1 00:00:00,000 --> 00:00:05,430 place to do it and also simultaneously 2 00:00:02,460 --> 00:00:07,950 next door in the lamar room at midnight 3 00:00:05,430 --> 00:00:10,678 is a screening of dark web produced 4 00:00:07,950 --> 00:00:13,320 directed by alex winter about the Silk 5 00:00:10,679 --> 00:00:16,740 Road case but right now i'm going to 6 00:00:13,320 --> 00:00:19,289 introduce alyssa chevensky AKA lady boss 7 00:00:16,740 --> 00:00:21,119 who was going to be a talk doing a talk 8 00:00:19,289 --> 00:00:24,029 about censorship social media and the 9 00:00:21,119 --> 00:00:26,250 presidential election very topical so 10 00:00:24,029 --> 00:00:30,570 take it away lady boss 11 00:00:26,250 --> 00:00:32,549 thank you thank you so much and thank 12 00:00:30,570 --> 00:00:34,950 you all for coming 13 00:00:32,549 --> 00:00:37,709 I'm really glad that were discussing 14 00:00:34,950 --> 00:00:41,730 this this is such an important topic 15 00:00:37,710 --> 00:00:44,160 increasingly so with each election so 16 00:00:41,730 --> 00:00:47,430 I'm really glad for us to talk about it 17 00:00:44,160 --> 00:00:50,669 here at Hope where we can be uh how do I 18 00:00:47,430 --> 00:00:53,879 say we can be very honest i posted a 19 00:00:50,670 --> 00:00:57,120 tweet that I might destroy like all my 20 00:00:53,879 --> 00:00:59,640 future likely job prospects in this one 21 00:00:57,120 --> 00:01:02,190 talk so I'm glad that you're here to be 22 00:00:59,640 --> 00:01:03,989 here with me for this as I complain 23 00:01:02,190 --> 00:01:05,820 about pretty much all the major 24 00:01:03,989 --> 00:01:09,929 political and technological 25 00:01:05,820 --> 00:01:11,548 organizations any bias in this talk is 26 00:01:09,930 --> 00:01:15,210 brought to you by me 27 00:01:11,549 --> 00:01:18,090 that's my hashtag and my handle on 28 00:01:15,210 --> 00:01:21,419 Twitter some disclosures I was formerly 29 00:01:18,090 --> 00:01:24,780 on Rand Paul's tech Advisory Council I'm 30 00:01:21,420 --> 00:01:27,990 talking to folks who ya Dona folks who 31 00:01:24,780 --> 00:01:30,299 worked for both Clinton and Trump I'm 32 00:01:27,990 --> 00:01:32,009 very strongly pro-clinton but just you 33 00:01:30,299 --> 00:01:35,460 would not be able to tell this from this 34 00:01:32,009 --> 00:01:36,780 talk because I'm pretty upset about the 35 00:01:35,460 --> 00:01:40,798 bias in her favor 36 00:01:36,780 --> 00:01:44,310 we accept political endorsements by the 37 00:01:40,799 --> 00:01:46,619 press we're used to that it's fine 38 00:01:44,310 --> 00:01:49,380 or if it's not fine we're certainly used 39 00:01:46,619 --> 00:01:53,159 to it and one thing that is really nice 40 00:01:49,380 --> 00:01:55,380 is they disclose their bias the post 41 00:01:53,159 --> 00:01:58,079 officially endorses Donald Trump they 42 00:01:55,380 --> 00:02:01,140 declare and the Daily News explains why 43 00:01:58,079 --> 00:02:05,758 they choose Hillary hill over burning 44 00:02:01,140 --> 00:02:08,489 and we see opposites that reflect this 45 00:02:05,759 --> 00:02:11,280 there's no pretense of neutrality 46 00:02:08,489 --> 00:02:12,920 there's the understanding that there's 47 00:02:11,280 --> 00:02:17,120 editorship and 48 00:02:12,920 --> 00:02:18,290 duration but social media is supposed to 49 00:02:17,120 --> 00:02:21,230 be neutral 50 00:02:18,290 --> 00:02:26,450 it's this beautiful fantasy that we've 51 00:02:21,230 --> 00:02:28,549 created our Facebook pages and indeed of 52 00:02:26,450 --> 00:02:30,920 the company's support this 53 00:02:28,550 --> 00:02:33,920 facebook has officially denied that it 54 00:02:30,920 --> 00:02:36,619 would ever try to influence the election 55 00:02:33,920 --> 00:02:38,299 they haven't denied that they ever would 56 00:02:36,620 --> 00:02:40,459 influence the election 57 00:02:38,300 --> 00:02:45,950 they've denied that they would try not 58 00:02:40,459 --> 00:02:47,630 that they would succeed so tell us more 59 00:02:45,950 --> 00:02:51,048 about this neutrality Mark Zuckerberg 60 00:02:47,630 --> 00:02:53,450 well we have guidelines they have 61 00:02:51,049 --> 00:02:55,580 guidelines so it's fine we have rigorous 62 00:02:53,450 --> 00:02:58,160 guidelines that do not permit the 63 00:02:55,580 --> 00:03:01,190 prioritisation of one viewpoint over 64 00:02:58,160 --> 00:03:03,650 another for the suppression of political 65 00:03:01,190 --> 00:03:06,380 perspectives 66 00:03:03,650 --> 00:03:10,850 tell me more well we have the trending 67 00:03:06,380 --> 00:03:13,670 section as you can see Mark Cuban is 68 00:03:10,850 --> 00:03:15,410 criticizing Donald Trump the nominee 69 00:03:13,670 --> 00:03:18,500 says he won't accept Ted Cruz's 70 00:03:15,410 --> 00:03:20,540 endorsement you know it's all about 71 00:03:18,500 --> 00:03:24,230 what's going on in the Republican Party 72 00:03:20,540 --> 00:03:27,200 Nazi salute some of these don't seem 73 00:03:24,230 --> 00:03:28,548 like the things that are most popular 74 00:03:27,200 --> 00:03:31,369 but perhaps the things that are most 75 00:03:28,549 --> 00:03:33,500 interesting to the team of curators so 76 00:03:31,370 --> 00:03:36,079 how did they come up with this list of 77 00:03:33,500 --> 00:03:37,640 what's trending well Facebook has an 78 00:03:36,079 --> 00:03:40,400 answer for this in their handy help 79 00:03:37,640 --> 00:03:43,670 section trending shows your topics that 80 00:03:40,400 --> 00:03:45,200 have recently become popular the topics 81 00:03:43,670 --> 00:03:48,500 you've seen are based on factors 82 00:03:45,200 --> 00:03:51,260 including engagement timeliness pages 83 00:03:48,500 --> 00:03:52,310 you've liked and your location if only 84 00:03:51,260 --> 00:03:55,040 that were true 85 00:03:52,310 --> 00:03:58,010 if that were true then perhaps we really 86 00:03:55,040 --> 00:04:03,260 would have a neutral ish facebook 87 00:03:58,010 --> 00:04:06,798 instead what we have are manually 88 00:04:03,260 --> 00:04:08,480 curated trends can i get a show of hands 89 00:04:06,799 --> 00:04:10,910 of folks in this room who's kind of 90 00:04:08,480 --> 00:04:13,040 familiar with the story of how Facebook 91 00:04:10,910 --> 00:04:15,799 was suppressing conservative news or 92 00:04:13,040 --> 00:04:17,060 allegedly suppressing conservative news 93 00:04:15,799 --> 00:04:19,608 ok so that's some of you but not 94 00:04:17,060 --> 00:04:23,840 everyone so to bring you all up to speed 95 00:04:19,608 --> 00:04:26,330 on this trending section that you see 96 00:04:23,840 --> 00:04:29,539 here that's actually 97 00:04:26,330 --> 00:04:32,270 not algorithmic that is curated by 98 00:04:29,539 --> 00:04:34,068 people people like you and me who are 99 00:04:32,270 --> 00:04:38,030 not employees but contractors of 100 00:04:34,069 --> 00:04:39,949 facebook one or more of them complain to 101 00:04:38,030 --> 00:04:42,438 news organization about their work 102 00:04:39,949 --> 00:04:44,569 conditions and a subset of that 103 00:04:42,439 --> 00:04:47,960 complaint was that conservative news was 104 00:04:44,569 --> 00:04:51,710 being suppressed that seems possible 105 00:04:47,960 --> 00:04:55,549 whether or not it's actually true the 106 00:04:51,710 --> 00:04:59,239 fact does remain that this is human 107 00:04:55,550 --> 00:05:00,860 curated there is inherent bias it is not 108 00:04:59,240 --> 00:05:03,440 what's most popular it's what the 109 00:05:00,860 --> 00:05:05,330 curation team deems to be worth viewing 110 00:05:03,440 --> 00:05:07,639 which of course inherently is biased 111 00:05:05,330 --> 00:05:10,669 this is what they're too looks like the 112 00:05:07,639 --> 00:05:13,129 review tool for deciding what goes up or 113 00:05:10,669 --> 00:05:14,719 down and there's a 28-page document 114 00:05:13,129 --> 00:05:16,729 that's been released its an internal 115 00:05:14,719 --> 00:05:18,800 facebook document and it talks about all 116 00:05:16,729 --> 00:05:23,120 of their policies and procedures and 117 00:05:18,800 --> 00:05:25,789 blacklisting of this topic or that topic 118 00:05:23,120 --> 00:05:28,969 gizmodo headline we routinely suppressed 119 00:05:25,789 --> 00:05:31,068 conservative news depending on who was 120 00:05:28,969 --> 00:05:34,430 on shift things would be blacklisted or 121 00:05:31,069 --> 00:05:36,110 trending says a former curator that's 122 00:05:34,430 --> 00:05:37,639 the danger with curation right like 123 00:05:36,110 --> 00:05:41,719 there are humans doing it how can it 124 00:05:37,639 --> 00:05:44,839 possibly be free of bias the company 125 00:05:41,719 --> 00:05:48,500 does state that it's neutral but the 126 00:05:44,839 --> 00:05:50,449 leadership isn't neutral here's one 127 00:05:48,500 --> 00:05:53,719 headline Mark Zuckerberg throw shade at 128 00:05:50,449 --> 00:05:56,449 Donald Trump during his speech he did 129 00:05:53,719 --> 00:06:00,020 insight trumped by name but it is it's 130 00:05:56,449 --> 00:06:01,940 clear his affiliations lie and it's 131 00:06:00,020 --> 00:06:05,240 clear where the affiliations of the 132 00:06:01,940 --> 00:06:07,879 employees like on the hey they went so 133 00:06:05,240 --> 00:06:11,060 far as to say in an internal meeting can 134 00:06:07,879 --> 00:06:13,940 we prevent president Trump and they also 135 00:06:11,060 --> 00:06:17,300 asked are we obligated to which is 136 00:06:13,940 --> 00:06:21,610 interesting if you could make sure all 137 00:06:17,300 --> 00:06:25,870 phones are off please 138 00:06:21,610 --> 00:06:28,150 it's dark we could have a dance party we 139 00:06:25,870 --> 00:06:30,819 can do whatever we want i have this 140 00:06:28,150 --> 00:06:33,310 stage I'm biased 141 00:06:30,819 --> 00:06:35,199 I'll do what I want huh with your 142 00:06:33,310 --> 00:06:37,990 consent the right this is hope it should 143 00:06:35,199 --> 00:06:41,770 be it should be all of us together one 144 00:06:37,990 --> 00:06:45,280 ideally so this was very interesting to 145 00:06:41,770 --> 00:06:47,530 me that the facebook employees asked not 146 00:06:45,280 --> 00:06:51,188 um know what their ethical 147 00:06:47,530 --> 00:06:53,198 responsibilities were as a company or is 148 00:06:51,189 --> 00:06:55,479 a supposedly neutral company but they 149 00:06:53,199 --> 00:06:57,909 asked if they were obligated to prevent 150 00:06:55,479 --> 00:07:02,289 one candidate from becoming successful 151 00:06:57,909 --> 00:07:04,479 and Mark Zuckerberg said no no we are 152 00:07:02,289 --> 00:07:09,400 not going to officially prevent a 153 00:07:04,479 --> 00:07:10,810 president Trump but still how much how 154 00:07:09,400 --> 00:07:13,270 much is the company controlling the 155 00:07:10,810 --> 00:07:16,090 employees and how much are they able to 156 00:07:13,270 --> 00:07:19,870 influence in the ways that they feel are 157 00:07:16,090 --> 00:07:22,448 ethical one thing we can say is that the 158 00:07:19,870 --> 00:07:24,550 company is officially neutral and the 159 00:07:22,449 --> 00:07:28,270 people inside of it are all over the 160 00:07:24,550 --> 00:07:32,770 place board member Peter teal is now 161 00:07:28,270 --> 00:07:36,039 famously a delegate for Trump and spoke 162 00:07:32,770 --> 00:07:38,289 at the RNC so one thing we can rely on 163 00:07:36,039 --> 00:07:41,259 within many of these companies is even 164 00:07:38,289 --> 00:07:43,089 the ones who are leaning heavily in one 165 00:07:41,259 --> 00:07:48,190 direction or another it's it's still a 166 00:07:43,089 --> 00:07:50,139 mixed bag of people it would be legal if 167 00:07:48,190 --> 00:07:52,089 Facebook wanted to impact the elections 168 00:07:50,139 --> 00:07:55,389 the same way it's legal for the New York 169 00:07:52,089 --> 00:07:57,250 Times or The Washington Post Facebook 170 00:07:55,389 --> 00:07:59,110 does not have an obligation to 171 00:07:57,250 --> 00:08:00,909 neutrality it's a business decision that 172 00:07:59,110 --> 00:08:03,639 they are making that they want to do 173 00:08:00,909 --> 00:08:05,620 things in this way it is arguably part 174 00:08:03,639 --> 00:08:07,539 of free speech they do need to stay 175 00:08:05,620 --> 00:08:10,419 within campaign finance regulations but 176 00:08:07,539 --> 00:08:13,659 that is not so difficult to do according 177 00:08:10,419 --> 00:08:16,359 to UCLA law professor Eugene Volokh who 178 00:08:13,659 --> 00:08:17,860 told Gizmodo facebook has the same First 179 00:08:16,360 --> 00:08:21,129 Amendment rights as the new york times 180 00:08:17,860 --> 00:08:22,990 you could argue that Google does not 181 00:08:21,129 --> 00:08:24,849 have these rights that google has an 182 00:08:22,990 --> 00:08:27,759 obligation to provide the most accurate 183 00:08:24,849 --> 00:08:29,800 or highest quality search results this 184 00:08:27,759 --> 00:08:32,709 law professor argues that Google does 185 00:08:29,800 --> 00:08:35,110 not have that obligation you could argue 186 00:08:32,708 --> 00:08:37,598 either way I think either other 187 00:08:35,110 --> 00:08:41,529 dave is valid I'm not a locks for though 188 00:08:37,599 --> 00:08:43,990 as for twitter i don't know is it 189 00:08:41,529 --> 00:08:47,380 business is a bias or is it like just a 190 00:08:43,990 --> 00:08:50,380 shitshow I don't know it's a good 191 00:08:47,380 --> 00:08:51,399 question i think about this a lot i 192 00:08:50,380 --> 00:08:55,990 wonder 193 00:08:51,399 --> 00:08:59,890 let's look at their curation so this is 194 00:08:55,990 --> 00:09:02,380 a sample of the moments that were on 195 00:08:59,890 --> 00:09:04,750 Twitter today earlier today 196 00:09:02,380 --> 00:09:09,490 it includes funniest gifs and memes to 197 00:09:04,750 --> 00:09:14,440 come from the RNC Donald Trump is now 198 00:09:09,490 --> 00:09:17,500 the official nominee guest RNC but what 199 00:09:14,440 --> 00:09:19,899 does that mean it seems like whoever 200 00:09:17,500 --> 00:09:22,839 created this was looking for the most 201 00:09:19,899 --> 00:09:27,010 entertaining not necessarily the most 202 00:09:22,839 --> 00:09:30,279 newsworthy or the most tweeted compared 203 00:09:27,010 --> 00:09:34,240 to this set of moments which is fairly 204 00:09:30,279 --> 00:09:37,660 dry i think um but is supposed to be 205 00:09:34,240 --> 00:09:40,269 what was the top three tweeted moments 206 00:09:37,660 --> 00:09:43,149 this came from the Twitter government 207 00:09:40,269 --> 00:09:46,149 account as opposed to the moments 208 00:09:43,149 --> 00:09:48,160 account they're both Twitter curated 209 00:09:46,149 --> 00:09:52,329 accounts and I think they're interesting 210 00:09:48,160 --> 00:09:55,569 as examples of of generation which just 211 00:09:52,329 --> 00:09:58,599 much like Facebook Twitter moments is 212 00:09:55,570 --> 00:10:01,839 curated by people people who have biases 213 00:09:58,600 --> 00:10:03,399 and the lack of transparency around how 214 00:10:01,839 --> 00:10:10,029 all of this is dawn leads to questions 215 00:10:03,399 --> 00:10:12,490 and skepticism why I'm not voting for 216 00:10:10,029 --> 00:10:15,370 hillary was the number-one trending 217 00:10:12,490 --> 00:10:18,519 hashtag and then it wasn't it was 218 00:10:15,370 --> 00:10:22,029 replaced by another hashtag Hillary 2016 219 00:10:18,519 --> 00:10:24,640 now it's possible that one hashtag just 220 00:10:22,029 --> 00:10:28,240 overtook the other but Twitter's lack of 221 00:10:24,640 --> 00:10:30,370 transparency leads to skepticism and 222 00:10:28,240 --> 00:10:33,699 questions and just a general lack of 223 00:10:30,370 --> 00:10:34,990 trust and I wonder whether bands will 224 00:10:33,699 --> 00:10:37,599 impact the election 225 00:10:34,990 --> 00:10:40,089 twitter is banning conservatives who in 226 00:10:37,600 --> 00:10:45,010 some cases may vary rightfully deserve 227 00:10:40,089 --> 00:10:48,160 it but nonetheless it could potentially 228 00:10:45,010 --> 00:10:48,550 impact the election if those folks are 229 00:10:48,160 --> 00:10:50,740 no 230 00:10:48,550 --> 00:10:53,560 longer tweeting you could say that's 231 00:10:50,740 --> 00:10:56,680 good or you could say that's bad but it 232 00:10:53,560 --> 00:10:59,979 does seem clear that Twitter's bands are 233 00:10:56,680 --> 00:11:02,349 more heavily conservative than in other 234 00:10:59,980 --> 00:11:03,610 directions i'm not saying whether that's 235 00:11:02,350 --> 00:11:06,490 good or bad but i do think it would 236 00:11:03,610 --> 00:11:09,459 impact the election so let's shift gears 237 00:11:06,490 --> 00:11:11,920 a bit and talk about Google apparently 238 00:11:09,459 --> 00:11:14,979 it wasn't a glitch when a State 239 00:11:11,920 --> 00:11:17,709 Department video was cut from youtube 240 00:11:14,980 --> 00:11:21,670 this is very interesting 241 00:11:17,709 --> 00:11:23,619 it's very interesting classified email 242 00:11:21,670 --> 00:11:25,569 suggest that Hillary Clinton worked with 243 00:11:23,620 --> 00:11:28,690 google and youtube to block those videos 244 00:11:25,570 --> 00:11:33,579 and we do know that Google and the White 245 00:11:28,690 --> 00:11:35,860 House are very very friendly or books is 246 00:11:33,579 --> 00:11:37,959 here and there are vending this book 247 00:11:35,860 --> 00:11:40,540 which is an interesting read on the 248 00:11:37,959 --> 00:11:42,670 topic from Julian Assange which 249 00:11:40,540 --> 00:11:44,920 documents his conversations with eric 250 00:11:42,670 --> 00:11:51,610 schmidt which documents Eric Schmidt's 251 00:11:44,920 --> 00:11:55,029 closeness with the White House eric 252 00:11:51,610 --> 00:11:57,550 schmidt started a start-up which has the 253 00:11:55,029 --> 00:12:00,220 aim of helping the Clinton campaign 254 00:11:57,550 --> 00:12:02,620 which actually you know it's like pretty 255 00:12:00,220 --> 00:12:06,820 cool the way that campaigns are becoming 256 00:12:02,620 --> 00:12:08,560 much more technical but not a lot of 257 00:12:06,820 --> 00:12:10,839 transparency there either 258 00:12:08,560 --> 00:12:13,899 this is the webpage for the groundwork 259 00:12:10,839 --> 00:12:15,399 the startup started by eric schmidt that 260 00:12:13,899 --> 00:12:17,920 is working closely with the Clinton 261 00:12:15,399 --> 00:12:20,290 campaign it's so stealthy the that's all 262 00:12:17,920 --> 00:12:21,310 that's there that's it that's that's 263 00:12:20,290 --> 00:12:23,230 everything 264 00:12:21,310 --> 00:12:25,180 the only reason I why have any 265 00:12:23,230 --> 00:12:27,790 background information on the groundwork 266 00:12:25,180 --> 00:12:30,399 is because I'm it's been documented 267 00:12:27,790 --> 00:12:33,849 elsewhere by reporters and researchers 268 00:12:30,399 --> 00:12:36,100 and I'm actually quite fond of how the 269 00:12:33,850 --> 00:12:38,140 Clinton campaign has handled many things 270 00:12:36,100 --> 00:12:40,570 i think that they seem to be a very 271 00:12:38,140 --> 00:12:43,060 intelligent and well run campaign 272 00:12:40,570 --> 00:12:45,910 nonetheless it is noteworthy that on 273 00:12:43,060 --> 00:12:48,609 many of their top people do help from 274 00:12:45,910 --> 00:12:51,819 google there's a some kind of a close 275 00:12:48,610 --> 00:12:54,190 relationship there that's that search 276 00:12:51,820 --> 00:12:58,060 does seem legit questions were raised 277 00:12:54,190 --> 00:13:01,290 about whether Google search results were 278 00:12:58,060 --> 00:13:03,239 sanitized on behalf of Hillary Clinton I 279 00:13:01,290 --> 00:13:06,120 this looked as if it was some kind of a 280 00:13:03,240 --> 00:13:09,870 scandal I mean if you look at the 281 00:13:06,120 --> 00:13:11,519 results they're they're not as 282 00:13:09,870 --> 00:13:13,709 interesting as the results you get with 283 00:13:11,519 --> 00:13:16,079 the same query on being on on other 284 00:13:13,709 --> 00:13:20,819 search engines but the same is true for 285 00:13:16,079 --> 00:13:22,888 Trump and in fact google autocomplete is 286 00:13:20,819 --> 00:13:26,189 dull ache for everyone 287 00:13:22,889 --> 00:13:28,019 you name a criminal named your favorite 288 00:13:26,190 --> 00:13:30,420 terminal and type your favorite 289 00:13:28,019 --> 00:13:32,430 criminals name into Google and they just 290 00:13:30,420 --> 00:13:36,569 won't autocomplete crime or criminal 291 00:13:32,430 --> 00:13:38,579 these algorithms are strange google 292 00:13:36,569 --> 00:13:40,920 doesn't want you searching for a 293 00:13:38,579 --> 00:13:43,620 person's name and then the word criminal 294 00:13:40,920 --> 00:13:45,630 for whatever reason maybe someone who 295 00:13:43,620 --> 00:13:47,040 worked on the algorithm decided that 296 00:13:45,630 --> 00:13:47,970 like those just weren't great search 297 00:13:47,040 --> 00:13:51,180 results 298 00:13:47,970 --> 00:13:54,779 I think if there's any one theme that 299 00:13:51,180 --> 00:13:56,969 comes out from this talk by knee today 300 00:13:54,779 --> 00:14:01,230 it's that this lack of transparency 301 00:13:56,970 --> 00:14:02,639 raises questions and is frustrating and 302 00:14:01,230 --> 00:14:03,959 and confusing 303 00:14:02,639 --> 00:14:07,500 I mean we really don't know the reason 304 00:14:03,959 --> 00:14:09,750 why Google works this way and because we 305 00:14:07,500 --> 00:14:12,930 don't know why sometimes we imagined or 306 00:14:09,750 --> 00:14:15,209 create theories about why that aren't 307 00:14:12,930 --> 00:14:17,099 very favorable and the truth is we 308 00:14:15,209 --> 00:14:19,680 actually don't know whether this is 309 00:14:17,100 --> 00:14:21,540 secretly to help one group or another 310 00:14:19,680 --> 00:14:28,050 group we just don't know the reasons 311 00:14:21,540 --> 00:14:30,029 what did you set maybe maybe it's to 312 00:14:28,050 --> 00:14:31,979 help conceal criminals who knows the 313 00:14:30,029 --> 00:14:34,740 lack of transparency for me is extremely 314 00:14:31,980 --> 00:14:37,470 frustrating but I it does seem clear 315 00:14:34,740 --> 00:14:40,889 that it's it's not specifically to help 316 00:14:37,470 --> 00:14:42,930 say Clinton or Trump google autocomplete 317 00:14:40,889 --> 00:14:48,480 doll for everyone 318 00:14:42,930 --> 00:14:50,489 let's go back to facebook this actually 319 00:14:48,480 --> 00:14:54,600 makes me a little worried and nervous 320 00:14:50,490 --> 00:14:58,949 and sad facebook runs experiments all 321 00:14:54,600 --> 00:15:03,509 the time they don't always tell us they 322 00:14:58,949 --> 00:15:05,639 don't have to google messed with your 323 00:15:03,509 --> 00:15:09,029 feed in 2012 324 00:15:05,639 --> 00:15:12,000 that's not the only time they did that 325 00:15:09,029 --> 00:15:13,880 to promote voting and they also messed 326 00:15:12,000 --> 00:15:15,830 with our feet in 2010 327 00:15:13,880 --> 00:15:18,860 which is there right right like they get 328 00:15:15,830 --> 00:15:22,640 to do whatever they want this is their 329 00:15:18,860 --> 00:15:24,080 stage they are in charge i think the 330 00:15:22,640 --> 00:15:27,170 issue with facebook is it feels like 331 00:15:24,080 --> 00:15:28,970 it's my page but it's very heavily 332 00:15:27,170 --> 00:15:31,969 curated what they're going to show for 333 00:15:28,970 --> 00:15:34,850 my friends and then they've started to 334 00:15:31,970 --> 00:15:37,310 really do those messages at the top of 335 00:15:34,850 --> 00:15:41,840 the page like suggesting that someone 336 00:15:37,310 --> 00:15:44,479 goes out and votes they dip is 260 337 00:15:41,840 --> 00:15:46,910 million users that's a lot of users that 338 00:15:44,480 --> 00:15:50,540 actually really will impact panel an 339 00:15:46,910 --> 00:15:52,640 election this was the paper that they 340 00:15:50,540 --> 00:15:54,380 published about it at least they come 341 00:15:52,640 --> 00:15:55,130 and they wrote about it it was after the 342 00:15:54,380 --> 00:15:58,370 fact 343 00:15:55,130 --> 00:16:00,110 here we report results from a randomized 344 00:15:58,370 --> 00:16:02,090 control trial of political mobilization 345 00:16:00,110 --> 00:16:04,490 messages delivered to 61 million 346 00:16:02,090 --> 00:16:09,830 facebook users during the Toulon during 347 00:16:04,490 --> 00:16:12,080 the congressional elections so it's a 348 00:16:09,830 --> 00:16:13,970 kind of digital gerrymandering if they 349 00:16:12,080 --> 00:16:15,620 were to show this to some people and not 350 00:16:13,970 --> 00:16:19,460 to others and that actually was the 351 00:16:15,620 --> 00:16:21,440 experiment some people got shown buttons 352 00:16:19,460 --> 00:16:22,700 that were more encouraging to vote than 353 00:16:21,440 --> 00:16:24,980 other people 354 00:16:22,700 --> 00:16:28,280 now we could just say civic engagement 355 00:16:24,980 --> 00:16:30,410 is good and that it's neutral and for 356 00:16:28,280 --> 00:16:35,120 the most part it's hard to argue against 357 00:16:30,410 --> 00:16:39,020 that but when I spoke to someone who 358 00:16:35,120 --> 00:16:42,020 works in Democratic campaigns and i said 359 00:16:39,020 --> 00:16:44,600 i would like to be helpful this election 360 00:16:42,020 --> 00:16:47,030 cycle to the Democrats you know how can 361 00:16:44,600 --> 00:16:50,870 i what can I do I'm here I'm here for 362 00:16:47,030 --> 00:16:53,300 you and they said that they didn't want 363 00:16:50,870 --> 00:16:56,240 me to promote their specific candidate 364 00:16:53,300 --> 00:16:59,420 that the most helpful thing I could do 365 00:16:56,240 --> 00:17:02,600 for the Democratic Party would be to 366 00:16:59,420 --> 00:17:06,649 just register more voters which was 367 00:17:02,600 --> 00:17:09,050 strange to me I I thought that the way 368 00:17:06,650 --> 00:17:13,339 to help a candidate win was to help the 369 00:17:09,050 --> 00:17:15,319 candidate went and and they said alyssa 370 00:17:13,339 --> 00:17:18,198 you know you've worked on Republican 371 00:17:15,319 --> 00:17:20,240 races because I helped her and Paul but 372 00:17:18,199 --> 00:17:22,880 democrats democrats we run races by 373 00:17:20,240 --> 00:17:27,140 registering voters that's how we went so 374 00:17:22,880 --> 00:17:27,660 you know when I see companies working 375 00:17:27,140 --> 00:17:30,030 together 376 00:17:27,660 --> 00:17:32,670 the vote i I'm a little bit skeptical 377 00:17:30,030 --> 00:17:34,889 because Democrats win elections by 378 00:17:32,670 --> 00:17:37,290 increasing voter turnout so if a 379 00:17:34,890 --> 00:17:38,880 platform like Facebook is working to 380 00:17:37,290 --> 00:17:40,649 increase voter turnout like on the one 381 00:17:38,880 --> 00:17:43,740 hand of course that's great on the other 382 00:17:40,650 --> 00:17:48,030 hand is it neutral maybe not 383 00:17:43,740 --> 00:17:49,440 and his neutrality an impossible goal i 384 00:17:48,030 --> 00:17:53,430 don't i don't know 385 00:17:49,440 --> 00:17:55,830 here's another curation escape his 386 00:17:53,430 --> 00:17:57,510 caricature Donald Trump passes up the 387 00:17:55,830 --> 00:18:02,429 chance 388 00:17:57,510 --> 00:18:06,390 it'sit's just not neutral and something 389 00:18:02,430 --> 00:18:12,330 much more favorable to clinton today so 390 00:18:06,390 --> 00:18:14,880 the next place i take that is this very 391 00:18:12,330 --> 00:18:18,659 hope like critique where I think the 392 00:18:14,880 --> 00:18:22,290 chance that we have for having actual 393 00:18:18,660 --> 00:18:24,960 neutrality on our platforms on is if 394 00:18:22,290 --> 00:18:27,300 we've removed the centralized curators 395 00:18:24,960 --> 00:18:30,210 and can actually just exchanged 396 00:18:27,300 --> 00:18:32,340 one-on-one the censorship is is 397 00:18:30,210 --> 00:18:34,410 problematic because there's just no 398 00:18:32,340 --> 00:18:35,490 transparency we just really don't know 399 00:18:34,410 --> 00:18:39,360 what's going on 400 00:18:35,490 --> 00:18:42,150 can I start opening it up for questions 401 00:18:39,360 --> 00:18:44,040 although it's very dark so I can't see 402 00:18:42,150 --> 00:18:47,790 anyone i might need some help 403 00:18:44,040 --> 00:18:53,250 could i get some help from one of the 404 00:18:47,790 --> 00:18:59,789 conference organizers thanks 405 00:18:53,250 --> 00:19:20,879 alright let's open it up for questions 406 00:18:59,789 --> 00:19:23,908 so I have worked for google for traders 407 00:19:20,879 --> 00:19:26,039 but these are in no way of presenting of 408 00:19:23,909 --> 00:19:29,759 use of either Google or Twitter okay 409 00:19:26,039 --> 00:19:32,070 positive negative Lee so a I think 410 00:19:29,759 --> 00:19:36,029 you're making a couple of logical 411 00:19:32,070 --> 00:19:39,418 mistakes here in your analyses one is 412 00:19:36,029 --> 00:19:41,429 you're ignoring the echo chamber effect 413 00:19:39,419 --> 00:19:44,759 when you when i'm looking at facebook 414 00:19:41,429 --> 00:19:48,149 pages and what facebook would show me 415 00:19:44,759 --> 00:19:51,119 and as interesting for me is what i have 416 00:19:48,149 --> 00:19:54,539 said what my friends have said now most 417 00:19:51,119 --> 00:19:58,289 of my friends are not hardcore 418 00:19:54,539 --> 00:20:00,450 Republicans will vote for Trump so it is 419 00:19:58,289 --> 00:20:04,289 natural except it's going into this guy 420 00:20:00,450 --> 00:20:07,679 here so it's natural that facebook won't 421 00:20:04,289 --> 00:20:10,320 show me stuff about Trump now if I were 422 00:20:07,679 --> 00:20:14,340 to just create a sock puppet that would 423 00:20:10,320 --> 00:20:16,320 that that would listen to the Westboro 424 00:20:14,340 --> 00:20:19,439 Baptist Church and rudy giuliani and 425 00:20:16,320 --> 00:20:21,960 Trump and another such things 426 00:20:19,440 --> 00:20:23,789 maybe i would get a difference in the 427 00:20:21,960 --> 00:20:28,139 results have you tried that experiment 428 00:20:23,789 --> 00:20:29,999 of creating a persona that does not fit 429 00:20:28,139 --> 00:20:31,949 the model you're trying to accuse or the 430 00:20:29,999 --> 00:20:34,499 modalities that you're suggesting that 431 00:20:31,950 --> 00:20:38,369 facebook is promoting and seeing what 432 00:20:34,499 --> 00:20:41,009 the results you get our I think that's 433 00:20:38,369 --> 00:20:44,189 right i think that it would be helpful 434 00:20:41,009 --> 00:20:47,039 to see it from different vantage point 435 00:20:44,190 --> 00:20:50,279 right but because some of that it 436 00:20:47,039 --> 00:20:52,200 doesn't change that the moments for the 437 00:20:50,279 --> 00:20:56,099 most part are universal 438 00:20:52,200 --> 00:20:57,840 I I don't that's true for twitter and i 439 00:20:56,099 --> 00:21:01,859 believe it's true for facebook as well 440 00:20:57,840 --> 00:21:05,179 that these aren't as highly customized 441 00:21:01,859 --> 00:21:20,899 as we would think 442 00:21:05,180 --> 00:21:23,420 everyone that's true for your newsfeed 443 00:21:20,900 --> 00:21:27,200 ah i'm curious about how true it is for 444 00:21:23,420 --> 00:21:30,110 what's trending at large 445 00:21:27,200 --> 00:21:31,010 can you come to my place I'm not done 446 00:21:30,110 --> 00:21:38,449 good 447 00:21:31,010 --> 00:21:41,300 doesn't that doesn't apply to it what it 448 00:21:38,450 --> 00:21:44,270 applies to is is each person's newsfeed 449 00:21:41,300 --> 00:21:49,040 right that is different for every person 450 00:21:44,270 --> 00:21:57,110 that's right what I'm referring to here 451 00:21:49,040 --> 00:22:04,970 is the trending section here 452 00:21:57,110 --> 00:22:08,120 yes the newsfeeds very customized well 453 00:22:04,970 --> 00:22:10,280 the newsfeed will be customized for the 454 00:22:08,120 --> 00:22:12,469 sort of articles i would be likely to 455 00:22:10,280 --> 00:22:14,870 read the show to click on them look at 456 00:22:12,470 --> 00:22:17,120 the ads and then make Facebook or Google 457 00:22:14,870 --> 00:22:19,370 or Twitter money they wouldn't rope if 458 00:22:17,120 --> 00:22:22,370 they showed me an ad about oh I don't 459 00:22:19,370 --> 00:22:24,379 know the starving children of africa i 460 00:22:22,370 --> 00:22:26,659 would probably not click on it and not 461 00:22:24,380 --> 00:22:28,100 read the article and you know not get 462 00:22:26,660 --> 00:22:29,720 the money so of course they won't do 463 00:22:28,100 --> 00:22:33,199 that they do a very good job of 464 00:22:29,720 --> 00:22:36,170 customizing those yeah so how can you 465 00:22:33,200 --> 00:22:38,870 differentiate what what is the the your 466 00:22:36,170 --> 00:22:43,550 you're out your algorithm for 467 00:22:38,870 --> 00:22:47,479 differentiating between a feed that is 468 00:22:43,550 --> 00:22:50,149 easily curated to only show those those 469 00:22:47,480 --> 00:22:53,060 even people that you don't want that 470 00:22:50,150 --> 00:22:54,740 you're accusing of up let me Tony down 471 00:22:53,060 --> 00:22:58,250 you know how can you differentiate 472 00:22:54,740 --> 00:23:01,820 between a feed that is customized 473 00:22:58,250 --> 00:23:05,990 according to some guidelines from above 474 00:23:01,820 --> 00:23:08,899 vs feed that happens to shoot to appear 475 00:23:05,990 --> 00:23:09,950 that way because of all your prior post 476 00:23:08,900 --> 00:23:12,620 in history 477 00:23:09,950 --> 00:23:15,020 well we need to separate news feed which 478 00:23:12,620 --> 00:23:16,998 is highly personalized and highly 479 00:23:15,020 --> 00:23:20,449 algorithmic vs 480 00:23:16,999 --> 00:23:23,269 what you can see here which is like 481 00:23:20,449 --> 00:23:26,989 national trending topics which is 482 00:23:23,269 --> 00:23:28,459 curated by humans skin discoloration 483 00:23:26,989 --> 00:23:32,419 that you're that you're talking about 484 00:23:28,459 --> 00:23:32,899 are the people who have you mentioned 485 00:23:32,419 --> 00:23:36,469 that 486 00:23:32,899 --> 00:23:37,458 oh excuse later said that and the rumor 487 00:23:36,469 --> 00:23:41,929 said that 488 00:23:37,459 --> 00:23:44,509 how have you actually are verified that 489 00:23:41,929 --> 00:23:47,809 this is indeed the case you know I heard 490 00:23:44,509 --> 00:23:50,299 from from from my dog yesterday that you 491 00:23:47,809 --> 00:23:52,579 know what Mark Zuckerberg girl don't 492 00:23:50,299 --> 00:23:55,009 need something like up just because you 493 00:23:52,579 --> 00:23:59,149 said it executes later said that did you 494 00:23:55,009 --> 00:24:01,639 reveal who ever publish that story did 495 00:23:59,149 --> 00:24:03,918 they go out and actually see that his 496 00:24:01,639 --> 00:24:05,718 head was actually an executor you know 497 00:24:03,919 --> 00:24:07,849 the the journalistic standards that we 498 00:24:05,719 --> 00:24:09,289 hope that we usually hold up of the 499 00:24:07,849 --> 00:24:12,859 Washington Post and The New York Times 500 00:24:09,289 --> 00:24:15,709 hundred years ago or 50 years ago our 501 00:24:12,859 --> 00:24:17,329 are a lot different that that and we 502 00:24:15,709 --> 00:24:19,279 expect when when you get the New York 503 00:24:17,329 --> 00:24:21,948 Times The Washington Post says you know 504 00:24:19,279 --> 00:24:23,749 about something happened that would be 505 00:24:21,949 --> 00:24:26,689 generally have actually check check I 506 00:24:23,749 --> 00:24:30,919 think that's fair that we just don't 507 00:24:26,689 --> 00:24:33,649 know what's going on with the curation 508 00:24:30,919 --> 00:24:36,559 and the trans and I certainly don't mean 509 00:24:33,649 --> 00:24:39,859 to say that we know for sure that 510 00:24:36,559 --> 00:24:42,199 there's this Lander that slant I think 511 00:24:39,859 --> 00:24:47,749 that it's compelling the evidence that 512 00:24:42,199 --> 00:24:51,199 there was the evidence that i was going 513 00:24:47,749 --> 00:24:53,359 on in this team that was a contract team 514 00:24:51,199 --> 00:24:55,279 that wasn't being given a lot of 515 00:24:53,359 --> 00:24:58,039 attention that like there may have been 516 00:24:55,279 --> 00:25:00,199 some things happening there that weren't 517 00:24:58,039 --> 00:25:02,419 up to facebook standards or up to the 518 00:25:00,199 --> 00:25:04,279 standards that we would want but that 519 00:25:02,419 --> 00:25:06,559 that's a lot of the point that's a lot 520 00:25:04,279 --> 00:25:08,239 of the point of this talk is that uh so 521 00:25:06,559 --> 00:25:10,069 much of this is a black box and we just 522 00:25:08,239 --> 00:25:13,549 don't know and that it would be possible 523 00:25:10,069 --> 00:25:15,979 for social media companies to have an 524 00:25:13,549 --> 00:25:18,319 outsized impact on the elections and in 525 00:25:15,979 --> 00:25:22,970 the absence of transparency about their 526 00:25:18,319 --> 00:25:27,139 methods were left not knowing 527 00:25:22,970 --> 00:25:29,600 thank you you got next week we give up 528 00:25:27,140 --> 00:25:31,280 the mic and come back thank you for your 529 00:25:29,600 --> 00:25:34,850 question and thank you for the cheering 530 00:25:31,280 --> 00:25:37,010 I always was particularly interested in 531 00:25:34,850 --> 00:25:38,600 your last slide we had a p2p model not 532 00:25:37,010 --> 00:25:40,460 kind of yes forgetting what is most 533 00:25:38,600 --> 00:25:42,709 interesting to you know like kind of 534 00:25:40,460 --> 00:25:44,390 figuring out in a decentralized way what 535 00:25:42,710 --> 00:25:46,909 is most interesting come to show people 536 00:25:44,390 --> 00:25:48,530 um I do kind of wonder how that how you 537 00:25:46,909 --> 00:25:50,150 would defend against creating echo 538 00:25:48,530 --> 00:25:51,408 chambers are just little pods of people 539 00:25:50,150 --> 00:25:53,570 to share the same news amongst each 540 00:25:51,409 --> 00:25:54,530 other and kind of always thinking that I 541 00:25:53,570 --> 00:25:56,510 was thinking you know there's a lot of 542 00:25:54,530 --> 00:25:58,070 talk about how everyone getting news 543 00:25:56,510 --> 00:26:00,950 from a different source kind of like 544 00:25:58,070 --> 00:26:01,939 helps to just make just many sizes just 545 00:26:00,950 --> 00:26:04,220 don't want to talk to each other anymore 546 00:26:01,940 --> 00:26:06,320 and how like having fewer new sources 547 00:26:04,220 --> 00:26:08,090 like everyone would watch one or two of 548 00:26:06,320 --> 00:26:09,649 them and they're like more common 549 00:26:08,090 --> 00:26:13,100 watching and so you would have less 550 00:26:09,650 --> 00:26:15,980 extreme is divisiveness in kind of hold 551 00:26:13,100 --> 00:26:17,719 political process so i do think there's 552 00:26:15,980 --> 00:26:20,870 a there's a kind of a part to play for 553 00:26:17,720 --> 00:26:23,120 these outlets to maybe show multiple 554 00:26:20,870 --> 00:26:24,830 like a diversity of things that may be 555 00:26:23,120 --> 00:26:26,030 interesting so they aren't just making 556 00:26:24,830 --> 00:26:28,970 echo chambers 557 00:26:26,030 --> 00:26:31,460 um it's going to be a question if i 558 00:26:28,970 --> 00:26:35,000 understand it is how would we do a 559 00:26:31,460 --> 00:26:37,400 peer-to-peer model that would bypass the 560 00:26:35,000 --> 00:26:40,460 echo chamber problem yes quite right 561 00:26:37,400 --> 00:26:42,200 huh that's a really hard problem that's 562 00:26:40,460 --> 00:26:43,909 like a human problem right 563 00:26:42,200 --> 00:26:47,450 I think you solved that in a human way 564 00:26:43,909 --> 00:26:50,360 uh I have a lot of friends who have 565 00:26:47,450 --> 00:26:54,380 different points of view and I engage 566 00:26:50,360 --> 00:26:57,530 with them and I've learned to stop doing 567 00:26:54,380 --> 00:26:59,690 that as much because I've got a negative 568 00:26:57,530 --> 00:27:01,100 feedback because I engage with people 569 00:26:59,690 --> 00:27:03,080 are very different points of view 570 00:27:01,100 --> 00:27:04,760 sometimes they will troll me or 571 00:27:03,080 --> 00:27:08,240 sometimes they'll be very mean or the 572 00:27:04,760 --> 00:27:10,070 whole my friends or people who have a 573 00:27:08,240 --> 00:27:12,020 similar point of view to me will get mad 574 00:27:10,070 --> 00:27:14,689 at me for interacting with people who 575 00:27:12,020 --> 00:27:17,450 they think are wrong so I actually 576 00:27:14,690 --> 00:27:18,710 wanted to be sure to started out as 577 00:27:17,450 --> 00:27:23,120 someone who was mentored by ethan 578 00:27:18,710 --> 00:27:25,190 Zuckerman who wrote rewire like his old 579 00:27:23,120 --> 00:27:26,570 thesis was that the internet is an echo 580 00:27:25,190 --> 00:27:30,350 chamber and we need to break out of that 581 00:27:26,570 --> 00:27:33,260 so like I came into the internet with 582 00:27:30,350 --> 00:27:35,449 that mentality being mentored by him and 583 00:27:33,260 --> 00:27:39,110 now I'm becoming smaller 584 00:27:35,450 --> 00:27:40,970 because of all the bad feedback that I'm 585 00:27:39,110 --> 00:27:43,189 getting is I actually engage with 586 00:27:40,970 --> 00:27:45,350 diverse views and and I say that in a 587 00:27:43,190 --> 00:27:47,090 sheepish way I don't want to I don't 588 00:27:45,350 --> 00:27:48,799 want my worldview to become smaller I 589 00:27:47,090 --> 00:27:50,899 don't want to become smaller as a person 590 00:27:48,799 --> 00:27:53,899 but it's what's happening 591 00:27:50,899 --> 00:27:55,850 sure i started deleting posts on my 592 00:27:53,899 --> 00:27:57,739 facebook last week for the first time 593 00:27:55,850 --> 00:28:01,730 kind of apologizing to people and 594 00:27:57,740 --> 00:28:03,649 explaining why so i think one first step 595 00:28:01,730 --> 00:28:07,970 is for us to be nicer to each other 596 00:28:03,649 --> 00:28:10,850 sure to agree to disagree yeah like if 597 00:28:07,970 --> 00:28:13,490 we could agree to disagree then it would 598 00:28:10,850 --> 00:28:16,010 be a lot easier to free exchange 599 00:28:13,490 --> 00:28:18,230 um yeah I think the bottom line for me 600 00:28:16,010 --> 00:28:21,440 is that's like a human problem in a 601 00:28:18,230 --> 00:28:23,690 kindness problem and I like can we be 602 00:28:21,440 --> 00:28:25,039 interested in views beyond our own with 603 00:28:23,690 --> 00:28:26,600 the humbleness that there's something to 604 00:28:25,039 --> 00:28:28,309 learn problem 605 00:28:26,600 --> 00:28:29,510 sure that that's a really big ass was 606 00:28:28,309 --> 00:28:31,340 awesome and i love it i think it should 607 00:28:29,510 --> 00:28:33,049 happen in the meantime maybe do you 608 00:28:31,340 --> 00:28:36,439 think it's possible for like Twitter and 609 00:28:33,049 --> 00:28:38,539 Facebook um to use their kind of power 610 00:28:36,440 --> 00:28:40,580 that they have just by doing these feeds 611 00:28:38,539 --> 00:28:42,260 not feasible like the trending topics um 612 00:28:40,580 --> 00:28:43,879 to try and force that hand a little bit 613 00:28:42,260 --> 00:28:46,010 like you think they have the capacity to 614 00:28:43,880 --> 00:28:47,480 be more even-handed i guess i'll just be 615 00:28:46,010 --> 00:28:49,669 like the BBC or something like that 616 00:28:47,480 --> 00:28:51,529 power is dry but it tends to be a little 617 00:28:49,669 --> 00:28:53,600 more neutral i think that Twitter and 618 00:28:51,529 --> 00:28:54,620 Facebook just have tremendous power 619 00:28:53,600 --> 00:28:57,350 right now 620 00:28:54,620 --> 00:28:59,750 oh yeah and it's and part of what i came 621 00:28:57,350 --> 00:29:01,580 away with as they did the research for 622 00:28:59,750 --> 00:29:04,429 the talk was just the respect for how 623 00:29:01,580 --> 00:29:07,789 much power Twitter and Facebook and you 624 00:29:04,429 --> 00:29:11,630 know Instagram and snapchat have for 625 00:29:07,789 --> 00:29:15,169 getting us to take actions and for 626 00:29:11,630 --> 00:29:16,730 changing our lens on world events and 627 00:29:15,169 --> 00:29:18,200 that's what's frustrating to me about 628 00:29:16,730 --> 00:29:21,049 the notion that they're neutral they're 629 00:29:18,200 --> 00:29:24,710 just so powerful and I want to see more 630 00:29:21,049 --> 00:29:27,260 acknowledgement of that and as for like 631 00:29:24,710 --> 00:29:29,720 how they could use that for good i think 632 00:29:27,260 --> 00:29:31,429 the first step might be for them to kind 633 00:29:29,720 --> 00:29:32,990 of acknowledge that there are entities 634 00:29:31,429 --> 00:29:35,779 that are trying to do things in the 635 00:29:32,990 --> 00:29:38,899 world as long as Facebook and Twitter 636 00:29:35,779 --> 00:29:40,880 take a stand and say we're neutral and 637 00:29:38,899 --> 00:29:42,529 we're not really doing much but by the 638 00:29:40,880 --> 00:29:44,120 way we're gonna run this experiment and 639 00:29:42,529 --> 00:29:47,360 some of you will vote in some of you 640 00:29:44,120 --> 00:29:48,658 want based on what we show you as long 641 00:29:47,360 --> 00:29:53,158 as there's this 642 00:29:48,659 --> 00:29:54,659 pretense that they're like bystanders um 643 00:29:53,159 --> 00:29:57,330 I think it would be hard for them to 644 00:29:54,659 --> 00:30:02,099 become best actors right to thank you 645 00:29:57,330 --> 00:30:05,399 thank you thank you 646 00:30:02,099 --> 00:30:07,979 um I do I probably should confess before 647 00:30:05,399 --> 00:30:10,049 I start is that I'm not unconsciously 648 00:30:07,979 --> 00:30:12,090 not on facebook because i am dismayed 649 00:30:10,049 --> 00:30:14,729 about the amount of power they have 650 00:30:12,090 --> 00:30:17,549 but what I'm saying is you're in good 651 00:30:14,729 --> 00:30:20,429 company hear what I like to point out is 652 00:30:17,549 --> 00:30:22,320 that no private company has transparency 653 00:30:20,429 --> 00:30:24,720 unless they want to i mean just think 654 00:30:22,320 --> 00:30:26,309 about the credit reporting agencies and 655 00:30:24,720 --> 00:30:31,139 the problem that I have with this 656 00:30:26,309 --> 00:30:33,599 Facebook story is they have no 657 00:30:31,139 --> 00:30:35,459 obligation to be neutral and the truth 658 00:30:33,599 --> 00:30:37,830 is if you're getting your news and 659 00:30:35,460 --> 00:30:41,129 you're allowing companies like Facebook 660 00:30:37,830 --> 00:30:42,418 or Twitter or snapchat app to mold your 661 00:30:41,129 --> 00:30:44,459 political views 662 00:30:42,419 --> 00:30:45,960 there's something wrong right there you 663 00:30:44,460 --> 00:30:47,519 need to take it back from these 664 00:30:45,960 --> 00:30:50,159 companies they have only as much power 665 00:30:47,519 --> 00:30:52,440 as we give to them and why you're giving 666 00:30:50,159 --> 00:30:55,379 these companies the power to mold your 667 00:30:52,440 --> 00:30:59,820 political views is beyond me 668 00:30:55,379 --> 00:31:02,340 I think yeah yeah I think that's a good 669 00:30:59,820 --> 00:31:04,649 lesson I think that's that was my 670 00:31:02,340 --> 00:31:08,009 takeaway from Snowden's revelations from 671 00:31:04,649 --> 00:31:10,049 so many other moments where we've seen 672 00:31:08,009 --> 00:31:11,879 these companies aggregating data and 673 00:31:10,049 --> 00:31:14,279 becoming very powerful that's power that 674 00:31:11,879 --> 00:31:17,609 they only have that we give them that's 675 00:31:14,279 --> 00:31:21,059 a very important point I i would argue 676 00:31:17,609 --> 00:31:23,879 that it's my hope that companies will be 677 00:31:21,059 --> 00:31:26,940 true to their word and it's my hope that 678 00:31:23,879 --> 00:31:29,219 companies will be true to the power and 679 00:31:26,940 --> 00:31:31,320 responsibility they have that's a very 680 00:31:29,220 --> 00:31:35,580 optimistic view they don't need to do 681 00:31:31,320 --> 00:31:37,889 that but we can call on them and and and 682 00:31:35,580 --> 00:31:41,249 you know come together and say we think 683 00:31:37,889 --> 00:31:45,389 it matters where we can you know I I 684 00:31:41,249 --> 00:31:54,629 admire your stance 685 00:31:45,389 --> 00:31:58,589 okay thanks Sharky niña like anyway i'm 686 00:31:54,629 --> 00:32:01,109 sitting because I'm like actually that 687 00:31:58,589 --> 00:32:02,458 the previous speaker actually had one of 688 00:32:01,109 --> 00:32:04,499 the things I wanted to say about the 689 00:32:02,459 --> 00:32:06,149 other thing i want to say is i mean i 690 00:32:04,499 --> 00:32:08,039 know a lot of people are suspicious of 691 00:32:06,149 --> 00:32:09,629 government because you know you know 692 00:32:08,039 --> 00:32:11,820 don't you know they're all these ones i 693 00:32:09,629 --> 00:32:13,559 want to go into why but I'm just 694 00:32:11,820 --> 00:32:15,269 thinking of that I think we should also 695 00:32:13,559 --> 00:32:17,190 be suspicious of private business 696 00:32:15,269 --> 00:32:19,320 because you know that's that's what they 697 00:32:17,190 --> 00:32:21,570 are and that people are just way too 698 00:32:19,320 --> 00:32:23,519 trusting and I don't I really don't 699 00:32:21,570 --> 00:32:27,239 think it's reasonable to expect that 700 00:32:23,519 --> 00:32:28,469 these companies are going to you know be 701 00:32:27,239 --> 00:32:30,149 what we want them to be 702 00:32:28,469 --> 00:32:32,159 I mean we have to decide you know like 703 00:32:30,149 --> 00:32:34,468 you said you know decide to engage or 704 00:32:32,159 --> 00:32:36,929 not you know but with our eyes open 705 00:32:34,469 --> 00:32:39,299 they're in it to make money they can do 706 00:32:36,929 --> 00:32:42,119 what they want and just you know that 707 00:32:39,299 --> 00:32:44,849 that's basically where is that they're 708 00:32:42,119 --> 00:32:47,399 in it to make money and in some cases uh 709 00:32:44,849 --> 00:32:50,189 I believe that technologist at these 710 00:32:47,399 --> 00:32:52,859 companies come in either with like a 711 00:32:50,190 --> 00:32:55,679 genuinely ethical point of view from 712 00:32:52,859 --> 00:32:58,109 their point of view you know like is it 713 00:32:55,679 --> 00:32:59,909 our obligation to stop Trump that's an 714 00:32:58,109 --> 00:33:02,309 employee question that's not about money 715 00:32:59,909 --> 00:33:03,899 it's what we can do about civilization i 716 00:33:02,309 --> 00:33:06,178 guess that's probably one of those core 717 00:33:03,899 --> 00:33:08,428 value things that people just do because 718 00:33:06,179 --> 00:33:10,440 of who they are you know when and when 719 00:33:08,429 --> 00:33:13,169 you look at people running experiments 720 00:33:10,440 --> 00:33:14,909 that's coming out of curiosity but with 721 00:33:13,169 --> 00:33:16,769 the so I think we got experiment but the 722 00:33:14,909 --> 00:33:19,049 thing is that it when it's when it's a 723 00:33:16,769 --> 00:33:20,489 private company doing experiments it's a 724 00:33:19,049 --> 00:33:22,109 kind of a different but I guess that's 725 00:33:20,489 --> 00:33:23,759 the thing it isn't that people know 726 00:33:22,109 --> 00:33:25,889 about it now I guess the only other 727 00:33:23,759 --> 00:33:27,570 thing I would mention is back in the day 728 00:33:25,889 --> 00:33:29,879 when we used to have like comment 729 00:33:27,570 --> 00:33:32,549 sections that were not moderated not 730 00:33:29,879 --> 00:33:34,079 curated any of that stuff one of the put 731 00:33:32,549 --> 00:33:37,559 one of those things that happens a lot 732 00:33:34,079 --> 00:33:40,889 is that you get the trolls types you get 733 00:33:37,559 --> 00:33:43,889 wall-to-wall shit there's and then 734 00:33:40,889 --> 00:33:45,029 everybody leaves everybody leaves and 735 00:33:43,889 --> 00:33:47,879 you've got nothing 736 00:33:45,029 --> 00:33:49,860 well maybe we do the smaller groups that 737 00:33:47,879 --> 00:33:52,260 it's an interesting question you 738 00:33:49,860 --> 00:33:55,229 and one that I hope people will continue 739 00:33:52,260 --> 00:34:00,629 of what else we could be doing anyway 740 00:33:55,230 --> 00:34:04,230 ok 741 00:34:00,630 --> 00:34:06,059 hi this is not like really more of a 742 00:34:04,230 --> 00:34:07,950 questions more of a comment and i want 743 00:34:06,059 --> 00:34:10,020 to live off of what the first man was 744 00:34:07,950 --> 00:34:12,389 speaking about but I want to give more 745 00:34:10,020 --> 00:34:15,239 of a differing outlook he was saying 746 00:34:12,389 --> 00:34:17,010 that maybe if you do follow conservative 747 00:34:15,239 --> 00:34:19,618 pages things like that maybe you do get 748 00:34:17,010 --> 00:34:22,560 different search results and I myself I 749 00:34:19,619 --> 00:34:25,080 support down from I follow the GOP on 750 00:34:22,560 --> 00:34:26,310 Twitter Trump and the other candidates 751 00:34:25,080 --> 00:34:29,909 when they were still competing in the 752 00:34:26,310 --> 00:34:31,500 primaries and I might has tagged the 753 00:34:29,909 --> 00:34:35,310 trends everything is still filled with 754 00:34:31,500 --> 00:34:37,500 Trump hate things like that and so I 755 00:34:35,310 --> 00:34:38,969 really and I get Hillary adds all the 756 00:34:37,500 --> 00:34:40,918 time I really don't think has anything 757 00:34:38,969 --> 00:34:42,899 to do with who you follow or what you 758 00:34:40,918 --> 00:34:45,868 say cuz i'm always tweeting pro 759 00:34:42,899 --> 00:34:48,960 conservative things out on Twitter arm 760 00:34:45,869 --> 00:34:50,879 and I think that even if it was based 761 00:34:48,960 --> 00:34:53,730 off of a majority of what people 762 00:34:50,879 --> 00:34:56,639 treating about I still think I disagree 763 00:34:53,730 --> 00:34:58,350 because I don't think if Trump would 764 00:34:56,639 --> 00:35:00,090 have won the primaries by such a 765 00:34:58,350 --> 00:35:00,900 landslide he obviously has such a large 766 00:35:00,090 --> 00:35:02,580 following 767 00:35:00,900 --> 00:35:04,140 and I don't think that the majority of 768 00:35:02,580 --> 00:35:05,819 people tweeting would always be tweeting 769 00:35:04,140 --> 00:35:08,400 such awful things about him 770 00:35:05,820 --> 00:35:11,130 I like I barely ever see if I've ever 771 00:35:08,400 --> 00:35:14,670 seen a positive hashtag about him and 772 00:35:11,130 --> 00:35:18,330 the trends so i really do think that the 773 00:35:14,670 --> 00:35:20,400 large social medias are trying to turn 774 00:35:18,330 --> 00:35:23,549 people left-leaning that's my opinion 775 00:35:20,400 --> 00:35:26,220 twitter in particular there's more 776 00:35:23,550 --> 00:35:28,530 evidence that there is an anti 777 00:35:26,220 --> 00:35:31,230 conservative bias and that shows up 778 00:35:28,530 --> 00:35:33,869 primarily and who they choose to ban and 779 00:35:31,230 --> 00:35:36,359 I'm certainly not defending that because 780 00:35:33,869 --> 00:35:40,290 the band people who are trolling people 781 00:35:36,359 --> 00:35:42,270 but I would argue that I haven't seen 782 00:35:40,290 --> 00:35:46,410 the same number of bands on the other 783 00:35:42,270 --> 00:35:48,240 side who uh in in other social media 784 00:35:46,410 --> 00:35:54,509 organizations like I believe personally 785 00:35:48,240 --> 00:35:56,100 that that's less clear but there I i do 786 00:35:54,510 --> 00:36:00,690 believe that there are legitimate 787 00:35:56,100 --> 00:36:02,700 concerns about bias at twitter twitter 788 00:36:00,690 --> 00:36:04,710 has made less statements and facebook 789 00:36:02,700 --> 00:36:06,660 about being neutral and if I could just 790 00:36:04,710 --> 00:36:09,780 like the the trouble yourself campaign 791 00:36:06,660 --> 00:36:12,390 that Hillary Clinton started on if you 792 00:36:09,780 --> 00:36:15,300 actually go to our website and look in 793 00:36:12,390 --> 00:36:19,200 that thing that she just started it says 794 00:36:15,300 --> 00:36:22,740 how he doesn't want to support lesbian 795 00:36:19,200 --> 00:36:24,750 and gay marriage but like I've seen him 796 00:36:22,740 --> 00:36:26,129 speak live in species board he said that 797 00:36:24,750 --> 00:36:28,950 he wants to protect people like that 798 00:36:26,130 --> 00:36:31,020 from hateful ideologies far an ideology 799 00:36:28,950 --> 00:36:34,259 is not domestic even foreign they're 800 00:36:31,020 --> 00:36:35,849 such a large number of people outside 801 00:36:34,260 --> 00:36:38,339 the united states that still have such 802 00:36:35,849 --> 00:36:41,790 barbaric points of views calling those 803 00:36:38,339 --> 00:36:45,180 ideals for you know that ideas of the 804 00:36:41,790 --> 00:36:48,810 domestic for as it for in themselves 805 00:36:45,180 --> 00:36:50,790 well we should probably move to next 806 00:36:48,810 --> 00:36:54,390 question last week unless we get off 807 00:36:50,790 --> 00:36:56,369 track but it's helpful to hear that your 808 00:36:54,390 --> 00:36:57,960 trends don't seem to reflect your 809 00:36:56,369 --> 00:37:03,240 ideology that's helpful 810 00:36:57,960 --> 00:37:05,369 um I don't know that until this current 811 00:37:03,240 --> 00:37:07,259 careful with the facebooks trending news 812 00:37:05,369 --> 00:37:11,280 stories that either facebook or twitter 813 00:37:07,260 --> 00:37:13,770 or any of these companies a professed 814 00:37:11,280 --> 00:37:17,099 that there or like really made a big 815 00:37:13,770 --> 00:37:18,329 deal out of like the neutrality of the 816 00:37:17,099 --> 00:37:21,060 topics that they're showing in there 817 00:37:18,329 --> 00:37:22,680 trending sections right already wasn't 818 00:37:21,060 --> 00:37:28,078 as important when there is less curation 819 00:37:22,680 --> 00:37:30,569 well but i think i'm a i think that what 820 00:37:28,079 --> 00:37:32,460 people often mistook for what people are 821 00:37:30,569 --> 00:37:35,970 mistaking for neutrality is like oh 822 00:37:32,460 --> 00:37:38,730 that's 100% algorithm powered no human 823 00:37:35,970 --> 00:37:42,779 ever touches it and I think that you 824 00:37:38,730 --> 00:37:44,579 would be really every yes that's very 825 00:37:42,780 --> 00:37:46,920 naive i mean i think that you would be 826 00:37:44,579 --> 00:37:48,450 incredibly as everyone was surprised 827 00:37:46,920 --> 00:37:50,819 that Facebook has dozens of people 828 00:37:48,450 --> 00:37:52,140 working around the clock to like look at 829 00:37:50,819 --> 00:37:54,089 the stuff that's coming through the 830 00:37:52,140 --> 00:38:01,920 algorithms and sort it all out and like 831 00:37:54,089 --> 00:38:03,480 make it not totally garbage like all of 832 00:38:01,920 --> 00:38:08,160 these companies have staffs of those 833 00:38:03,480 --> 00:38:11,339 people i think that the idea that an 834 00:38:08,160 --> 00:38:15,730 algorithm is like has no bias because 835 00:38:11,339 --> 00:38:19,180 its computer powered is like 836 00:38:15,730 --> 00:38:22,359 well but I mean but that it's an 837 00:38:19,180 --> 00:38:24,549 algorithm that it's I wouldn't say it's 838 00:38:22,359 --> 00:38:25,869 naive because in the facebook help 839 00:38:24,550 --> 00:38:27,640 section 840 00:38:25,869 --> 00:38:30,730 here's the question how does facebook 841 00:38:27,640 --> 00:38:32,560 determine topics and these sure you're 842 00:38:30,730 --> 00:38:34,810 gonna die today they don't mention 843 00:38:32,560 --> 00:38:37,029 anything about it being an algorithm or 844 00:38:34,810 --> 00:38:38,830 it being curated at all right and most 845 00:38:37,030 --> 00:38:40,450 people aren't going to read through this 846 00:38:38,830 --> 00:38:42,279 to cash or they're trying to focus on 847 00:38:40,450 --> 00:38:44,439 their lives and they're not necessarily 848 00:38:42,280 --> 00:38:47,260 diving in sharpie 849 00:38:44,440 --> 00:38:49,750 I'm just telling you that like I don't I 850 00:38:47,260 --> 00:38:52,960 don't want to live in a world in 2016 851 00:38:49,750 --> 00:38:55,720 where an algorithm a pure algorithm with 852 00:38:52,960 --> 00:38:57,250 no human oversight whatsoever is 853 00:38:55,720 --> 00:38:58,810 controlling everything that comes 854 00:38:57,250 --> 00:39:00,640 through my social media feeds because 855 00:38:58,810 --> 00:39:03,130 like who wrote the algorithm some 856 00:39:00,640 --> 00:39:05,589 squishy human with biases wrote that too 857 00:39:03,130 --> 00:39:07,270 and you know what there's like much less 858 00:39:05,590 --> 00:39:11,530 diversity of like thought and opinion 859 00:39:07,270 --> 00:39:13,180 and like background in those algorithm 860 00:39:11,530 --> 00:39:15,880 writers then there are amongst the 861 00:39:13,180 --> 00:39:17,649 editors and people who are like going 862 00:39:15,880 --> 00:39:22,119 through and picking out the trends and 863 00:39:17,650 --> 00:39:24,850 whatever and that might be one of the 864 00:39:22,119 --> 00:39:27,910 reasons why no one likes to talk about 865 00:39:24,850 --> 00:39:29,200 yeah i mean like that might be one of 866 00:39:27,910 --> 00:39:31,480 the reasons that no one likes to talk 867 00:39:29,200 --> 00:39:34,060 about the fact that their technology is 868 00:39:31,480 --> 00:39:36,400 not a hundred percent perfect and still 869 00:39:34,060 --> 00:39:38,470 turns up regular garbage if you look at 870 00:39:36,400 --> 00:39:42,730 what is the most popular on any social 871 00:39:38,470 --> 00:39:45,730 network i guarantee you it is gar bitch 872 00:39:42,730 --> 00:39:47,350 yeah that's a good point we need the 873 00:39:45,730 --> 00:39:50,080 curators we do 874 00:39:47,350 --> 00:39:52,660 that's a good point hello first of all 875 00:39:50,080 --> 00:39:54,130 before i get to my point i want to 876 00:39:52,660 --> 00:39:56,379 commit you for saying that 877 00:39:54,130 --> 00:39:57,730 overall people should be able to have 878 00:39:56,380 --> 00:39:59,680 different political views without 879 00:39:57,730 --> 00:40:01,150 jumping on each other and getting all 880 00:39:59,680 --> 00:40:03,700 crazy whether you're for Trump Hillary 881 00:40:01,150 --> 00:40:09,280 whoever thank you yeah that's just 882 00:40:03,700 --> 00:40:11,710 unfair you know i'm a registered into 883 00:40:09,280 --> 00:40:13,180 and you're very patient with the first 884 00:40:11,710 --> 00:40:15,940 guy by the way to I commend you for that 885 00:40:13,180 --> 00:40:18,190 but I'm registered independent and you 886 00:40:15,940 --> 00:40:20,230 know I don't follow parties i kinda go 887 00:40:18,190 --> 00:40:24,280 with who I like at the time you know 888 00:40:20,230 --> 00:40:27,970 concerns the country at the time but you 889 00:40:24,280 --> 00:40:30,400 know I signed on LOL i still have a well 890 00:40:27,970 --> 00:40:32,980 and there's a giant picture Hillary time 891 00:40:30,400 --> 00:40:35,110 with her and I don't get that's their 892 00:40:32,980 --> 00:40:37,630 right to do it i get on facebook and I 893 00:40:35,110 --> 00:40:39,400 have you know i know my feet I get my 894 00:40:37,630 --> 00:40:42,370 friends and things were into movies and 895 00:40:39,400 --> 00:40:46,330 stuff but the sides are inundated with 896 00:40:42,370 --> 00:40:49,450 you know since reflect you it doesn't 897 00:40:46,330 --> 00:40:51,549 reflect me so obviously they're not it's 898 00:40:49,450 --> 00:40:53,740 not being catered to me so it's not 899 00:40:51,550 --> 00:40:55,990 being catered to me what why is it all 900 00:40:53,740 --> 00:40:58,450 stupid things like Laura Ingraham giving 901 00:40:55,990 --> 00:41:01,270 his Nazi salute you know Republicans 902 00:40:58,450 --> 00:41:02,620 must be racist or you know or even you 903 00:41:01,270 --> 00:41:04,870 know even when you get stuff about the 904 00:41:02,620 --> 00:41:07,390 Democrats are Hillary's you know jacket 905 00:41:04,870 --> 00:41:10,569 cost six thousand dollars it's just 906 00:41:07,390 --> 00:41:12,759 stupidity and I think it it wasn't 907 00:41:10,570 --> 00:41:14,980 fabulous it was like it looked like 908 00:41:12,760 --> 00:41:17,230 mount state owns a jacket back in the 909 00:41:14,980 --> 00:41:19,690 day it's what the curators seems to be 910 00:41:17,230 --> 00:41:21,820 newsworthy what is what I look at that 911 00:41:19,690 --> 00:41:24,310 don't I see like like facebook like you 912 00:41:21,820 --> 00:41:25,900 know if you're if you're an adult who is 913 00:41:24,310 --> 00:41:28,240 smart enough to like you know not be 914 00:41:25,900 --> 00:41:30,400 affected by what facebook tells you what 915 00:41:28,240 --> 00:41:33,850 about the youth you know I think in 916 00:41:30,400 --> 00:41:36,850 general in the media music all facets 917 00:41:33,850 --> 00:41:39,970 there's a certain slant to one party 918 00:41:36,850 --> 00:41:41,710 compared to another and you know the 919 00:41:39,970 --> 00:41:44,529 slant is kind of fortified with that 920 00:41:41,710 --> 00:41:46,750 notion that well if you follow Trump you 921 00:41:44,530 --> 00:41:49,090 must you know be against gays you must 922 00:41:46,750 --> 00:41:51,310 be against african-americans and I think 923 00:41:49,090 --> 00:41:54,880 that is that kind of it kind of like 924 00:41:51,310 --> 00:41:56,259 subliminally or subconsciously sustains 925 00:41:54,880 --> 00:41:58,480 that kind of schism that we have going 926 00:41:56,260 --> 00:42:00,340 on now it doesn't do any good so whether 927 00:41:58,480 --> 00:42:02,590 it's going to help Trump or help Hillary 928 00:42:00,340 --> 00:42:04,450 in general you know it's their right to 929 00:42:02,590 --> 00:42:07,180 do it but it's just it's troublesome 930 00:42:04,450 --> 00:42:09,069 it's concerning maybe it's hurting us in 931 00:42:07,180 --> 00:42:11,200 terms of our stereotypes of each other I 932 00:42:09,070 --> 00:42:12,610 definitely think so cuz I you know I 933 00:42:11,200 --> 00:42:14,890 can't wait till this election is over 934 00:42:12,610 --> 00:42:16,210 just because my friends 935 00:42:14,890 --> 00:42:20,029 it's like everybody's hating each other 936 00:42:16,210 --> 00:42:24,770 well thank you thank you very much 937 00:42:20,030 --> 00:42:26,030 I thanks for talking i'm just wondering 938 00:42:24,770 --> 00:42:27,050 you talked a little about the 939 00:42:26,030 --> 00:42:28,160 peer-to-peer thing and I know that was 940 00:42:27,050 --> 00:42:29,720 kind of touched on with the filter 941 00:42:28,160 --> 00:42:31,040 bubble thing i'm just wondering if you 942 00:42:29,720 --> 00:42:33,020 thought more if there's any like way of 943 00:42:31,040 --> 00:42:35,090 sort of getting the critical mass for an 944 00:42:33,020 --> 00:42:36,440 alternative to a decent return to 945 00:42:35,090 --> 00:42:38,270 facebook or do you think that's just 946 00:42:36,440 --> 00:42:40,580 sort of too difficult to the only thing 947 00:42:38,270 --> 00:42:42,290 we can do is to just put pressure on 948 00:42:40,580 --> 00:42:43,880 facebook and twitter to improve and I 949 00:42:42,290 --> 00:42:45,890 don't want to say it's too difficult I 950 00:42:43,880 --> 00:42:47,900 think that it's absolutely worthwhile 951 00:42:45,890 --> 00:42:50,359 so the aspirant tried it and isn't doing 952 00:42:47,900 --> 00:42:52,460 all that well well you know it and I was 953 00:42:50,360 --> 00:42:55,070 making social apps for a long time I 954 00:42:52,460 --> 00:42:57,710 stopped and and but the reason why I did 955 00:42:55,070 --> 00:43:00,110 it was because I thought you could get 956 00:42:57,710 --> 00:43:01,730 new companies and new social networks 957 00:43:00,110 --> 00:43:04,670 off the ground and i still believe that 958 00:43:01,730 --> 00:43:07,700 we can it just it's a really big 959 00:43:04,670 --> 00:43:12,170 endeavor you have to find some way to 960 00:43:07,700 --> 00:43:14,120 get so many people on it right like if 961 00:43:12,170 --> 00:43:16,640 you really want it to take off and 962 00:43:14,120 --> 00:43:20,420 sustained as a platform it has to be 963 00:43:16,640 --> 00:43:23,450 appealing to a wide group of people and 964 00:43:20,420 --> 00:43:27,170 that's tough but I think that we may get 965 00:43:23,450 --> 00:43:28,640 their Google+ and how am I gonna do I i 966 00:43:27,170 --> 00:43:31,460 think that's the challenge right because 967 00:43:28,640 --> 00:43:34,250 I'm so many people here at Hope have the 968 00:43:31,460 --> 00:43:36,230 technical ability and and those 969 00:43:34,250 --> 00:43:38,390 platforms in the open source code to 970 00:43:36,230 --> 00:43:41,090 build them that's already out there 971 00:43:38,390 --> 00:43:44,390 the tricky pieces like how do we make it 972 00:43:41,090 --> 00:43:46,670 stick so we can have something that's 973 00:43:44,390 --> 00:43:49,009 more ours and and it's a really hard 974 00:43:46,670 --> 00:43:52,310 problem but i do think that we could get 975 00:43:49,010 --> 00:43:54,020 there you know I really do and things 976 00:43:52,310 --> 00:43:55,160 are changing very fast look at the world 977 00:43:54,020 --> 00:43:56,870 look at the things that you never 978 00:43:55,160 --> 00:43:58,490 thought would have happened we can for 979 00:43:56,870 --> 00:44:02,089 sure make a new social network i'm not 980 00:43:58,490 --> 00:44:05,959 sure how just yet but I think that's 981 00:44:02,090 --> 00:44:08,570 absolutely a worthwhile topic for us to 982 00:44:05,960 --> 00:44:10,820 keep noodling on and i have to believe 983 00:44:08,570 --> 00:44:12,320 it's possible you know I like to think 984 00:44:10,820 --> 00:44:14,210 that the future's going to do you think 985 00:44:12,320 --> 00:44:15,500 we can go to another one that's not just 986 00:44:14,210 --> 00:44:17,990 like that doesn't just become another 987 00:44:15,500 --> 00:44:19,790 corporate facebook and twitter and let 988 00:44:17,990 --> 00:44:21,560 ya like building a decentralized 989 00:44:19,790 --> 00:44:24,410 building like another corporate one is 990 00:44:21,560 --> 00:44:27,320 no can't ya because that the thing is 991 00:44:24,410 --> 00:44:29,230 you look at google rate this is actually 992 00:44:27,320 --> 00:44:31,869 really important to me 993 00:44:29,230 --> 00:44:35,200 our diocese what if so you remember the 994 00:44:31,869 --> 00:44:38,410 Kickstarter maybe hopefully it was a 995 00:44:35,200 --> 00:44:41,500 while ago about dice para they built a 996 00:44:38,410 --> 00:44:42,700 centralized showing that morning I mean 997 00:44:41,500 --> 00:44:46,869 they feel that they built it but did 998 00:44:42,700 --> 00:44:48,640 anybody come but there are people on 999 00:44:46,869 --> 00:44:51,700 even though I killed himself 1000 00:44:48,640 --> 00:44:54,670 oh I would argue this about what we 1001 00:44:51,700 --> 00:44:58,029 could do if you look at google google 1002 00:44:54,670 --> 00:44:59,830 became evil you could argue it certainly 1003 00:44:58,030 --> 00:45:03,100 created problems that it didn't 1004 00:44:59,830 --> 00:45:05,020 anticipate the google solve many of the 1005 00:45:03,100 --> 00:45:07,390 problems we wanted to solve in the early 1006 00:45:05,020 --> 00:45:09,580 days in terms of like what a work 1007 00:45:07,390 --> 00:45:12,910 environment could be like what what 1008 00:45:09,580 --> 00:45:16,180 google did so many wonderful things they 1009 00:45:12,910 --> 00:45:17,770 just they didn't foresee the dark side 1010 00:45:16,180 --> 00:45:19,868 and I think that's what we can do I 1011 00:45:17,770 --> 00:45:21,550 think we can solve this and we're gonna 1012 00:45:19,869 --> 00:45:24,100 mess up in some way that we won't 1013 00:45:21,550 --> 00:45:27,010 foresee but it's still like a process of 1014 00:45:24,100 --> 00:45:29,410 evolution and things get better so like 1015 00:45:27,010 --> 00:45:31,750 I'm down for that i'm down for us to 1016 00:45:29,410 --> 00:45:33,100 solve this problem unintentionally 1017 00:45:31,750 --> 00:45:34,600 create something that we can foresee 1018 00:45:33,100 --> 00:45:40,060 because it didn't exist yet 1019 00:45:34,600 --> 00:45:43,990 and then it keeps going 1020 00:45:40,060 --> 00:45:45,880 so like also like other things that they 1021 00:45:43,990 --> 00:45:48,100 do that really got is google Summer of 1022 00:45:45,880 --> 00:45:49,720 Code what gets college student 1023 00:45:48,100 --> 00:45:52,359 yeah and then they have google code in 1024 00:45:49,720 --> 00:45:54,100 which actually gets a doula it reaches 1025 00:45:52,359 --> 00:45:57,940 out to middle school to high school 1026 00:45:54,100 --> 00:45:59,770 students have been thinking get any car 1027 00:45:57,940 --> 00:46:02,350 so yeah you want to change the way 1028 00:45:59,770 --> 00:46:04,330 Google and Facebook do their algorithms 1029 00:46:02,350 --> 00:46:07,660 I have a very simple solution you know 1030 00:46:04,330 --> 00:46:10,270 go work for them they're hiring both of 1031 00:46:07,660 --> 00:46:19,779 them as the street I think although they 1032 00:46:10,270 --> 00:46:21,280 get no no no no that that is you 1033 00:46:19,780 --> 00:46:24,490 obviously were not on the debate team 1034 00:46:21,280 --> 00:46:25,930 that was an idiotic remark if you want 1035 00:46:24,490 --> 00:46:31,600 to counter your welcome to come up to 1036 00:46:25,930 --> 00:46:33,520 the mic so i challenge I want to 1037 00:46:31,600 --> 00:46:38,710 challenge the original assumption that 1038 00:46:33,520 --> 00:46:41,530 facebook even has a responsibility or we 1039 00:46:38,710 --> 00:46:42,789 even have a the right to demand of them 1040 00:46:41,530 --> 00:46:44,650 to be neutral 1041 00:46:42,789 --> 00:46:46,419 you know you look at things that are 1042 00:46:44,650 --> 00:46:48,130 supposed to be neutral because their 1043 00:46:46,419 --> 00:46:49,989 journal that they claim to be 1044 00:46:48,130 --> 00:46:52,869 generalists organizations like folks 1045 00:46:49,989 --> 00:46:55,329 news who are anything but no they do is 1046 00:46:52,869 --> 00:46:58,779 promote hatred and nobody's actually 1047 00:46:55,329 --> 00:47:02,049 taking them up to the task so why should 1048 00:46:58,779 --> 00:47:06,369 somebody who doesn't even have the UH 1049 00:47:02,049 --> 00:47:09,069 the the uh I would be more comfortable 1050 00:47:06,369 --> 00:47:11,829 at Facebook admitted any bias that it 1051 00:47:09,069 --> 00:47:14,859 has that's my point but but don't but is 1052 00:47:11,829 --> 00:47:16,839 why is your comfort even consider why 1053 00:47:14,859 --> 00:47:19,119 should your comfort be a consideration 1054 00:47:16,839 --> 00:47:21,640 for facebook if you don't like Facebook 1055 00:47:19,119 --> 00:47:24,249 do what the the previous speaker said 1056 00:47:21,640 --> 00:47:24,669 don't be on Facebook and you know you 1057 00:47:24,249 --> 00:47:26,979 can 1058 00:47:24,669 --> 00:47:30,519 there is life outside of facebook fair 1059 00:47:26,979 --> 00:47:33,009 enough next question 1060 00:47:30,519 --> 00:47:34,089 so where I I've been thinking about this 1061 00:47:33,009 --> 00:47:35,439 a little bit and hearing about all the 1062 00:47:34,089 --> 00:47:37,869 curfuffle and it seemed like a lot of 1063 00:47:35,439 --> 00:47:39,578 people that we're talking about it kind 1064 00:47:37,869 --> 00:47:41,799 of warrant really techie there was no 1065 00:47:39,579 --> 00:47:43,689 they're saying things like you can't 1066 00:47:41,799 --> 00:47:44,349 change the algorithm at all you know 1067 00:47:43,689 --> 00:47:45,609 ever 1068 00:47:44,349 --> 00:47:47,499 yeah it was very much like you know 1069 00:47:45,609 --> 00:47:50,529 whatever facebook did last Tuesday well 1070 00:47:47,499 --> 00:47:52,359 that was perfect and that kind of didn't 1071 00:47:50,529 --> 00:47:53,619 seem to go along with how we have tech 1072 00:47:52,359 --> 00:47:55,509 companies now that you know they're 1073 00:47:53,619 --> 00:47:57,939 doing a be test their changing little 1074 00:47:55,509 --> 00:47:59,890 things all the time and facebook is 1075 00:47:57,939 --> 00:48:01,719 running tests all the time yeah and 1076 00:47:59,890 --> 00:48:03,489 every you know everything you know 1077 00:48:01,719 --> 00:48:04,989 twitter is an every big you know tech 1078 00:48:03,489 --> 00:48:07,209 companies like how they work 1079 00:48:04,989 --> 00:48:09,699 so what and I was saying about the mom 1080 00:48:07,209 --> 00:48:11,828 almost not sure what do you see what 1081 00:48:09,699 --> 00:48:13,119 would have more transparent facebook be 1082 00:48:11,829 --> 00:48:16,209 how would they be able to communicate 1083 00:48:13,119 --> 00:48:18,880 that you know easily to people that are 1084 00:48:16,209 --> 00:48:21,489 tacky and non-techie well if they 1085 00:48:18,880 --> 00:48:24,009 removed the curated section that would 1086 00:48:21,489 --> 00:48:27,219 remove a lot of this problem off the bat 1087 00:48:24,009 --> 00:48:30,069 because i think the way that they 1088 00:48:27,219 --> 00:48:32,380 handled the algorithm for your feed is 1089 00:48:30,069 --> 00:48:34,269 generally more transparent and less 1090 00:48:32,380 --> 00:48:36,759 problematic people don't seem concerned 1091 00:48:34,269 --> 00:48:38,589 about what showing up in their feet 1092 00:48:36,759 --> 00:48:40,719 outside of search engine optimization 1093 00:48:38,589 --> 00:48:42,939 for example if you're a non-profit 1094 00:48:40,719 --> 00:48:44,529 brands it's hard for you to show up in 1095 00:48:42,939 --> 00:48:49,328 the feed but that's not really 1096 00:48:44,529 --> 00:48:52,179 politicized and and you can choose what 1097 00:48:49,329 --> 00:48:54,440 shows up in your feet so for me like i 1098 00:48:52,179 --> 00:48:55,700 have my top 10 or 20 friends and I've 1099 00:48:54,440 --> 00:48:58,579 wasn't that they should get shown first 1100 00:48:55,700 --> 00:49:01,220 and they do get shown first and also 1101 00:48:58,579 --> 00:49:04,250 know that seems to work well enough not 1102 00:49:01,220 --> 00:49:06,439 perfect but well enough 1103 00:49:04,250 --> 00:49:09,170 it's these moments where there's 1104 00:49:06,440 --> 00:49:15,319 curation against a claim of neutrality 1105 00:49:09,170 --> 00:49:18,050 that I'm seeing uh it's just BS thank 1106 00:49:15,319 --> 00:49:20,359 you 1107 00:49:18,050 --> 00:49:21,890 hello first off I'd like to thank you 1108 00:49:20,359 --> 00:49:24,680 for being patient with this somewhat 1109 00:49:21,890 --> 00:49:27,230 rowdy group of people i'm already group 1110 00:49:24,680 --> 00:49:29,149 of people good good you're good company 1111 00:49:27,230 --> 00:49:31,910 so are you looking for these companies 1112 00:49:29,150 --> 00:49:33,170 to be completely neutral or maybe it's 1113 00:49:31,910 --> 00:49:34,910 ok if they have a stance but they're 1114 00:49:33,170 --> 00:49:37,250 just more transparent about what it is 1115 00:49:34,910 --> 00:49:39,828 oh well you know I'm coming out of 1116 00:49:37,250 --> 00:49:42,020 having worked in more open source 1117 00:49:39,829 --> 00:49:44,300 software or the last few years and I 1118 00:49:42,020 --> 00:49:46,910 like transparency I think that 1119 00:49:44,300 --> 00:49:48,710 transparency is and it's a nice practice 1120 00:49:46,910 --> 00:49:50,420 you know you make mistakes but you're 1121 00:49:48,710 --> 00:49:53,089 open about it and we can have 1122 00:49:50,420 --> 00:49:55,040 conversations about it as a community so 1123 00:49:53,089 --> 00:49:57,980 that would be my preference my 1124 00:49:55,040 --> 00:50:00,259 preference would be for companies to 1125 00:49:57,980 --> 00:50:03,230 just admit that they're imperfect and 1126 00:50:00,260 --> 00:50:05,420 it's a process and solicit feedback and 1127 00:50:03,230 --> 00:50:07,400 for us to say yeah we'd rather have you 1128 00:50:05,420 --> 00:50:09,710 curate and we understand that humans are 1129 00:50:07,400 --> 00:50:11,210 messy and the curations not perfect and 1130 00:50:09,710 --> 00:50:14,089 for facebook to say something like 1131 00:50:11,210 --> 00:50:16,940 neutrality is a value but it's like we 1132 00:50:14,089 --> 00:50:20,630 get as close as we can I think something 1133 00:50:16,940 --> 00:50:22,069 like that would probably I'd probably 1134 00:50:20,630 --> 00:50:25,069 like that a lot 1135 00:50:22,069 --> 00:50:26,930 um it'sit's the conflict between when a 1136 00:50:25,069 --> 00:50:28,069 company says a thing and it's just know 1137 00:50:26,930 --> 00:50:30,799 what they're doing that's what troubles 1138 00:50:28,069 --> 00:50:32,300 me I'm also just troubled by the 1139 00:50:30,800 --> 00:50:34,490 enormous power that these companies have 1140 00:50:32,300 --> 00:50:37,310 an elections and because of the lack of 1141 00:50:34,490 --> 00:50:39,439 transparency we just don't know i think 1142 00:50:37,310 --> 00:50:41,390 it's unlikely that Facebook or Twitter 1143 00:50:39,440 --> 00:50:44,420 will swing the selection or certainly 1144 00:50:41,390 --> 00:50:47,598 that they'll try but it remains true 1145 00:50:44,420 --> 00:50:51,290 that they could and I think that's worth 1146 00:50:47,599 --> 00:50:53,030 it's worth our attention and that's what 1147 00:50:51,290 --> 00:50:55,250 did I answer your question yeah also 1148 00:50:53,030 --> 00:50:57,349 follow up which is using neutrality is a 1149 00:50:55,250 --> 00:51:00,410 good value is there some limited 1150 00:50:57,349 --> 00:51:02,030 neutrality like we we can agree that if 1151 00:51:00,410 --> 00:51:04,040 they choose Democrats Republicans that's 1152 00:51:02,030 --> 00:51:06,020 a bummer but should they be censoring 1153 00:51:04,040 --> 00:51:07,710 Isis posts like is there anything that's 1154 00:51:06,020 --> 00:51:09,930 so objectionable that is at 1155 00:51:07,710 --> 00:51:12,180 equally desirable for them to step in 1156 00:51:09,930 --> 00:51:14,220 and say no we have an opinion on this 1157 00:51:12,180 --> 00:51:15,868 content with this viewpoint when i take 1158 00:51:14,220 --> 00:51:17,520 a step back and I think about facebook 1159 00:51:15,869 --> 00:51:20,400 and i think about Twitter I'm seeing 1160 00:51:17,520 --> 00:51:22,230 things that I've selected to see so I 1161 00:51:20,400 --> 00:51:23,940 shouldn't see Isis and left i've 1162 00:51:22,230 --> 00:51:25,349 selected to see them and then we get 1163 00:51:23,940 --> 00:51:27,570 into the question of whether facebook 1164 00:51:25,349 --> 00:51:32,460 should be hosting that content and 1165 00:51:27,570 --> 00:51:35,520 that's something that I'm unsure about 1166 00:51:32,460 --> 00:51:38,430 that's a free-speech question that I'm 1167 00:51:35,520 --> 00:51:40,140 not Eddie's with yet myself and I'm 1168 00:51:38,430 --> 00:51:43,529 interested in what other people have to 1169 00:51:40,140 --> 00:51:44,940 say I if I ran one of those companies i 1170 00:51:43,530 --> 00:51:47,910 wouldn't want to host certain kinds of 1171 00:51:44,940 --> 00:51:49,830 speech but it is difficult to figure out 1172 00:51:47,910 --> 00:51:52,020 where you draw certain lines I suppose 1173 00:51:49,830 --> 00:51:54,210 Isis might be an easy line to draw other 1174 00:51:52,020 --> 00:51:57,420 lines are less easy 1175 00:51:54,210 --> 00:52:00,480 um yeah these are good questions 1176 00:51:57,420 --> 00:52:03,420 yeah I thank you thank you 1177 00:52:00,480 --> 00:52:10,920 this topic seems to attract a lot of 1178 00:52:03,420 --> 00:52:12,150 man's planning I to add onto that not I 1179 00:52:10,920 --> 00:52:14,849 wasn't calling on anyone in particular 1180 00:52:12,150 --> 00:52:16,920 what maybe my whole talk with 1181 00:52:14,849 --> 00:52:20,940 mansplaining and I apologize for that 1182 00:52:16,920 --> 00:52:24,750 no huh arms so my question was to add to 1183 00:52:20,940 --> 00:52:26,369 someone who asked earlier about how out 1184 00:52:24,750 --> 00:52:29,520 the ribbons can have bias and i wanted 1185 00:52:26,369 --> 00:52:32,790 to add onto that and say like algorithms 1186 00:52:29,520 --> 00:52:34,740 are you know often like simulation so if 1187 00:52:32,790 --> 00:52:39,000 you think about i was reading an article 1188 00:52:34,740 --> 00:52:41,069 about the algorithms used to predict 1189 00:52:39,000 --> 00:52:44,310 whether someone will wind up in jail 1190 00:52:41,070 --> 00:52:47,520 again after being released and it was 1191 00:52:44,310 --> 00:52:50,430 was really interesting to see how like 1192 00:52:47,520 --> 00:52:51,690 the inputs into an algorithm really 1193 00:52:50,430 --> 00:52:54,960 affect the outcome and can actually 1194 00:52:51,690 --> 00:52:56,280 cause like racism or bias in one way or 1195 00:52:54,960 --> 00:52:58,200 another depending on how the algorithm 1196 00:52:56,280 --> 00:53:01,410 is crafted so do you have any ideas 1197 00:52:58,200 --> 00:53:02,730 about how we can ask for transparency 1198 00:53:01,410 --> 00:53:05,250 especially from like government 1199 00:53:02,730 --> 00:53:07,589 institutions about the like machine 1200 00:53:05,250 --> 00:53:09,599 learning algorithms or or the inputs 1201 00:53:07,589 --> 00:53:11,910 that they use into those algorithms uh 1202 00:53:09,599 --> 00:53:14,700 to to solve these problems that are 1203 00:53:11,910 --> 00:53:17,098 really complicated and you know they're 1204 00:53:14,700 --> 00:53:19,290 uh you know a machine so how can machine 1205 00:53:17,099 --> 00:53:20,790 have bias but the inputs into those 1206 00:53:19,290 --> 00:53:22,740 systems 1207 00:53:20,790 --> 00:53:24,810 so I as the cheeks who i've been working 1208 00:53:22,740 --> 00:53:25,950 with lon press stuff is giving a talk 1209 00:53:24,810 --> 00:53:28,259 tomorrow 1210 00:53:25,950 --> 00:53:30,210 uh that deals with them how they make 1211 00:53:28,260 --> 00:53:34,350 changes in oakland around 1212 00:53:30,210 --> 00:53:36,720 anti-surveillance and oh hi yeah so I 1213 00:53:34,350 --> 00:53:38,310 sunday my apologies not tomorrow sunday 1214 00:53:36,720 --> 00:53:41,549 so i recommend going and seeing that 1215 00:53:38,310 --> 00:53:43,560 talk to see how citizens can impact that 1216 00:53:41,550 --> 00:53:45,450 sort of thing 1217 00:53:43,560 --> 00:53:47,790 what I've seen specifically an open 1218 00:53:45,450 --> 00:53:49,230 source and transparency is the election 1219 00:53:47,790 --> 00:53:50,640 of certain candidates who promote those 1220 00:53:49,230 --> 00:53:55,500 values and that's starting to happen 1221 00:53:50,640 --> 00:53:57,660 then Kalos in Manhattan is one example 1222 00:53:55,500 --> 00:54:00,090 Rand Paul was interested in open source 1223 00:53:57,660 --> 00:54:03,029 when i was working with him so I think 1224 00:54:00,090 --> 00:54:05,460 we can vote for those candidates and 1225 00:54:03,030 --> 00:54:07,770 make it clear that we care and perhaps 1226 00:54:05,460 --> 00:54:09,240 maybe like become a voting bloc is 1227 00:54:07,770 --> 00:54:11,250 hackers or something like I think that 1228 00:54:09,240 --> 00:54:14,100 there might be political ways that we 1229 00:54:11,250 --> 00:54:17,580 can communicate to the reps that we care 1230 00:54:14,100 --> 00:54:19,140 and i think we can become louder on 1231 00:54:17,580 --> 00:54:20,490 specific issues like if people are 1232 00:54:19,140 --> 00:54:22,770 getting more jail time because an 1233 00:54:20,490 --> 00:54:24,899 algorithm determines that that seems 1234 00:54:22,770 --> 00:54:26,820 like something that citizens should be 1235 00:54:24,900 --> 00:54:29,490 able to see all right 1236 00:54:26,820 --> 00:54:34,020 no thanks thank you i think we have time 1237 00:54:29,490 --> 00:54:40,640 for one more question no one comes up 1238 00:54:34,020 --> 00:54:40,640 alright thank you so much