, so they all knowingly participate in this. I 0:08:18.000 --> 0:08:21.680 think Facebook has been the least responsible about it, but 0:08:21.760 --> 0:08:24.920 that doesn't that shouldn't be taken as praise of anybody, 0:08:25.120 --> 0:08:28.080 like saying Twitter. Saying Twitter has done the best is 0:08:28.120 --> 0:08:32.040 saying like, well, we were all drunk driving, but John 0:08:32.160 --> 0:08:36.080 could actually walk most of a straight line before vomiting, 0:08:36.240 --> 0:08:38.840 so he was the least irresponsible of us who drunk 0:08:38.920 --> 0:08:40.800 drove that night. Just to put it in terms that 0:08:40.840 --> 0:08:43.120 I understand, I was it sounds like Twitter is the 0:08:43.120 --> 0:08:45.679 Backstreet boy. That's like, look, I don't believe in Q 0:08:45.920 --> 0:08:50.280 and on, but I see their points. That's that's kind 0:08:50.320 --> 0:08:53.960 of um the vibe I'm getting fair enough. Um So, 0:08:54.240 --> 0:08:57.000 when deciding what which posts should show up more often 0:08:57.000 --> 0:08:59.520 in the feeds of other users. Facebook's algorithm ways a 0:08:59.600 --> 0:09:02.000 number of actors. The end goal is always the same, 0:09:02.040 --> 0:09:03.960 to get the most people to spend the most time 0:09:04.000 --> 0:09:06.920 interacting with the site. For years, this was done by 0:09:06.960 --> 0:09:09.839 calculating the different reactions of post god and weighing it 0:09:09.920 --> 0:09:13.360 based on what responses people had to it. Again, for years, 0:09:13.400 --> 0:09:15.679 the reaction that carried the most weight was anger. The 0:09:15.720 --> 0:09:18.320 little snarling smiley face icon you could click under a 0:09:18.360 --> 0:09:21.200 post it was at one point being waited five times 0:09:21.200 --> 0:09:24.160 more than just like a like really like the again 0:09:24.200 --> 0:09:26.000 when I'm when I'm saying this was all intentional. They 0:09:26.000 --> 0:09:29.720 were like people who respond angrily to post that keeps 0:09:29.760 --> 0:09:31.640 them on the site more. That's they spend the most 0:09:31.679 --> 0:09:33.880 time engaging with things that make them angry. So, when 0:09:33.880 --> 0:09:38.040 it comes to determining what by which methody like how 0:09:38.280 --> 0:09:41.320 we choose to have the algorithm present people with posts, 0:09:41.760 --> 0:09:44.199 the post that are making people angriest is the post 0:09:44.240 --> 0:09:46.640 our algorithm will send to the most people. That's a 0:09:46.679 --> 0:09:50.080 conscious choice. That's a conscious Yeah, it's so funny, how 0:09:50.120 --> 0:09:52.920 I mean not funny, it's tragic and upsetting, But just 0:09:53.080 --> 0:09:55.680 how specific the Facebook audience is that it's like you 0:09:55.679 --> 0:09:57.600 would have to be the kind of person who would 0:09:57.960 --> 0:10:02.120 be like, I'd better react angry as specific as possible. 0:10:02.120 --> 0:10:04.280 In my feedback to this post, which is farm bill 0:10:04.440 --> 0:10:08.040 moms and yeah it's boomers, it's boomers, and yeah, um 0:10:08.760 --> 0:10:12.040 and yeah, they just kind of knowingly set off a 0:10:12.080 --> 0:10:14.720 bomb and a lot of people's fucking brains. Um, they're 0:10:14.760 --> 0:10:19.560 addicted to telling on themselves for Yeah, why why Facebook 0:10:19.600 --> 0:10:22.160 has something called the Integrity Department and these are the 0:10:22.160 --> 0:10:25.400 people with the unenviable task of trying to fight misinformation 0:10:25.440 --> 0:10:28.800 and radicalization on the platform. They noted in July that 0:10:28.920 --> 0:10:33.880 is so embarrassing. Yeah, just going on the first day 0:10:33.960 --> 0:10:37.000 be like, I work for the Facebook Integrity Department, Like, yeah, 0:10:37.080 --> 0:10:40.360 good fucking luck. Yeah, I work for thee in your life. 0:10:40.400 --> 0:10:42.839 My job is to go door to door and apologize 0:10:42.840 --> 0:10:45.920 to people after we bomb them. We have gift baskets 0:10:46.080 --> 0:10:50.480 for the survivors, you know, Like it's that that's the gig. Really. Um, yeah, 0:10:50.640 --> 0:10:53.360 I send edible arrangements to people who have been droned 0:10:53.400 --> 0:10:57.600 striight like, oh Jesus awful. There's was one of my 0:10:57.640 --> 0:11:01.200 favorite follows on twitters, Brook Minkowski Um, who used to 0:11:01.240 --> 0:11:03.079 work for Facebook and was like one of the people 0:11:03.080 --> 0:11:06.559 early on who was trying to warn them about disinformation 0:11:06.559 --> 0:11:09.720 and radicalization on the platform years ago and left because 0:11:09.720 --> 0:11:11.720 like it was clear they didn't actually give a shit. 0:11:12.360 --> 0:11:15.040 UM And and a lot of the Integrity department people 0:11:15.040 --> 0:11:18.600 are actually like really good people who are a little 0:11:18.600 --> 0:11:21.160 bit optimistic and kind of young and coming and like, Okay, 0:11:21.200 --> 0:11:23.280 I'll make it's my job to make this huge and 0:11:23.320 --> 0:11:27.880 important thing a lot safer. Um. And these people get 0:11:28.160 --> 0:11:31.320 chewed up and spit out very very quickly. Um. And 0:11:31.760 --> 0:11:34.600 the members of the integrity team, Um, we're kind of 0:11:34.640 --> 0:11:38.520 analyzing the impact of weighing angry content so much. And 0:11:38.640 --> 0:11:41.640 some of them noted in July the extra weight given 0:11:41.679 --> 0:11:45.240 to the anger reaction was a huge problem. They recommended 0:11:45.240 --> 0:11:47.800 the company stop weighing at extra in order to stop 0:11:47.800 --> 0:11:50.720 the spread of harmful content. Their own tests showed that 0:11:50.840 --> 0:11:53.160 dialing the weight of anger back to zero so it 0:11:53.240 --> 0:11:55.800 was no more influential than any other reaction, would stop 0:11:55.840 --> 0:11:58.960 rege inducing content from being shared and spread nearly as widely. 0:11:59.280 --> 0:12:03.040 This led to a five percent reduction and hate speech, misinformation, bullying, 0:12:03.040 --> 0:12:05.160 and posts with violent threats. And when you consider how 0:12:05.160 --> 0:12:08.000 many billions of Facebook posts there are that's a lot 0:12:08.160 --> 0:12:11.280 less nasty shit, um, some of which is going to 0:12:11.320 --> 0:12:13.440 translate into real world violence. And again this was kind 0:12:13.440 --> 0:12:15.120 of a limited study, so who knows how it would 0:12:15.120 --> 0:12:18.000 have actually affected things in the long run. So Facebook 0:12:18.080 --> 0:12:22.200 made this well less money question mark. Yeah, this actually 0:12:22.280 --> 0:12:24.520 was kind of a win for them. Um. Facebook did 0:12:24.640 --> 0:12:27.760 make this change. They pushed it out in September. UM, 0:12:27.760 --> 0:12:30.400 and the employees responsible deserve real credit. Again, there's people 0:12:30.400 --> 0:12:34.080 within Um, there's people within Facebook who did things that 0:12:34.160 --> 0:12:38.760 really actually we're good Like changing this was probably like 0:12:38.920 --> 0:12:41.920 made the world a bit healthier. That said, the fact 0:12:41.920 --> 0:12:44.839 that it had been weighted this way for years, you 0:12:45.160 --> 0:12:48.440 don't undo that just by dialing it back now. For 0:12:48.520 --> 0:12:51.520 one thing, anger has become such an aspect of the 0:12:51.600 --> 0:12:55.080 culture of Facebook that even without weighing the anger emoji, 0:12:55.160 --> 0:12:57.160 most of the content that goes viral is still stuff 0:12:57.200 --> 0:12:59.400 that makes pisces people off, because that's just become what 0:12:59.440 --> 0:13:02.319 Facebook is, because that's what they selected for for years. 0:13:02.920 --> 0:13:05.880 Like it Also, like, who knows, like if they've done 0:13:05.920 --> 0:13:08.360 this years ago, if they never waited anger more, it 0:13:08.440 --> 0:13:10.840 might be a very different platform with like a very 0:13:10.840 --> 0:13:13.400 different impact on the brains of for example, our aunts 0:13:13.400 --> 0:13:16.840 and uncles. Um I think that that's really interesting too, 0:13:16.880 --> 0:13:20.800 because that timeline lines up pretty quickly with or pretty 0:13:20.800 --> 0:13:22.800 exactly with where it feels like a lot of younger 0:13:22.840 --> 0:13:27.600 people were leaving that platform and the platforms became associated 0:13:27.679 --> 0:13:30.559 with older people. Because I feel like I don't think 0:13:30.559 --> 0:13:34.880 I was using Facebook consistently after twenties seventeen. I want 0:13:34.880 --> 0:13:37.240 to say it was maybe my last Facebook year. Yeah 0:13:37.320 --> 0:13:41.200 I stopped. I mean I I stopped visiting it super 0:13:41.200 --> 0:13:45.800 regularly a while back. Um yeah, maybe around So in 0:13:45.920 --> 0:13:49.880 April of Facebook employees came up with another recommendation. This 0:13:49.920 --> 0:13:52.920 one wouldn't be as successful as as changing you know, 0:13:53.000 --> 0:13:57.000 the reaction of the algorithm to the angry reaction spurred 0:13:57.000 --> 0:13:58.960 by the lockdown on the sudden surge of q and 0:13:58.960 --> 0:14:02.680 on Boogaloo and he lockdown groups urging real world violence. 0:14:02.920 --> 0:14:05.880 It was suggested by internal employees that the news feed 0:14:05.880 --> 0:14:09.040 algorithm de prioritize the posting of content based on the 0:14:09.080 --> 0:14:12.040 behavior of people's Facebook friends. So the basic idea is this, 0:14:12.200 --> 0:14:16.160 um a new what Facebook was doing was you would 0:14:16.240 --> 0:14:19.200 if normally like the way you'd think it would work, right, 0:14:19.280 --> 0:14:21.760 is that like your friends post something and you see 0:14:21.760 --> 0:14:23.880 that in your news feed, right, like the posts of 0:14:23.880 --> 0:14:25.960 the people that you've chosen to follow and say are 0:14:26.000 --> 0:14:28.400 your friends? Right? That's how you you would you would 0:14:28.400 --> 0:14:30.080 want it to work. Is that what worked? At one 0:14:30.080 --> 0:14:32.920 point they made a change a few years back where 0:14:33.160 --> 0:14:35.800 they started sending you things not because someone you followed 0:14:35.840 --> 0:14:38.840 had said something, but because they'd like to think um, 0:14:38.960 --> 0:14:41.800 or they'd commented like not so even commented just like 0:14:41.800 --> 0:14:43.840 liked a thing, Like if they'd reacted to a thing 0:14:44.240 --> 0:14:47.360 you would get you would get that sent to your 0:14:47.400 --> 0:14:51.240 news feed UM. And members of the Integrity Team start 0:14:51.280 --> 0:14:54.160 to recognize, like, this has some problems um in it 0:14:54.240 --> 0:14:56.160 from one thing, it results in a lot of people 0:14:56.200 --> 0:15:00.520 getting exposed to dangerous bullshit UM. So they uh, they 0:15:01.160 --> 0:15:04.080 start looking into like the impact of this and how 0:15:04.520 --> 0:15:06.880 how just sharing the kind of things your friends are 0:15:06.920 --> 0:15:09.400 reacting to influences what you see and what that does 0:15:09.440 --> 0:15:13.520 to you on Facebook. The Integrity Team experimented without changing 0:15:13.520 --> 0:15:15.840 this might work, and their early experiments found that fixing 0:15:15.880 --> 0:15:19.920 this would reduce the spread of violence, insight and content UM. 0:15:20.000 --> 0:15:23.640 For one thing, what they found is that like normally 0:15:23.800 --> 0:15:27.400 if you hadn't seen someone like a post about something 0:15:27.400 --> 0:15:29.720 that was maybe like violent or aggressive, your conspiratory, like 0:15:29.760 --> 0:15:32.160 a flat earth post, your appost urging the execution of 0:15:32.160 --> 0:15:34.680 an elected leader. If you hadn't seen anyone that you 0:15:34.760 --> 0:15:37.440 knew react to that post, even if you saw it, 0:15:37.480 --> 0:15:39.880 you wouldn't comment on it or share it. But they 0:15:39.880 --> 0:15:41.920 found that, like if you just saw that a friend 0:15:41.920 --> 0:15:43.840 had liked it, you were more likely to share it, 0:15:43.880 --> 0:15:47.360 which increases exponentially the spread of this kind of violent content. 0:15:47.680 --> 0:15:50.760 And it's it's this idea, like the whole people weren't 0:15:50.760 --> 0:15:53.240 stopped being afraid to be racist at a certain points 0:15:53.400 --> 0:15:55.360 as much as they had been earlier, and it led 0:15:55.360 --> 0:15:58.080 to this surge in real world violence. It is kind 0:15:58.080 --> 0:16:01.160 of the same thing people felt by seeing their friends 0:16:01.200 --> 0:16:03.800 react to this, they felt permission to react to it too. 0:16:03.800 --> 0:16:05.240 In a way, maybe they would have liked, well, I 0:16:05.280 --> 0:16:07.360 don't want to like, maybe I'm interested in flat earthship, 0:16:07.360 --> 0:16:08.720 but I'm just going to ignore this because like I 0:16:08.760 --> 0:16:10.720 don't want to seem like a kok that is so 0:16:10.840 --> 0:16:14.560 fucking upsetting and and fascinating in the way that it 0:16:14.600 --> 0:16:16.840 affects your mind. Is is Yeah, there there was a 0:16:16.880 --> 0:16:23.880 time where you would if you were you know, racist, misogynist, homophobic, 0:16:23.920 --> 0:16:26.440 whatever you were, but you just didn't talk about it. 0:16:26.480 --> 0:16:29.680 But then all of a sudden, there's this confirmation that like, hey, 0:16:29.680 --> 0:16:32.040 this person you know and see all the time feels 0:16:32.040 --> 0:16:34.880 the same fucking way you do. So why be quiet 0:16:34.960 --> 0:16:38.640 about it. Let's discuss. Like it's just that's so dark. 0:16:38.800 --> 0:16:41.480 It's really dark. And so the the the integrity team 0:16:41.520 --> 0:16:44.040 sees this and they're like, we should change this. Um, 0:16:44.120 --> 0:16:46.280 we should only show We shouldn't be showing people just 0:16:46.360 --> 0:16:48.680 like the reactions their friends have had to content because 0:16:48.720 --> 0:16:52.720 it seems to be bad for everybody. Um. And they 0:16:52.760 --> 0:16:54.480 do find in some of their you know, because when 0:16:54.480 --> 0:16:57.080 the experiment, they're like that, we'll take this country or 0:16:57.120 --> 0:16:59.120 this city and we'll we'll roll this change out in 0:16:59.120 --> 0:17:01.520 this limited geograph location to like try and see how 0:17:01.560 --> 0:17:03.920 it might affect its scale. And they do this and 0:17:03.960 --> 0:17:07.000 they see that like, oh, changing this significantly reduces the 0:17:07.040 --> 0:17:10.320 spread of specifically violence and citing contents. So they're like, hey, 0:17:10.359 --> 0:17:14.879 we should roll this out service wide. Zuckerberg himself steps in. 0:17:14.960 --> 0:17:19.119 According to Francis Hoggan, the whistleblower and quote rejected this 0:17:19.200 --> 0:17:21.280 intervention that could have reduced the risk of violence in 0:17:21.320 --> 0:17:25.920 the election from The Atlantic Quote, an internal message characterizing 0:17:25.960 --> 0:17:28.639 Zuckerberg's reasoning says he wanted to avoid new features that 0:17:28.680 --> 0:17:31.960 would get in the way of meaningful social interactions. But 0:17:32.000 --> 0:17:35.600 according to Facebook's definition, its employees say engagement is considered 0:17:35.640 --> 0:17:38.240 meaningful even if it entails bullying, hate speech, and re 0:17:38.320 --> 0:17:42.080 shares of harmful content. The episode, like Facebook's response to 0:17:42.080 --> 0:17:45.360 the incitement that proliferated between the election in January six, 0:17:45.520 --> 0:17:48.800 reveals a fundamental problem with the platform. Facebook's mega scale 0:17:48.800 --> 0:17:51.080 allows the company to influence the speech and thought patterns 0:17:51.080 --> 0:17:53.200 of billions of people. What the world is seeing now 0:17:53.200 --> 0:17:55.560 through the window provided by reams of internal documents is 0:17:55.600 --> 0:17:58.439 that Facebook catalogs and studies the harm it inflicts on people, 0:17:58.680 --> 0:18:01.800 and then it keeps harming anyway. See that that is 0:18:01.840 --> 0:18:06.200 so that's always um so interesting to hear. And by interesting, 0:18:06.240 --> 0:18:12.560 I mean, you know, psychologically harmful, but uh, because it's like, yes, 0:18:12.680 --> 0:18:14.760 that is a fundamental flaw of the platform, but that's 0:18:14.800 --> 0:18:17.639 also very entrenched into like what the DNA of the 0:18:17.680 --> 0:18:23.480 platform always was, which was based on harshly judging other people. 0:18:23.600 --> 0:18:28.000 Like that's why Mark Zuckerberg created Facebook, was to harshly 0:18:28.080 --> 0:18:33.439 judge women in his community. So it's like, I I 0:18:33.520 --> 0:18:36.600 know that it is you know, on a bajillion scale 0:18:36.600 --> 0:18:39.440 at this point, but I'm always kind of stunned at 0:18:40.080 --> 0:18:43.280 how people are like, oh, it's so weird that this 0:18:43.359 --> 0:18:45.679 went you know, the way that it did. It's like, well, 0:18:45.920 --> 0:18:50.239 to an extent um, it was always like that, and 0:18:50.280 --> 0:18:52.600 maybe it was like cosplaying as not being like that, 0:18:52.680 --> 0:18:55.640 and for certain people there were eras in Facebook where 0:18:55.640 --> 0:18:58.399 your user experience wouldn't be like that. But it you know, 0:18:58.680 --> 0:19:00.960 this goes back almost twenty is at this point of 0:19:02.480 --> 0:19:06.840 this being in the DNA of this um this ship show. Yeah, 0:19:06.960 --> 0:19:10.440 and it's it's really bleak. Um. It's just really bleak. 0:19:10.480 --> 0:19:13.760 And it also goes to show like the It was 0:19:13.800 --> 0:19:16.040 one of the things Zuckerberg will say repeatedly when he 0:19:16.440 --> 0:19:18.640 talks about when he when he does admit like, yes, 0:19:18.640 --> 0:19:20.960 there are problems and there have been like negatives associated 0:19:20.960 --> 0:19:23.280 with the site, and we're we're aware of that. They're humbling, 0:19:23.320 --> 0:19:25.399 but like, you know, you also have to include all 0:19:25.440 --> 0:19:27.040 the good that we're doing, all of the meaning and 0:19:27.040 --> 0:19:28.639 the way he always phrases this is like all of 0:19:28.640 --> 0:19:31.480 the meaningful social interactions that wouldn't have happened otherwise. And 0:19:31.480 --> 0:19:34.160 then you realize when every time he says that meaningful 0:19:34.200 --> 0:19:37.439 social interactions that have taken place on Facebook, when he 0:19:37.480 --> 0:19:41.400 says he's including as these internal blocuments, he includes bullying 0:19:41.440 --> 0:19:43.600 and people like making death threats and like talking about 0:19:43.600 --> 0:19:46.560 their desire to murder people like, that's a meaningful interaction. 0:19:46.600 --> 0:19:49.399 People getting angry and trying to inside violence together is 0:19:49.440 --> 0:19:53.400 a meaningful social interaction, which I guess yet I mean 0:19:53.560 --> 0:19:57.320 not meaningless that it that has meaning. Plan meetings were 0:19:57.520 --> 0:20:01.359 meaningful social interactions, you know, Um, you gotta give the 0:20:01.440 --> 0:20:07.399 KKK that. Uh, the Nuremberg rally was meaning interacting. The 0:20:07.480 --> 0:20:10.639 last meaningful interaction I had on Twitter led to like 0:20:11.040 --> 0:20:14.160 a rebound. I was dating coming to my grandma's funeral, 0:20:14.280 --> 0:20:21.920 blackout drunk, so you know, just man, it's been too 0:20:21.960 --> 0:20:24.640 long since I've shown up at a funeral just too 0:20:24.720 --> 0:20:28.960 drunk to stand. It was still a one of my 0:20:29.000 --> 0:20:31.639 favorite mag with my family to this day. They're like, 0:20:31.680 --> 0:20:33.840 who is this guy? And I'm like, I don't really know. 0:20:35.280 --> 0:20:39.639 He's drunk. He came on the megabus. He did the 0:20:39.680 --> 0:20:46.080 megabus getting drunk from a camel back on the megabus. Yeah, 0:20:46.119 --> 0:20:48.040 that would be when I used to do a lot 0:20:48.080 --> 0:20:50.320 of bus trips, like when I was traveling and stuff. 0:20:50.320 --> 0:20:52.040 That would be one of the tactics as you feel 0:20:52.080 --> 0:20:56.919 like a thermist or a camel back with like cranberry juice, 0:20:57.040 --> 0:21:01.520 six liquor and and just oh, I mean get that. 0:21:02.200 --> 0:21:05.680 It's I'm not above getting funked up on a mega 0:21:06.600 --> 0:21:09.119 you know, on your way to my grandma's funeral. That was. 0:21:09.440 --> 0:21:11.919 That was a move. Me and my friends got like 0:21:12.119 --> 0:21:15.120 wasted in San Francisco one day, just like going shopping 0:21:15.200 --> 0:21:17.359 in broad daylight with a camel back, where we would 0:21:17.560 --> 0:21:20.479 we would get a bottle of orange flavored Trader Joe's 0:21:20.600 --> 0:21:23.520 Patron tequila, and we would get a half dozen lime 0:21:24.000 --> 0:21:26.720 popsicles and you just throw the popsicles in with the 0:21:26.760 --> 0:21:29.040