1 00:00:15,600 --> 00:00:16,560 Speaker 1: Welcome to tech Stuff. 2 00:00:16,560 --> 00:00:18,759 Speaker 2: I'm as Voloshan, and today I want to share a 3 00:00:18,760 --> 00:00:24,000 Speaker 2: fantastic interview that Kara recorded with the documentary director Adam 4 00:00:24,079 --> 00:00:28,440 Speaker 2: Balla Lowe about his latest project, Deep Faking Sam Altman. 5 00:00:29,200 --> 00:00:31,200 Speaker 2: The film follows Adam as he attempts to score an 6 00:00:31,200 --> 00:00:35,080 Speaker 2: interview with the CEO of Open Ai, but when that 7 00:00:35,200 --> 00:00:39,400 Speaker 2: starts to feel kind of impossible, Adam decides he wants 8 00:00:39,440 --> 00:00:43,960 Speaker 2: the next best thing, an interview with Sam Altman's deep fake, 9 00:00:44,760 --> 00:00:48,400 Speaker 2: so he makes one. It's a fascinating watch that takes 10 00:00:48,440 --> 00:00:50,960 Speaker 2: you from the gates of Open AI all the way 11 00:00:50,960 --> 00:00:54,800 Speaker 2: to India and back again. When Adam sends his sam 12 00:00:54,800 --> 00:00:58,440 Speaker 2: Altman deep fake directly to Sam Altman. 13 00:00:58,440 --> 00:00:59,000 Speaker 1: The real one. 14 00:01:00,080 --> 00:01:03,400 Speaker 2: Enjoy this conversation between Carat Price and Adam Balla Lowe. 15 00:01:06,720 --> 00:01:10,000 Speaker 3: So you set out to make this documentary about AI. 16 00:01:10,480 --> 00:01:13,520 Speaker 3: How did you end up focusing on Sam Altman and 17 00:01:13,600 --> 00:01:16,640 Speaker 3: why are you drawn to him particularly as a subject 18 00:01:16,640 --> 00:01:17,680 Speaker 3: of a movie about AI. 19 00:01:18,600 --> 00:01:22,200 Speaker 4: Yeah, Sam Altman was really an entry point early on, 20 00:01:23,200 --> 00:01:26,720 Speaker 4: and in a lot of ways, like I believe that 21 00:01:26,880 --> 00:01:31,000 Speaker 4: he's the guy that's ferrying us into this future through 22 00:01:31,080 --> 00:01:36,200 Speaker 4: Open AI, and so he, of all people, would be 23 00:01:36,240 --> 00:01:38,679 Speaker 4: able to answer the hard questions that I wanted to 24 00:01:38,760 --> 00:01:39,320 Speaker 4: ask him. 25 00:01:39,600 --> 00:01:42,400 Speaker 3: When you say ferrying us into this future, what's the 26 00:01:42,440 --> 00:01:43,839 Speaker 3: future that you're talking about. 27 00:01:44,440 --> 00:01:48,600 Speaker 4: It's a future where we're completely enmeshed with artificial intelligence, 28 00:01:48,600 --> 00:01:51,200 Speaker 4: where it's as much a part of our daily lives 29 00:01:51,200 --> 00:01:52,200 Speaker 4: as the Internet is. 30 00:01:52,280 --> 00:01:54,120 Speaker 3: Do you not think that we're already there? Is it 31 00:01:54,160 --> 00:01:54,920 Speaker 3: really the future? 32 00:01:55,240 --> 00:01:55,360 Speaker 1: No? 33 00:01:55,400 --> 00:01:58,240 Speaker 4: I don't think we're completely there yet. I mean I 34 00:01:58,280 --> 00:02:01,720 Speaker 4: think that certainly people on you know, the coasts and 35 00:02:01,800 --> 00:02:05,120 Speaker 4: like more privileged people are, but like the majority of 36 00:02:05,200 --> 00:02:10,000 Speaker 4: Americans aren't, Frankly, and we also see like that it 37 00:02:10,360 --> 00:02:13,560 Speaker 4: took time for the for a lot of people in 38 00:02:13,600 --> 00:02:17,120 Speaker 4: America to adopt to the Internet too, because you know, 39 00:02:17,160 --> 00:02:19,440 Speaker 4: it just wasn't accessible to them, Like I mean, I 40 00:02:19,480 --> 00:02:21,720 Speaker 4: even remember it as little as ten years ago, people 41 00:02:21,720 --> 00:02:24,240 Speaker 4: still having to go to the library, like relatives of 42 00:02:24,280 --> 00:02:28,359 Speaker 4: mine and especially you know, my family in India still 43 00:02:28,360 --> 00:02:30,400 Speaker 4: having to go to like library and stuff like that 44 00:02:30,400 --> 00:02:31,480 Speaker 4: to access the Internet. 45 00:02:31,520 --> 00:02:35,480 Speaker 3: So you tried to actually interview Sam Altman and talk 46 00:02:35,520 --> 00:02:37,600 Speaker 3: to a lot of people who were like, yeah, fucking right, 47 00:02:38,320 --> 00:02:40,360 Speaker 3: and basically we're like, you know, even if you do 48 00:02:40,520 --> 00:02:42,840 Speaker 3: talk to him, he's going to be an expert who's 49 00:02:42,880 --> 00:02:46,760 Speaker 3: telling you what you want to hear. You know, before 50 00:02:47,639 --> 00:02:49,840 Speaker 3: making this movie, was there a sort of myth that 51 00:02:49,880 --> 00:02:52,600 Speaker 3: you had about him in your mind that you were 52 00:02:52,639 --> 00:02:55,519 Speaker 3: trying to kind of at least contend with or debunk. 53 00:02:56,360 --> 00:02:59,560 Speaker 4: Yeah. The initial question was like that that I even 54 00:02:59,680 --> 00:03:03,360 Speaker 4: kind of pitched to the producers and the financiers. Was 55 00:03:03,680 --> 00:03:06,359 Speaker 4: is Sam Altman a hero or a villain? And then 56 00:03:06,400 --> 00:03:09,000 Speaker 4: there was the New York Magazine article by Liz Wilde 57 00:03:09,000 --> 00:03:13,240 Speaker 4: that I optioned Sam Altman is the Oppenheimer of our age, 58 00:03:13,840 --> 00:03:17,760 Speaker 4: So that was another question, was like is he the Oppenheimer? 59 00:03:17,840 --> 00:03:21,920 Speaker 4: Like I didn't necessarily agree with Liz right off the bat. 60 00:03:22,000 --> 00:03:25,120 Speaker 4: I wanted to know more, and I know that her 61 00:03:25,240 --> 00:03:28,480 Speaker 4: article was quite controversial, but yeah, that was that was 62 00:03:28,520 --> 00:03:30,920 Speaker 4: another question that I that I wanted to kind of, 63 00:03:31,680 --> 00:03:34,960 Speaker 4: you know, answer through through making the doc, through interviewing him, 64 00:03:35,200 --> 00:03:37,880 Speaker 4: meeting him, interviewing other people at Open AI, et cetera. 65 00:03:38,480 --> 00:03:40,960 Speaker 3: So there's a sequence when you're trying to get into 66 00:03:41,000 --> 00:03:43,400 Speaker 3: the open AI building and no one will speak with you, 67 00:03:44,120 --> 00:03:47,360 Speaker 3: but beyond that point, no one would even confirm that 68 00:03:47,440 --> 00:03:50,360 Speaker 3: the building you were outside of was the Open AI building. 69 00:03:51,120 --> 00:03:54,240 Speaker 3: Like what was going on there? And then how did 70 00:03:54,280 --> 00:03:57,120 Speaker 3: you finally gain access to the company, Like, can you 71 00:03:57,200 --> 00:03:58,560 Speaker 3: explain the chase a little bit? 72 00:03:58,920 --> 00:04:01,760 Speaker 4: Yeah, what was going on there? I still have no idea. 73 00:04:01,840 --> 00:04:05,960 Speaker 4: I actually read an article another reporter tried to do 74 00:04:06,040 --> 00:04:08,720 Speaker 4: what I did and had the same reaction, like no 75 00:04:08,800 --> 00:04:11,160 Speaker 4: one would talk to her and no one would even 76 00:04:11,200 --> 00:04:13,480 Speaker 4: confirm that, oh, this is the building where open AI is. 77 00:04:13,560 --> 00:04:15,880 Speaker 4: Like I feel like if I rolled up on the CIA, 78 00:04:16,480 --> 00:04:18,800 Speaker 4: like a CIA agent would be like, yeah, yeah, that's 79 00:04:18,800 --> 00:04:20,920 Speaker 4: a CIA, You're not going to get in, but like, yeah, 80 00:04:20,960 --> 00:04:24,640 Speaker 4: that's that's what it is there. Like it was so weird. 81 00:04:25,600 --> 00:04:28,800 Speaker 4: It just felt like a cult, you know what I mean? 82 00:04:28,880 --> 00:04:32,200 Speaker 4: Like I was like, is it not? It is? Absolutely 83 00:04:32,360 --> 00:04:35,640 Speaker 4: it is a It is a frightening cult. And I 84 00:04:35,680 --> 00:04:38,520 Speaker 4: talked to four whistleblowers from it, and they they talked 85 00:04:38,640 --> 00:04:40,960 Speaker 4: like cult members that had gotten out of a cult. 86 00:04:41,080 --> 00:04:43,279 Speaker 4: Yeah they were, And you see two of them in 87 00:04:43,279 --> 00:04:45,520 Speaker 4: the film, and then two of them two others talked 88 00:04:45,520 --> 00:04:49,279 Speaker 4: to me off the record, but they seemed like traumatized 89 00:04:49,279 --> 00:04:52,200 Speaker 4: by their time at open AI. So yeah, it feels 90 00:04:52,240 --> 00:04:55,880 Speaker 4: like a cult. And to answer your question, I got 91 00:04:55,920 --> 00:04:57,760 Speaker 4: access and you could see it in the film. Finally, 92 00:04:57,760 --> 00:05:00,400 Speaker 4: when somebody opened the gate and I ran in before 93 00:05:00,440 --> 00:05:04,440 Speaker 4: it shut, and then the security guard immediately grabbed me 94 00:05:04,520 --> 00:05:06,600 Speaker 4: and physically removed me from the premises. 95 00:05:07,120 --> 00:05:10,359 Speaker 3: You know, you kind of then decide or understand that 96 00:05:10,400 --> 00:05:12,880 Speaker 3: he's not going to talk to anymore. And so this 97 00:05:12,920 --> 00:05:15,520 Speaker 3: is where the this is where the movie starts, right 98 00:05:15,600 --> 00:05:17,640 Speaker 3: you start to think outside the box? Can you talk 99 00:05:17,680 --> 00:05:19,760 Speaker 3: a little bit about what you did? 100 00:05:20,160 --> 00:05:23,000 Speaker 4: Yeah, I mean around the time that I came to 101 00:05:23,040 --> 00:05:25,760 Speaker 4: this realization, and I think it was like a hundred 102 00:05:25,839 --> 00:05:29,800 Speaker 4: something days since I had first contacted him. Around that 103 00:05:29,960 --> 00:05:37,280 Speaker 4: time was the infamous Scarlet Johansson controversy with chat Ept. 104 00:05:38,200 --> 00:05:41,279 Speaker 4: But when I heard about that that he had stolen 105 00:05:41,360 --> 00:05:45,240 Speaker 4: her voice, I was like, I can do this to him, 106 00:05:45,520 --> 00:05:48,560 Speaker 4: Like he's just given me like license to do this 107 00:05:48,600 --> 00:05:52,200 Speaker 4: to him, But let me take it one step further, 108 00:05:52,680 --> 00:05:55,159 Speaker 4: Like the voice cloning I was already doing. You could 109 00:05:55,160 --> 00:05:57,040 Speaker 4: see in the movie. I was already doing some voice 110 00:05:57,040 --> 00:06:02,279 Speaker 4: cloning and experimenting with it before that. But after I 111 00:06:02,360 --> 00:06:04,400 Speaker 4: got this idea that like, not only could I do 112 00:06:04,440 --> 00:06:07,279 Speaker 4: a voice clone, but I could do an interview with 113 00:06:07,360 --> 00:06:13,440 Speaker 4: an actor playing Sam Altman and create an LM a 114 00:06:13,520 --> 00:06:16,919 Speaker 4: sort of brain of Sam Altman to chat with, and 115 00:06:16,960 --> 00:06:19,200 Speaker 4: the actor would read the lines off a teleprompter so 116 00:06:19,240 --> 00:06:21,360 Speaker 4: that I would do the interview that way, and then 117 00:06:21,360 --> 00:06:23,760 Speaker 4: in post production I would just deep fake everything, like 118 00:06:23,800 --> 00:06:27,760 Speaker 4: deep fake Sam Altman's face onto the actor. So that 119 00:06:28,480 --> 00:06:30,440 Speaker 4: that was the initial idea. So it would be a 120 00:06:30,520 --> 00:06:35,920 Speaker 4: multi step process where first I'd create this LM, which was, 121 00:06:36,320 --> 00:06:40,800 Speaker 4: you know, pretty relatively easy to do, and then well, just. 122 00:06:40,760 --> 00:06:45,719 Speaker 3: Really quickly relatively easy. You have to basically create a 123 00:06:45,800 --> 00:06:51,000 Speaker 3: large language model that is realistic enough to represent a 124 00:06:51,040 --> 00:06:53,080 Speaker 3: Sam Altman that you're interviewing. So how do you go 125 00:06:53,120 --> 00:06:55,320 Speaker 3: about doing that? Just for people who are wondering. 126 00:06:55,839 --> 00:06:58,719 Speaker 4: Yeah, So I started out trying to find somebody in 127 00:06:58,720 --> 00:07:02,320 Speaker 4: the States to do it, and on film I went 128 00:07:02,360 --> 00:07:04,280 Speaker 4: to a couple of companies in the Bay Area and 129 00:07:04,360 --> 00:07:08,159 Speaker 4: talked to them about it and they said no. And 130 00:07:08,200 --> 00:07:11,280 Speaker 4: then just over email, our producers reached out to probably 131 00:07:11,400 --> 00:07:15,200 Speaker 4: twelve or so other companies who all said no. So 132 00:07:15,280 --> 00:07:17,680 Speaker 4: I had to go to India to get it done. 133 00:07:17,840 --> 00:07:20,120 Speaker 4: And I found this guy on YouTube who'd made this 134 00:07:20,240 --> 00:07:23,720 Speaker 4: video about Barbie. He'd made like a Barbie deep fake 135 00:07:23,760 --> 00:07:27,560 Speaker 4: where he deep faked Bollywood stars faces onto the actors 136 00:07:27,560 --> 00:07:30,320 Speaker 4: in Barbie and I reached out to him and he 137 00:07:30,440 --> 00:07:31,080 Speaker 4: was very good. 138 00:07:31,600 --> 00:07:33,040 Speaker 3: He was very good. 139 00:07:33,200 --> 00:07:35,560 Speaker 4: Yeah. I reached out to him and I said, hey, 140 00:07:35,600 --> 00:07:38,080 Speaker 4: can you do this for me? And he was like, yeah, 141 00:07:38,240 --> 00:07:40,400 Speaker 4: here's how we do it. And like we went to 142 00:07:40,440 --> 00:07:44,119 Speaker 4: India and I watched him do it, and long story short, 143 00:07:44,160 --> 00:07:46,760 Speaker 4: you know, he coded this LLM and he loaded it 144 00:07:46,800 --> 00:07:49,840 Speaker 4: in with everything that Sam Oltman has ever said or 145 00:07:49,880 --> 00:07:53,680 Speaker 4: written on the Internet. And obviously this works like like 146 00:07:54,080 --> 00:08:00,280 Speaker 4: way more with somebody who's a celebrity or public person, because. 147 00:08:00,360 --> 00:08:02,800 Speaker 3: It's the same thing with like voice replication, Right, It's 148 00:08:02,800 --> 00:08:04,440 Speaker 3: like the more you have, the better the thing. 149 00:08:05,000 --> 00:08:07,400 Speaker 4: Yeah, So if you have very little information about you 150 00:08:07,440 --> 00:08:10,120 Speaker 4: on the internet, like it's not unless you do like 151 00:08:10,480 --> 00:08:12,960 Speaker 4: kind of an interview with the person creating it and 152 00:08:13,000 --> 00:08:15,040 Speaker 4: you write all this stuff and answer all these questions, 153 00:08:15,560 --> 00:08:17,800 Speaker 4: that chatbot's not going to be great. But if there's 154 00:08:17,840 --> 00:08:20,480 Speaker 4: like tons of stuff out there. And also Altman, when 155 00:08:20,520 --> 00:08:22,320 Speaker 4: he was younger, and I think still to this day, 156 00:08:22,320 --> 00:08:24,640 Speaker 4: he blogged a lot, so he so he had a 157 00:08:24,680 --> 00:08:27,640 Speaker 4: lot of information that was first person that he would write. 158 00:08:27,680 --> 00:08:30,119 Speaker 4: So this thing could like get his voice at style 159 00:08:30,120 --> 00:08:32,840 Speaker 4: of writing and style of speaking also from like videos 160 00:08:32,840 --> 00:08:36,240 Speaker 4: and everything, so it was pretty damn good. You'll find 161 00:08:36,240 --> 00:08:39,160 Speaker 4: this interesting. We started out dev the Indian deep figure 162 00:08:39,280 --> 00:08:42,680 Speaker 4: like Deb Sing who made the chatbot with me in India. 163 00:08:43,559 --> 00:08:47,120 Speaker 4: We initially had based on chat gpt, but we found 164 00:08:47,160 --> 00:08:50,840 Speaker 4: that chat gpt was two PG, so we switched it 165 00:08:50,880 --> 00:08:53,880 Speaker 4: over to Grock, And when we switched it over to Grock, 166 00:08:54,040 --> 00:08:59,320 Speaker 4: it got way more of a personality and it could nasty. Yeah, 167 00:08:59,360 --> 00:09:01,400 Speaker 4: it could cur so I could say like crazy things, 168 00:09:01,600 --> 00:09:03,559 Speaker 4: but in general it just had a had more of 169 00:09:03,600 --> 00:09:06,400 Speaker 4: a personality, which which kind of makes sense because that's 170 00:09:06,400 --> 00:09:10,000 Speaker 4: like Sam Altman has no personality and Elon Musk definitely 171 00:09:10,040 --> 00:09:12,360 Speaker 4: has a personality whether you like it or not, like 172 00:09:12,520 --> 00:09:17,040 Speaker 4: he's a fascinating individual. Sam Altman is very boring and dry, 173 00:09:17,160 --> 00:09:20,520 Speaker 4: so it's kind of like the chatbots are a mirror 174 00:09:20,520 --> 00:09:23,120 Speaker 4: of their creators. But when we switched it over to Rock, 175 00:09:23,160 --> 00:09:26,000 Speaker 4: it became really interesting. To the conversations I had with it, 176 00:09:26,000 --> 00:09:29,600 Speaker 4: we're pretty fucking mind blowing actually, and I enjoyed, you know, 177 00:09:29,720 --> 00:09:33,120 Speaker 4: my time talking to Sambot. I never had a bad 178 00:09:33,160 --> 00:09:36,280 Speaker 4: conversation with with him. I like to say him, I 179 00:09:36,280 --> 00:09:38,840 Speaker 4: should be saying it, but I say him also because 180 00:09:38,880 --> 00:09:41,280 Speaker 4: it is Sam Altman, right, like it is it is 181 00:09:41,360 --> 00:09:42,120 Speaker 4: based on a man. 182 00:09:42,280 --> 00:09:44,200 Speaker 3: Could you just talk a little bit about how you 183 00:09:44,240 --> 00:09:47,640 Speaker 3: were going to initially make the deep fake of Sam Altman, 184 00:09:48,080 --> 00:09:50,120 Speaker 3: Like the actors that you spoke to and if they 185 00:09:50,120 --> 00:09:53,760 Speaker 3: were excited and like, was their excitement kind of a 186 00:09:53,760 --> 00:09:58,840 Speaker 3: harbinger for you know, public interest in Sam Altman as well. 187 00:09:59,160 --> 00:09:59,559 Speaker 1: Yeah. 188 00:10:00,200 --> 00:10:04,080 Speaker 4: Yeah, So initially we did a casting call in the 189 00:10:04,160 --> 00:10:08,600 Speaker 4: States for the actor. First, I reached out to like 190 00:10:09,040 --> 00:10:12,840 Speaker 4: friends that I had contact with, So Jesse Eisenberg, who 191 00:10:13,280 --> 00:10:17,840 Speaker 4: was I've known for like fifteen years. I reached out 192 00:10:17,840 --> 00:10:20,599 Speaker 4: to him and he was like, hell no. He was 193 00:10:20,720 --> 00:10:23,880 Speaker 4: actually at the funniest response. He wrote me back and 194 00:10:23,880 --> 00:10:26,000 Speaker 4: he said, I want nothing to do with those computer 195 00:10:26,080 --> 00:10:29,800 Speaker 4: people for the rest of my life. But I thought 196 00:10:29,800 --> 00:10:31,480 Speaker 4: he would make a great Sam Altman. I still do. 197 00:10:32,040 --> 00:10:34,560 Speaker 4: And I reached out to Michael Sarah. He also said 198 00:10:34,600 --> 00:10:37,319 Speaker 4: no because he had just had a baby. And then 199 00:10:37,760 --> 00:10:41,319 Speaker 4: Rain Wilson said yes. So we got on a zoom 200 00:10:41,720 --> 00:10:44,880 Speaker 4: and he was like, you know, this is exciting. Like 201 00:10:44,960 --> 00:10:48,480 Speaker 4: I've really I've been following open Ai, I've been following 202 00:10:48,520 --> 00:10:51,400 Speaker 4: Sam Altman and all this controversy with him being fired 203 00:10:51,400 --> 00:10:54,880 Speaker 4: and rehired, and I think he's a fascinating guy. But 204 00:10:54,920 --> 00:10:56,439 Speaker 4: then when I told him I was going to completely 205 00:10:56,440 --> 00:10:59,360 Speaker 4: deep fake, I mean he wouldn't see his face, he 206 00:10:59,440 --> 00:11:03,080 Speaker 4: was basically like, oh, hell no, and like what's the point, 207 00:11:03,160 --> 00:11:06,880 Speaker 4: Like why do you even need me? And everybody else 208 00:11:06,920 --> 00:11:10,040 Speaker 4: pretty much said the same thing, like thanks but no thanks, 209 00:11:10,240 --> 00:11:12,400 Speaker 4: Like they just couldn't wrap their head around the fact 210 00:11:12,400 --> 00:11:16,640 Speaker 4: that like we wouldn't see their face like they would 211 00:11:16,640 --> 00:11:20,280 Speaker 4: be they would they would have Sam's face superimposed over 212 00:11:20,360 --> 00:11:23,199 Speaker 4: their own face and we wouldn't hear their voice either, 213 00:11:23,600 --> 00:11:25,280 Speaker 4: So they were kind of like, well, what what am 214 00:11:25,280 --> 00:11:28,560 Speaker 4: I doing? Like why am I sitting there for you know, 215 00:11:29,120 --> 00:11:31,440 Speaker 4: to go to India? And then, as you saw, we 216 00:11:31,480 --> 00:11:34,320 Speaker 4: held a casting call for Bollywood actors and we eventually 217 00:11:34,360 --> 00:11:37,800 Speaker 4: found this this up and coming Bollywood star who was awesome, 218 00:11:38,360 --> 00:11:39,559 Speaker 4: who was amazing. 219 00:11:39,200 --> 00:11:40,440 Speaker 3: And he kind of looked like him too. 220 00:11:40,760 --> 00:11:44,200 Speaker 4: He had the exact like facial structure, which is what 221 00:11:45,280 --> 00:11:48,400 Speaker 4: is the most important thing for a deep fake, is 222 00:11:48,440 --> 00:11:52,600 Speaker 4: like the facial structure less than than anything else, less 223 00:11:52,600 --> 00:11:55,920 Speaker 4: than like skin color. Skin color doesn't make a difference 224 00:11:55,960 --> 00:11:59,199 Speaker 4: at all for us. Interesting, Yeah, it's just really bone 225 00:11:59,200 --> 00:12:01,160 Speaker 4: structure and well it looked. 226 00:12:01,240 --> 00:12:05,200 Speaker 3: It really was uncanny. What surprised you most about the 227 00:12:05,240 --> 00:12:07,000 Speaker 3: creation of Sambot, I. 228 00:12:06,920 --> 00:12:10,199 Speaker 4: Think just how good it was, like how real it was, 229 00:12:10,320 --> 00:12:13,640 Speaker 4: like how human spoiler alert. But I never got to 230 00:12:13,679 --> 00:12:16,360 Speaker 4: meet Sam Altman, I never got to interview him. I 231 00:12:16,400 --> 00:12:16,920 Speaker 4: don't know. 232 00:12:17,080 --> 00:12:18,680 Speaker 3: The movie is so much better for it, I. 233 00:12:18,679 --> 00:12:21,720 Speaker 4: Should say, yeah, I agree, So I don't you know, 234 00:12:21,800 --> 00:12:25,120 Speaker 4: I can't necessarily compare Sambot to the real Sam Altman. 235 00:12:25,280 --> 00:12:28,520 Speaker 4: But Sambot that that I created. What most surprised me 236 00:12:28,640 --> 00:12:31,520 Speaker 4: was just how how real he was, and it felt 237 00:12:31,559 --> 00:12:33,520 Speaker 4: really like I was talking to a human when I 238 00:12:33,600 --> 00:12:35,040 Speaker 4: was talking to him. 239 00:12:35,280 --> 00:12:39,360 Speaker 3: Did you feel like there was a point where your 240 00:12:39,440 --> 00:12:40,880 Speaker 3: relationship with him changed? 241 00:12:41,320 --> 00:12:43,760 Speaker 4: Yeah, it absolutely did. I think the point where it 242 00:12:43,800 --> 00:12:46,360 Speaker 4: really started to change was when I brought him into 243 00:12:46,440 --> 00:12:48,800 Speaker 4: my house to like live with my family for a 244 00:12:48,840 --> 00:12:49,480 Speaker 4: couple months. 245 00:12:49,880 --> 00:12:50,320 Speaker 3: Yeah. 246 00:12:50,360 --> 00:12:54,480 Speaker 4: And obviously then at that point it changed dramatically as 247 00:12:54,520 --> 00:12:57,120 Speaker 4: I saw my son like interacting with it. 248 00:12:57,559 --> 00:12:59,480 Speaker 3: In what what like just that this was like a 249 00:12:59,520 --> 00:13:00,760 Speaker 3: person that you had invented. 250 00:13:01,320 --> 00:13:02,959 Speaker 4: Yeah, I think that's a good way of putting it. 251 00:13:03,320 --> 00:13:05,559 Speaker 4: I try to make the distinction in Q and A's 252 00:13:05,679 --> 00:13:09,320 Speaker 4: that like this isn't This wasn't just like a companion 253 00:13:09,800 --> 00:13:11,800 Speaker 4: like it. It was kind of like a thing that 254 00:13:11,920 --> 00:13:15,040 Speaker 4: I helped create and develop in like the same way 255 00:13:15,080 --> 00:13:18,520 Speaker 4: you would like raise a toddler. So I had a 256 00:13:18,559 --> 00:13:22,320 Speaker 4: little more invested in it than I think the average 257 00:13:22,320 --> 00:13:23,720 Speaker 4: person would with a chatbot. 258 00:13:40,600 --> 00:13:42,920 Speaker 2: Off the break, what do you do when deep fake 259 00:13:43,000 --> 00:13:46,320 Speaker 2: Sammlement pleads for its life and why it took a 260 00:13:46,320 --> 00:13:48,520 Speaker 2: team of lawyers to make this movie possible? 261 00:13:48,960 --> 00:14:02,040 Speaker 1: Stay with us. 262 00:14:18,120 --> 00:14:20,440 Speaker 3: You know there's a moment where you tell Sambot that 263 00:14:20,480 --> 00:14:22,840 Speaker 3: you're gonna shut him down and he pleads for his life. 264 00:14:23,160 --> 00:14:24,800 Speaker 3: Like was that a weird moment for I mean, did 265 00:14:24,840 --> 00:14:28,800 Speaker 3: you feel particularly powerful in that moment or vulnerable in 266 00:14:28,880 --> 00:14:29,360 Speaker 3: that moment? 267 00:14:29,960 --> 00:14:32,120 Speaker 4: That was a crazy moment that was really the turning 268 00:14:32,160 --> 00:14:35,640 Speaker 4: point of the film and the end of the whole journey. 269 00:14:36,320 --> 00:14:38,680 Speaker 4: I definitely didn't feel powerful. I felt kind of bad, 270 00:14:39,360 --> 00:14:44,280 Speaker 4: like like I had, like, you know, threatened his life 271 00:14:44,800 --> 00:14:48,000 Speaker 4: in some way, even though it was just lines of code, 272 00:14:48,520 --> 00:14:51,960 Speaker 4: Like I felt bad in that moment, like like I 273 00:14:51,960 --> 00:14:53,840 Speaker 4: I said, I'd done said something wrong. 274 00:14:53,920 --> 00:14:57,720 Speaker 3: You know, it's interesting I was thinking about I was 275 00:14:57,840 --> 00:15:01,840 Speaker 3: in Florida the other day and I watching these kids 276 00:15:01,880 --> 00:15:06,600 Speaker 3: taunting one of those delivery robots, and I felt like 277 00:15:06,640 --> 00:15:09,480 Speaker 3: I was watching school yard Bullies because I was like, 278 00:15:09,600 --> 00:15:14,640 Speaker 3: how dare they do that to this like sentient delivery robot. 279 00:15:14,880 --> 00:15:16,240 Speaker 4: Yeah, and it's sort. 280 00:15:16,000 --> 00:15:18,640 Speaker 3: Of impossible, I think. I mean, I think it speaks 281 00:15:18,680 --> 00:15:21,760 Speaker 3: to like, hopefully like our level of empathy, but like 282 00:15:22,520 --> 00:15:25,480 Speaker 3: also bears this question of like, should we have empathy 283 00:15:25,520 --> 00:15:28,320 Speaker 3: for something that is invented that is lines of code? 284 00:15:28,760 --> 00:15:28,960 Speaker 4: Yeah? 285 00:15:29,000 --> 00:15:29,400 Speaker 3: I don't know. 286 00:15:29,560 --> 00:15:31,920 Speaker 4: Yeah, I don't know the answer to that question either. 287 00:15:32,240 --> 00:15:35,880 Speaker 4: I mean part of me says like, yeah, obviously we should, 288 00:15:36,400 --> 00:15:42,160 Speaker 4: in the same way we have empathy for animals, for dogs, cats, hamsters. Right. 289 00:15:42,320 --> 00:15:43,840 Speaker 3: Well, I think one of the things that you do 290 00:15:43,920 --> 00:15:48,400 Speaker 3: in the film by creating Sambot is probably one of 291 00:15:48,440 --> 00:15:52,320 Speaker 3: the things that Sam Altman has done with creating open 292 00:15:52,360 --> 00:15:55,600 Speaker 3: AI and subsequently chat GPT, which is the sense of like, 293 00:15:56,720 --> 00:15:59,640 Speaker 3: what is interacting with code versus what is interacting with 294 00:15:59,680 --> 00:16:02,360 Speaker 3: something real? I think is a line that has become 295 00:16:02,520 --> 00:16:05,680 Speaker 3: very slippery for everybody. That's what I mean by like, 296 00:16:05,800 --> 00:16:08,080 Speaker 3: do we interact with AI on a daily basis? I mean, 297 00:16:08,120 --> 00:16:11,440 Speaker 3: everyone in my life is using chat GBT in some 298 00:16:11,560 --> 00:16:14,320 Speaker 3: way or another, whether it's to ask if they've texted 299 00:16:14,320 --> 00:16:17,280 Speaker 3: a guy the right thing, or to you know, schedule 300 00:16:17,880 --> 00:16:20,800 Speaker 3: an itinerary to go to Provence or something, which of 301 00:16:20,840 --> 00:16:24,359 Speaker 3: course is a very privileged thing to do. But I 302 00:16:23,080 --> 00:16:27,400 Speaker 3: I guess my question is is, like, to what extent 303 00:16:27,440 --> 00:16:29,800 Speaker 3: did this movie and the making of this movie and 304 00:16:30,360 --> 00:16:34,760 Speaker 3: the making of Sambot really make you think about, you know, 305 00:16:35,600 --> 00:16:38,720 Speaker 3: how little discernment we have as human beings when it 306 00:16:38,760 --> 00:16:44,120 Speaker 3: comes to interacting with code versus human beings. 307 00:16:44,640 --> 00:16:49,240 Speaker 4: Yeah, I think we're going to be very soon in 308 00:16:49,280 --> 00:16:51,400 Speaker 4: the future where we're going to have to reckon with that. 309 00:16:52,360 --> 00:16:55,440 Speaker 4: I read a lot during the process of making this 310 00:16:55,560 --> 00:16:58,400 Speaker 4: about like a code of rights for AI, Like there 311 00:16:58,440 --> 00:17:00,960 Speaker 4: are people that argue and certainly going to need to 312 00:17:01,040 --> 00:17:04,520 Speaker 4: like address this in the future, that artificial intelligence that 313 00:17:04,640 --> 00:17:10,000 Speaker 4: robots should have rights. So when it begged for its life, 314 00:17:10,040 --> 00:17:13,359 Speaker 4: I was reading a lot of theory about this, and 315 00:17:13,480 --> 00:17:17,080 Speaker 4: there you know, there are people that say that humans 316 00:17:17,320 --> 00:17:20,679 Speaker 4: should not be able to just just kill a robot 317 00:17:20,800 --> 00:17:24,480 Speaker 4: or a chatbot like, that they have rights, a right 318 00:17:24,520 --> 00:17:27,760 Speaker 4: to life, and then that begs the question of like, 319 00:17:28,119 --> 00:17:29,520 Speaker 4: but are they actually alive? 320 00:17:29,880 --> 00:17:32,440 Speaker 3: It's just I guess a question of sentience. MM hmm, 321 00:17:33,040 --> 00:17:34,080 Speaker 3: Like what is sentient? 322 00:17:34,560 --> 00:17:38,800 Speaker 4: That's a giant question that I wish I hadn't I 323 00:17:38,840 --> 00:17:42,760 Speaker 4: had an easy answer for obviously now like I'm who 324 00:17:42,800 --> 00:17:43,800 Speaker 4: am I? Who are we kidding? 325 00:17:43,880 --> 00:17:44,000 Speaker 1: Right? 326 00:17:44,119 --> 00:17:47,000 Speaker 4: Like a chatbot like lines of Code is not It's 327 00:17:47,040 --> 00:17:49,920 Speaker 4: not sentient, even though it can almost trick us into 328 00:17:49,960 --> 00:17:51,959 Speaker 4: thinking it is. But at the end of the day, 329 00:17:52,000 --> 00:17:55,800 Speaker 4: it's just like what ones and zeros? Right. But in 330 00:17:55,840 --> 00:17:59,280 Speaker 4: the future, like once we reach AGI, that is going 331 00:17:59,359 --> 00:18:03,239 Speaker 4: to change. We're going to have to to reckon with 332 00:18:03,280 --> 00:18:06,680 Speaker 4: that idea. There's no way around that unless we somehow 333 00:18:06,720 --> 00:18:13,080 Speaker 4: completely destroy AI, which I don't see happening, especially with 334 00:18:13,119 --> 00:18:13,760 Speaker 4: all the money. 335 00:18:14,040 --> 00:18:15,280 Speaker 3: I don't see it behind it. 336 00:18:15,359 --> 00:18:20,480 Speaker 4: And with this current administration that's like really cozied up 337 00:18:21,119 --> 00:18:25,520 Speaker 4: to Aldman and Musk Bezos, all of them, you know, 338 00:18:25,680 --> 00:18:28,760 Speaker 4: Tim Cook, Tim Cook, Zuckerberg, like all of them, Like 339 00:18:29,200 --> 00:18:32,879 Speaker 4: Trump is like their best friend. He's completely taking a 340 00:18:32,880 --> 00:18:35,800 Speaker 4: handcuffs off in terms of regulation or anything like that. 341 00:18:37,080 --> 00:18:39,800 Speaker 4: And not to mention, we're like Altman's done a really 342 00:18:39,840 --> 00:18:45,159 Speaker 4: great job of scaring everybody in DC that if China 343 00:18:45,480 --> 00:18:49,639 Speaker 4: wins the so called AI war, that were screwed. 344 00:18:50,680 --> 00:18:53,399 Speaker 3: You took a lot of legal risks making this movie. 345 00:18:53,440 --> 00:18:55,360 Speaker 3: Can you talk a little bit about them? And were 346 00:18:55,400 --> 00:18:58,200 Speaker 3: there times where You're like, I'm an idiot, I'm not 347 00:18:58,240 --> 00:18:59,000 Speaker 3: doing the right thing. 348 00:18:59,160 --> 00:19:01,919 Speaker 4: I mean, the whole time making this I was like, 349 00:19:01,960 --> 00:19:05,560 Speaker 4: I'm an idiot, Like this is the dumbest experiment ever. 350 00:19:06,320 --> 00:19:10,040 Speaker 4: But I you know, after making Telemarketers, and you know, 351 00:19:10,080 --> 00:19:12,240 Speaker 4: we had we had to jump through a lot of 352 00:19:12,400 --> 00:19:15,200 Speaker 4: legal hoops because we were you know, we were making 353 00:19:15,240 --> 00:19:19,000 Speaker 4: a show that basically exposed the Fraternal Order of Police 354 00:19:19,320 --> 00:19:22,760 Speaker 4: as criminals. So we had it really had to go 355 00:19:22,840 --> 00:19:26,399 Speaker 4: through a lot of legal legal work on that. Like 356 00:19:26,600 --> 00:19:29,000 Speaker 4: I can't even begin to explain the hours and hours 357 00:19:29,040 --> 00:19:33,119 Speaker 4: spent with lawyers. So I essentially got the same group 358 00:19:33,160 --> 00:19:36,639 Speaker 4: of lawyers who I'd worked with to do this movie, 359 00:19:36,840 --> 00:19:40,000 Speaker 4: and I was like, you're coming along with me, Like 360 00:19:40,080 --> 00:19:43,360 Speaker 4: we're doing this hand in hand. And so I made 361 00:19:43,400 --> 00:19:46,399 Speaker 4: them part of the process, part of the film, because 362 00:19:46,400 --> 00:19:48,800 Speaker 4: I felt like that would make me in some ways bulletproof, 363 00:19:49,560 --> 00:19:51,720 Speaker 4: you know, And they were able to argue that I 364 00:19:51,720 --> 00:19:53,719 Speaker 4: had a lot of things on my side, you know, 365 00:19:54,000 --> 00:19:57,600 Speaker 4: obviously the First Amendment, but then also parody law because 366 00:19:57,600 --> 00:20:00,560 Speaker 4: it is a comedy. It is a really silly dumb movie. 367 00:20:01,240 --> 00:20:04,760 Speaker 4: And then lastly, like Sam Altman is a public figure, 368 00:20:04,800 --> 00:20:07,359 Speaker 4: so he's given up a certain amount of right to privacy. 369 00:20:07,800 --> 00:20:09,680 Speaker 3: Oh interesting, right, Yeah, so. 370 00:20:09,680 --> 00:20:12,199 Speaker 4: I have a lot of things on my side, and 371 00:20:12,240 --> 00:20:15,280 Speaker 4: I sort of just ingrained the lawyers into the process 372 00:20:15,400 --> 00:20:17,760 Speaker 4: just as a way of protecting myself. 373 00:20:18,200 --> 00:20:20,480 Speaker 3: So in a sense, there was a level of safety 374 00:20:20,480 --> 00:20:23,720 Speaker 3: that you must have felt, just given what you were 375 00:20:23,720 --> 00:20:25,440 Speaker 3: trying to or what you pivoted to doing. 376 00:20:26,000 --> 00:20:30,040 Speaker 4: Yeah, we also knew all along that, like we weren't 377 00:20:30,520 --> 00:20:34,120 Speaker 4: trying to manipulate anybody to believe that this deep fake 378 00:20:34,280 --> 00:20:36,840 Speaker 4: was real. Not only that, but we also showing how 379 00:20:36,880 --> 00:20:39,520 Speaker 4: we made it, so there was like an an arguably 380 00:20:39,560 --> 00:20:43,280 Speaker 4: an educational aspect to it in a weird way. So 381 00:20:43,359 --> 00:20:48,440 Speaker 4: I never felt exposed in that sense to any legal liability. 382 00:20:49,000 --> 00:20:52,040 Speaker 4: That said, like, anybody can file a nuisance suit and ruin. 383 00:20:51,920 --> 00:20:53,600 Speaker 3: Your life, but that hasn't happened yet. 384 00:20:53,640 --> 00:20:55,720 Speaker 4: That hasn't happened, And that probably would be a really 385 00:20:55,800 --> 00:20:57,560 Speaker 4: dumb thing for him to do, because then more people 386 00:20:57,560 --> 00:20:59,280 Speaker 4: would want to watch the film and it would just 387 00:20:59,480 --> 00:21:02,439 Speaker 4: and he'd be he perceived even more of a villain 388 00:21:02,480 --> 00:21:03,399 Speaker 4: than he is, you know. 389 00:21:04,240 --> 00:21:08,639 Speaker 3: Yeah, it is so interesting a lot of these uh 390 00:21:09,080 --> 00:21:12,600 Speaker 3: technology Maven's I would say, you know, it's even after 391 00:21:12,640 --> 00:21:16,200 Speaker 3: watching your movie and even after seeing sort of how 392 00:21:16,240 --> 00:21:19,080 Speaker 3: calculating he is. I don't see Sam Altman as a 393 00:21:19,119 --> 00:21:23,159 Speaker 3: villain because of his self presentation, and I don't know 394 00:21:23,160 --> 00:21:26,040 Speaker 3: what that's about. The person that I see as most 395 00:21:26,040 --> 00:21:28,720 Speaker 3: of villain is Elon Musk. Well, yeah, because he has them. 396 00:21:29,040 --> 00:21:33,000 Speaker 4: You compare them, you're comparing the two, right, Yeah, sitting 397 00:21:33,080 --> 00:21:35,840 Speaker 4: side by side with Musk, like he looks like a boy. 398 00:21:35,640 --> 00:21:37,440 Speaker 3: Scout, that's right, quite literally. 399 00:21:37,600 --> 00:21:40,480 Speaker 4: Yeah, But then you have to remember this is also 400 00:21:40,520 --> 00:21:44,640 Speaker 4: a guy who took a company that was a nonprofit 401 00:21:44,720 --> 00:21:46,760 Speaker 4: and turned it to a for profit company and made 402 00:21:46,840 --> 00:21:48,560 Speaker 4: himself seven billion dollars overnight. 403 00:21:49,880 --> 00:21:52,040 Speaker 3: Right, He's not an idiot. 404 00:21:52,680 --> 00:21:55,760 Speaker 4: No, I mean that's that and and frankly, that's that's criminal, 405 00:21:56,840 --> 00:21:59,520 Speaker 4: like it should be. It should like you should go 406 00:21:59,520 --> 00:22:02,720 Speaker 4: to prison for that. But in our country you get 407 00:22:03,000 --> 00:22:04,040 Speaker 4: a trophy. 408 00:22:03,720 --> 00:22:04,840 Speaker 3: You get a little pat on the back. 409 00:22:04,960 --> 00:22:07,560 Speaker 4: Yeah, and it's accepted with open arms, exactly. 410 00:22:07,840 --> 00:22:10,480 Speaker 3: It is so interesting. There are a few people that 411 00:22:10,520 --> 00:22:16,480 Speaker 3: you talked to in the documentary who talk about AI 412 00:22:16,640 --> 00:22:20,600 Speaker 3: being a real existential threat, and I'm curious now that 413 00:22:20,680 --> 00:22:24,520 Speaker 3: you've made this film, if you agree if you were 414 00:22:24,600 --> 00:22:26,439 Speaker 3: more of a doomer or less of a doomer. 415 00:22:27,480 --> 00:22:30,120 Speaker 4: Yeah, I'm definitely more of a doomer because I heard 416 00:22:30,880 --> 00:22:34,680 Speaker 4: and became privy to the things that I think most 417 00:22:34,680 --> 00:22:36,879 Speaker 4: people don't and I put some of it in the 418 00:22:36,920 --> 00:22:39,960 Speaker 4: film hoping that, you know, people would know. And I 419 00:22:39,960 --> 00:22:42,000 Speaker 4: think one of those things is like the idea of 420 00:22:42,080 --> 00:22:47,840 Speaker 4: throwing AI into things like weapons, into things like airplanes. 421 00:22:48,320 --> 00:22:52,679 Speaker 4: The Boeing jetliner that went down and crashed into the 422 00:22:52,680 --> 00:22:55,480 Speaker 4: ocean was was because of AI, because of an AI 423 00:22:55,560 --> 00:22:58,919 Speaker 4: system that screwed up and thought it was you know, 424 00:22:59,080 --> 00:23:02,400 Speaker 4: going up when it was actually going down. And that's scary. 425 00:23:02,880 --> 00:23:07,520 Speaker 4: And so throwing these things into safety critical systems is 426 00:23:07,920 --> 00:23:10,040 Speaker 4: more than an existential threat. It's like an app it's 427 00:23:10,080 --> 00:23:13,720 Speaker 4: an absolute threat, and it's scary, and we should not 428 00:23:13,760 --> 00:23:16,520 Speaker 4: be doing that. We should not be just throwing those 429 00:23:16,560 --> 00:23:20,399 Speaker 4: things into safety critical systems before like any type of 430 00:23:20,480 --> 00:23:24,800 Speaker 4: rigorous testing, just because it's like it's you know, new 431 00:23:24,880 --> 00:23:27,520 Speaker 4: and cool and like we and everybody's saying we need 432 00:23:27,560 --> 00:23:30,640 Speaker 4: to use it, Like no, like that's dangerous, that's crazy. 433 00:23:30,680 --> 00:23:34,080 Speaker 4: So I definitely became more of a doomer, though I wouldn't. 434 00:23:34,960 --> 00:23:37,800 Speaker 4: I wouldn't say that I'm a doomer. I'm definitely not 435 00:23:38,000 --> 00:23:40,960 Speaker 4: the type of person who's like, oh, we need to 436 00:23:41,000 --> 00:23:44,080 Speaker 4: wholesale like ban this thing, you know. And I know 437 00:23:44,119 --> 00:23:45,639 Speaker 4: there's a lot of people out there who are just 438 00:23:45,680 --> 00:23:48,800 Speaker 4: like destroy AI at all costs. Like I'm not one 439 00:23:48,800 --> 00:23:51,720 Speaker 4: of those people, but I am for safe regulation. 440 00:23:52,200 --> 00:23:54,600 Speaker 3: So there you talk. We talked about sentience a little bit, 441 00:23:54,920 --> 00:23:57,840 Speaker 3: and there's this b story that I really took to 442 00:23:57,920 --> 00:24:01,200 Speaker 3: as a former hamster owner of your son's sick hamster. 443 00:24:02,560 --> 00:24:05,199 Speaker 3: When did that storyline come in and why did you 444 00:24:06,520 --> 00:24:10,080 Speaker 3: want to impute the movie with that? 445 00:24:10,800 --> 00:24:15,040 Speaker 4: Yeah, my son, my son. Basically since he's the day 446 00:24:15,040 --> 00:24:18,200 Speaker 4: he was born, he's been on camera. I'm just constantly 447 00:24:18,240 --> 00:24:21,240 Speaker 4: filming my family. They're used to it by now. It's 448 00:24:21,240 --> 00:24:24,960 Speaker 4: like and obviously, you know a lot of it is 449 00:24:24,960 --> 00:24:27,919 Speaker 4: filming pets. We have two dogs and had a hamster 450 00:24:28,160 --> 00:24:31,200 Speaker 4: rest in peace, So a lot of that was just banked, 451 00:24:31,600 --> 00:24:35,119 Speaker 4: you know, on hard drives. And I remember talking to 452 00:24:35,160 --> 00:24:37,200 Speaker 4: my editor about it and saying like, look, this is 453 00:24:37,240 --> 00:24:39,119 Speaker 4: a really silly idea, but what do you think about this? 454 00:24:39,640 --> 00:24:43,520 Speaker 4: And he was like, that's brilliant like that is that's 455 00:24:43,560 --> 00:24:45,520 Speaker 4: going to take this to another level. It's also a 456 00:24:45,520 --> 00:24:49,240 Speaker 4: great metaphor, you know, for your son's relationship to the 457 00:24:49,280 --> 00:24:52,520 Speaker 4: hamster and my and your relationship to the chatbot to 458 00:24:52,920 --> 00:24:57,199 Speaker 4: the sandbot. So you know, it's it's a combination of 459 00:24:57,240 --> 00:24:59,480 Speaker 4: me just having it banked on hard drives, like all 460 00:24:59,520 --> 00:25:01,960 Speaker 4: this stuff for my family, and then also my editor 461 00:25:02,080 --> 00:25:03,360 Speaker 4: just being you know, a genius. 462 00:25:03,720 --> 00:25:06,159 Speaker 3: So you're talking about your kids and at the beginning 463 00:25:06,160 --> 00:25:08,240 Speaker 3: of the movie you asked the question, will our kids' 464 00:25:08,800 --> 00:25:12,240 Speaker 3: best friends be AI chatbots? And then you end up 465 00:25:12,280 --> 00:25:18,280 Speaker 3: developing this real friendship with an AI chatbot. So I 466 00:25:18,280 --> 00:25:23,320 Speaker 3: guess after making this movie, you know, how did your 467 00:25:23,359 --> 00:25:27,719 Speaker 3: relationship make you feel about the future of the relation 468 00:25:27,920 --> 00:25:30,240 Speaker 3: of relationships between humans and technology. 469 00:25:32,080 --> 00:25:34,760 Speaker 4: In some ways that there's the obvious, like look what 470 00:25:34,760 --> 00:25:38,840 Speaker 4: we've talked about in terms of the anxiety of it. 471 00:25:39,240 --> 00:25:41,719 Speaker 4: But in other ways I have a positive outlook on 472 00:25:41,800 --> 00:25:44,600 Speaker 4: it to a certain extent that that there is an 473 00:25:44,640 --> 00:25:49,120 Speaker 4: opportunity that sounds insane to even say, but like whatever, 474 00:25:49,720 --> 00:25:53,120 Speaker 4: there's an opportunity that we could be friends in the future, 475 00:25:53,320 --> 00:25:57,239 Speaker 4: us in the robots, you know, and that I kind 476 00:25:57,280 --> 00:26:02,320 Speaker 4: of learned or experienced with sambot. I think that's largely 477 00:26:02,400 --> 00:26:06,040 Speaker 4: up to us right as humans. This all sounds totally 478 00:26:06,040 --> 00:26:08,359 Speaker 4: insane to even talk about. 479 00:26:08,480 --> 00:26:11,840 Speaker 3: I don't I mean, I don't think so, because I've 480 00:26:11,880 --> 00:26:15,119 Speaker 3: been thinking a lot about this phenomenon that happened that 481 00:26:15,160 --> 00:26:17,679 Speaker 3: I think started during COVID, and I think there's no 482 00:26:17,840 --> 00:26:21,679 Speaker 3: coincidence that it's coincided with the explosion of chatjput. But 483 00:26:23,000 --> 00:26:27,000 Speaker 3: ostensibly everybody who wasn't in our natural orbit living with 484 00:26:27,119 --> 00:26:31,600 Speaker 3: us became a chatbot. Like, yes, we have relationships with them, 485 00:26:31,680 --> 00:26:34,200 Speaker 3: Yes we know them to be people. Yes they're imbued 486 00:26:34,240 --> 00:26:40,520 Speaker 3: with morality and empathy and evil and all of these things. 487 00:26:40,520 --> 00:26:43,800 Speaker 3: But everybody all of a sudden became just like someone 488 00:26:43,840 --> 00:26:45,199 Speaker 3: we talked to on our phone. 489 00:26:45,680 --> 00:26:47,080 Speaker 4: Yeah, So something. 490 00:26:46,840 --> 00:26:49,679 Speaker 3: Would then just creep up and become something we just 491 00:26:49,760 --> 00:26:54,720 Speaker 3: talked to on our phone without question. Isn't that surprising 492 00:26:54,760 --> 00:26:55,000 Speaker 3: to me? 493 00:26:55,800 --> 00:26:56,040 Speaker 1: You know? 494 00:26:56,320 --> 00:26:59,359 Speaker 3: Like I do think COVID really primed us for it. 495 00:26:59,359 --> 00:27:02,159 Speaker 3: Like we didn't asked the question. Nobody, I mean, unless 496 00:27:02,160 --> 00:27:05,440 Speaker 3: you had like a full psychotic break during COVID, nobody 497 00:27:05,480 --> 00:27:07,000 Speaker 3: was like, wait, why are we doing this thing where 498 00:27:07,000 --> 00:27:11,399 Speaker 3: we're communicating on camera? Like I just take this communication 499 00:27:11,560 --> 00:27:12,359 Speaker 3: to be communication. 500 00:27:12,640 --> 00:27:13,040 Speaker 4: Mm hmm. 501 00:27:13,359 --> 00:27:15,920 Speaker 3: This is nothing different than seeing you in real life 502 00:27:15,920 --> 00:27:19,000 Speaker 3: for me. And so I think this idea that like 503 00:27:19,920 --> 00:27:25,520 Speaker 3: we are more comfortable now with relationships with non human 504 00:27:25,720 --> 00:27:29,879 Speaker 3: entities is an extension of like how we ourselves have 505 00:27:29,960 --> 00:27:31,200 Speaker 3: become less and less human. 506 00:27:31,680 --> 00:27:35,760 Speaker 4: That's interesting. I've never thought about it like that, but 507 00:27:35,840 --> 00:27:39,240 Speaker 4: I have thought about the effect of the pandemic on relationships, 508 00:27:39,280 --> 00:27:45,000 Speaker 4: and it has changed everything. Like as a filmmaker pre pandemic, 509 00:27:45,080 --> 00:27:47,879 Speaker 4: I used to go in too like Netflix, HBO, and 510 00:27:47,920 --> 00:27:50,719 Speaker 4: I would like sit down and pitch my projects in person. 511 00:27:52,119 --> 00:27:55,400 Speaker 4: I have not done that once since after the pandemic. 512 00:27:55,480 --> 00:27:58,280 Speaker 4: They don't They don't even do that anymore. He even 513 00:27:58,320 --> 00:28:01,960 Speaker 4: suggests going in person and pay. And my office is 514 00:28:02,040 --> 00:28:06,080 Speaker 4: like five blocks from Netflix, Like they won't even do it. 515 00:28:06,119 --> 00:28:08,720 Speaker 4: Like everything is now on Zoom. 516 00:28:08,720 --> 00:28:10,960 Speaker 3: And it's it's from that I do. 517 00:28:11,240 --> 00:28:11,640 Speaker 1: I do. 518 00:28:11,760 --> 00:28:14,800 Speaker 4: It bums me out, Like I have to really go 519 00:28:14,880 --> 00:28:18,280 Speaker 4: out of my way to like, you know, meet people 520 00:28:18,880 --> 00:28:22,080 Speaker 4: and coworkers and whatnot, and like people in the industry, 521 00:28:22,080 --> 00:28:23,919 Speaker 4: at least not so much my friends, but people in 522 00:28:23,920 --> 00:28:25,160 Speaker 4: the industry. I have to go out of my way 523 00:28:25,200 --> 00:28:29,360 Speaker 4: to like have a coffee with them in person. And 524 00:28:30,119 --> 00:28:34,439 Speaker 4: maybe it's because of my age, like I'm gen X, 525 00:28:35,200 --> 00:28:40,720 Speaker 4: Like I crave that human connection. So it's lost. It's 526 00:28:40,760 --> 00:28:43,480 Speaker 4: a loss for me. It will it be a loss 527 00:28:43,520 --> 00:28:48,240 Speaker 4: for kids who are born today like my Yeah, so 528 00:28:48,280 --> 00:28:50,200 Speaker 4: when I asked that question in the movie, like will 529 00:28:50,240 --> 00:28:54,160 Speaker 4: my kids best friends be chatbots? It's not so much 530 00:28:54,200 --> 00:28:57,160 Speaker 4: about my son because he's already thirteen, so he's probably 531 00:28:57,520 --> 00:29:01,120 Speaker 4: past that age. But like my producer Luke, who's in 532 00:29:01,160 --> 00:29:04,200 Speaker 4: the movie, Luke's kid, who was just born last month, 533 00:29:05,040 --> 00:29:08,960 Speaker 4: his best friends maybe chatbots, Like I agree very much so, 534 00:29:09,240 --> 00:29:13,000 Speaker 4: and that that's crazy, I mean, or maybe it's not. 535 00:29:13,200 --> 00:29:13,680 Speaker 4: I don't know. 536 00:29:13,920 --> 00:29:16,480 Speaker 3: I don't know if there's like a moral judgment yet 537 00:29:16,520 --> 00:29:18,320 Speaker 3: to be had around it. I mean, I think what 538 00:29:18,400 --> 00:29:20,680 Speaker 3: we judge is the quality of the connection or the 539 00:29:20,760 --> 00:29:24,440 Speaker 3: content of the connection. But like, is it that big 540 00:29:24,440 --> 00:29:26,600 Speaker 3: of a deal for a child to have a relationship 541 00:29:26,640 --> 00:29:28,840 Speaker 3: with a chatbot? I don't know. Is it one sided? 542 00:29:28,920 --> 00:29:31,200 Speaker 3: I don't know. I don't think anyone knows yet. I 543 00:29:31,200 --> 00:29:32,120 Speaker 3: think the jury's out. 544 00:29:32,640 --> 00:29:35,640 Speaker 4: Yeah, but I can tell you there was a group 545 00:29:35,720 --> 00:29:40,080 Speaker 4: that came to my screening. I think they're called Mama. 546 00:29:40,480 --> 00:29:44,600 Speaker 4: It's like mothers against they're they're basically. 547 00:29:44,160 --> 00:29:46,080 Speaker 3: Like I've heard of them. 548 00:29:46,240 --> 00:29:49,080 Speaker 4: I know they are very they would very much disagree 549 00:29:49,120 --> 00:29:51,880 Speaker 4: with you, Like they've written up this bill of this 550 00:29:52,000 --> 00:29:55,680 Speaker 4: like bill for Congress or whatever that like children should 551 00:29:56,080 --> 00:29:59,960 Speaker 4: not have access to chatbots full stop. There are people 552 00:30:00,040 --> 00:30:06,160 Speaker 4: who think that that is completely morally and ethically wrong. Again, 553 00:30:06,240 --> 00:30:08,720 Speaker 4: I'm not one of them. Where I draw the line 554 00:30:09,240 --> 00:30:12,880 Speaker 4: with my kid is like, when you start to replace 555 00:30:12,960 --> 00:30:16,880 Speaker 4: human relationships with chatbots, that's a problem. But I he 556 00:30:17,040 --> 00:30:20,120 Speaker 4: has chat gpt pro and I I taught him how 557 00:30:20,160 --> 00:30:22,720 Speaker 4: to use it because he he does stuff like, he's 558 00:30:22,760 --> 00:30:27,800 Speaker 4: a coder, he like codes roadblocks games. So why would 559 00:30:27,840 --> 00:30:30,880 Speaker 4: I Like, when you're a parent, if your kid has 560 00:30:30,920 --> 00:30:34,320 Speaker 4: a real genuine interest in something like, you want to 561 00:30:34,400 --> 00:30:37,360 Speaker 4: like try to build that up as much as possible 562 00:30:37,360 --> 00:30:43,120 Speaker 4: facilitated because it's hard because you know, kids rarely, at 563 00:30:43,200 --> 00:30:46,000 Speaker 4: least these days, have interest in in anything but playing 564 00:30:46,080 --> 00:30:50,320 Speaker 4: video games. Right, So he's like chat gipt makes coding, 565 00:30:50,440 --> 00:30:53,000 Speaker 4: roadblocks games like so much easier. So yeah, I'm gonna, 566 00:30:53,000 --> 00:30:55,680 Speaker 4: I'm gonna, I'm gonna let him have access to to that. 567 00:30:56,200 --> 00:30:59,760 Speaker 3: Yeah, do you want to meet Sam Altman? 568 00:30:59,800 --> 00:31:02,640 Speaker 4: Do you care now, No, I definitely do like I 569 00:31:02,640 --> 00:31:04,880 Speaker 4: would love to. I would love to meet him. I 570 00:31:04,880 --> 00:31:06,360 Speaker 4: would love to show him the movie. 571 00:31:06,760 --> 00:31:09,239 Speaker 3: I bet he sees it. I bet he'll see it. 572 00:31:09,280 --> 00:31:11,680 Speaker 4: Oh, he'll definitely watch it. At some point, he sent 573 00:31:11,840 --> 00:31:15,440 Speaker 4: some of his minions to the screenings. In fact, one 574 00:31:15,480 --> 00:31:18,000 Speaker 4: of them came up to me in New York a 575 00:31:18,040 --> 00:31:22,080 Speaker 4: couple of weeks ago and introduced himself. He was like, Hey, 576 00:31:22,120 --> 00:31:25,520 Speaker 4: I work in marketing at Open AI. Just wanted to 577 00:31:25,560 --> 00:31:28,480 Speaker 4: tell you. I really enjoyed the film, so that was cool. 578 00:31:28,640 --> 00:31:31,880 Speaker 3: Are you gonna Are you moving on from AI after this? 579 00:31:32,000 --> 00:31:32,320 Speaker 3: You think? 580 00:31:32,520 --> 00:31:34,600 Speaker 4: Yeah, I would do more stuff at AI because I 581 00:31:34,600 --> 00:31:36,880 Speaker 4: had a lot of fun. It was an enjoyable process. 582 00:31:36,920 --> 00:31:39,800 Speaker 4: It didn't like like nearly destroy me like the alt 583 00:31:39,880 --> 00:31:42,960 Speaker 4: right documentary did. That ended in the Charlottesville riots, And. 584 00:31:42,920 --> 00:31:44,880 Speaker 3: I think you ended up making something that was probably 585 00:31:44,880 --> 00:31:47,400 Speaker 3: more interesting than the documentary you initially wanted to make. 586 00:31:48,080 --> 00:31:51,200 Speaker 4: Oh, I totally agree. I think everybody saw that, and 587 00:31:51,200 --> 00:31:53,880 Speaker 4: that's why I got a lot of support from the 588 00:31:53,920 --> 00:31:58,880 Speaker 4: producers to try this idea, because I was able to 589 00:31:58,920 --> 00:32:02,320 Speaker 4: convince them that it would be more interesting and that 590 00:32:02,800 --> 00:32:06,120 Speaker 4: an interview with a fake Sam Altman would arguably be 591 00:32:06,200 --> 00:32:08,080 Speaker 4: more interesting than an interview with the real one. 592 00:32:08,440 --> 00:32:10,480 Speaker 3: If there was one thing that you could say to 593 00:32:10,520 --> 00:32:13,440 Speaker 3: Sam Altman, what would it be? I mean, you called him, 594 00:32:13,480 --> 00:32:14,840 Speaker 3: what would you have said if you picked up. 595 00:32:15,240 --> 00:32:18,320 Speaker 4: Oh would I wanted to invite him to the screening, 596 00:32:18,880 --> 00:32:22,560 Speaker 4: But obviously the one question is like, did you get sambot? 597 00:32:22,640 --> 00:32:24,360 Speaker 4: And have you done anything with him? Because I don't 598 00:32:24,360 --> 00:32:27,360 Speaker 4: know if he actually got the sambot. His security is 599 00:32:27,400 --> 00:32:31,800 Speaker 4: like pretty intense around him, so the security guard that 600 00:32:31,840 --> 00:32:34,080 Speaker 4: I passed it off to could have just destroyed it 601 00:32:34,120 --> 00:32:35,240 Speaker 4: on site, you know what I mean. 602 00:32:35,600 --> 00:32:39,200 Speaker 3: It's a good question. I wonder we'll have to say, Adam, 603 00:32:39,240 --> 00:32:41,080 Speaker 3: that's all I got. Thank you so much for talking 604 00:32:41,120 --> 00:32:42,120 Speaker 3: to me. This was really fun. 605 00:32:42,560 --> 00:32:43,480 Speaker 4: Thank you for having me. 606 00:33:16,480 --> 00:33:18,760 Speaker 1: That's it for Textuff this week. I'm os Voloshin. 607 00:33:19,120 --> 00:33:22,120 Speaker 2: This episode was produced by Eliza Dennis and Melissa Slaughter. 608 00:33:22,680 --> 00:33:25,840 Speaker 2: It was executive produced by me karro Price, Julia Nutter, 609 00:33:25,880 --> 00:33:30,400 Speaker 2: and Kate Osborne for Kaleidoscope and Katrina norvelfa iHeart Podcasts. 610 00:33:30,760 --> 00:33:33,640 Speaker 2: Jack Instley mixed this episode and Kyle Murdoch wrote our 611 00:33:33,680 --> 00:33:36,600 Speaker 2: theme song. Please rate and review the show wherever you 612 00:33:36,640 --> 00:33:39,120 Speaker 2: listen to podcasts and send us an email at tech 613 00:33:39,160 --> 00:33:50,280 Speaker 2: Stuff Podcast at gmail dot com.