1 00:00:15,396 --> 00:00:32,836 Speaker 1: Pushkin. I'm Jacob Goldstein, and this is What's Your Problem? 2 00:00:33,016 --> 00:00:36,236 Speaker 1: My guest today is Luke Aragoni. He is the co-founder 3 00:00:36,256 --> 00:00:39,506 Speaker 1: and CEO of a company called Lodi. And Lodi is 4 00:00:39,526 --> 00:00:42,846 Speaker 1: an acronym, and it stands for Likeness on the Internet. 5 00:00:43,456 --> 00:00:46,846 Speaker 1: Tens of thousands of famous people pay Lodi to scour 6 00:00:46,886 --> 00:00:51,726 Speaker 1: the internet and take down deepfakes and unapproved endorsements. If 7 00:00:51,746 --> 00:00:53,726 Speaker 1: you've heard of a person, if a person is a 8 00:00:53,766 --> 00:00:56,226 Speaker 1: little bit famous, there is a very good chance they 9 00:00:56,266 --> 00:00:59,376 Speaker 1: are one of Lodi's clients. But in Luke's telling, this 10 00:00:59,516 --> 00:01:02,736 Speaker 1: is just the beginning. The company will soon start offering 11 00:01:02,796 --> 00:01:06,456 Speaker 1: a service for celebrities to license the use of their image. 12 00:01:06,976 --> 00:01:09,716 Speaker 1: In other words, ordinary people will be able to pay 13 00:01:09,736 --> 00:01:12,666 Speaker 1: Lodi to create stuff with AI using the faces and 14 00:01:12,736 --> 00:01:16,946 Speaker 1: voices of real celebrities. Lodi also plans to start offering 15 00:01:16,986 --> 00:01:20,426 Speaker 1: its services to ordinary people. The plan is to include 16 00:01:20,446 --> 00:01:23,406 Speaker 1: a free tier for people who are not famous, but 17 00:01:23,446 --> 00:01:26,096 Speaker 1: who are worried about what's happening with their likeness on 18 00:01:26,136 --> 00:01:29,896 Speaker 1: the internet. At a macro level, the biggest level, Luke's 19 00:01:29,936 --> 00:01:33,656 Speaker 1: problem is this. How do you turn the internet into 20 00:01:33,716 --> 00:01:37,046 Speaker 1: a place where people have more control over their name, image, 21 00:01:37,186 --> 00:01:40,936 Speaker 1: and likeness? In our conversation, Luke and I talked about 22 00:01:41,056 --> 00:01:45,376 Speaker 1: the technology behind Lodi. We talked about federal legislation that 23 00:01:45,436 --> 00:01:48,746 Speaker 1: may soon make it easier for people to fight deep fakes. 24 00:01:49,426 --> 00:01:52,146 Speaker 1: And we also talk about AI and the future of creativity. 25 00:01:52,846 --> 00:01:54,666 Speaker 1: To start, I asked Luke to give me the state 26 00:01:54,686 --> 00:01:58,476 Speaker 1: of play for what's happening right now with unauthorized use 27 00:01:58,536 --> 00:02:00,556 Speaker 1: of people's likeness on the internet. 28 00:02:01,416 --> 00:02:03,636 Speaker 2: So you have kind of two different categories of people. 29 00:02:03,796 --> 00:02:06,776 Speaker 2: If you're really famous and your likeness is going to 30 00:02:06,816 --> 00:02:10,126 Speaker 2: get used to sell random things. It could be to 31 00:02:10,146 --> 00:02:13,086 Speaker 2: sell a scam. It's like an impersonation, a fake Instagram 32 00:02:13,126 --> 00:02:15,786 Speaker 2: of a celebrity, and they're selling you anything from VIP 33 00:02:15,906 --> 00:02:19,106 Speaker 2: access to the next tour to a romance scam that 34 00:02:19,126 --> 00:02:21,766 Speaker 2: they're falling in love with you, right? And all of 35 00:02:21,806 --> 00:02:25,506 Speaker 2: them have some kind of economic extraction from it. Or 36 00:02:25,566 --> 00:02:28,086 Speaker 2: it's kind of more straightforward. It's like, hey, I've come 37 00:02:28,126 --> 00:02:30,516 Speaker 2: up with a brand new shampoo, and this famous celebrity 38 00:02:30,596 --> 00:02:33,596 Speaker 2: uses the shampoo. Here's a deepfake video talking about how 39 00:02:33,636 --> 00:02:37,556 Speaker 2: they love it. So that's probably the most common way 40 00:02:37,596 --> 00:02:39,356 Speaker 2: in which it's used. And then you have a kind 41 00:02:39,376 --> 00:02:44,676 Speaker 2: of more incendiary thing where everyone else, like me, for example, 42 00:02:44,956 --> 00:02:47,036 Speaker 2: what we have to be concerned about is someone deepfaking 43 00:02:47,116 --> 00:02:51,286 Speaker 2: us for any kind of authorization reason. And then there's 44 00:02:51,306 --> 00:02:53,526 Speaker 2: an even darker way in which it's used that we 45 00:02:53,586 --> 00:02:56,686 Speaker 2: saw at some high schools this year, which is a 46 00:02:56,726 --> 00:02:59,286 Speaker 2: lot of folks are getting deepfaked to make images of 47 00:02:59,326 --> 00:03:02,906 Speaker 2: them that were never taken, but have a very harmful 48 00:03:02,986 --> 00:03:04,426 Speaker 2: nature to them and their reputation. 49 00:03:04,996 --> 00:03:08,416 Speaker 1: Do you mean revenge porn? You're talking around revenge porn there. 50 00:03:08,436 --> 00:03:12,236 Speaker 2: Yes. We like to call it non-consensual intimate images, which 51 00:03:12,316 --> 00:03:14,396 Speaker 2: is a lot better of a term than revenge porn. 52 00:03:14,456 --> 00:03:17,256 Speaker 2: But colloquially, we all call it revenge porn, and so 53 00:03:17,276 --> 00:03:19,176 Speaker 2: we even have to kind of mix the terms together. 54 00:03:19,196 --> 00:03:21,196 Speaker 2: But just to kind of re-educate a lot of people, 55 00:03:21,716 --> 00:03:23,456 Speaker 2: NCII is a much better term for it. 56 00:03:23,876 --> 00:03:28,686 Speaker 1: Okay, fair enough. So this is a potential problem for everybody. 57 00:03:29,426 --> 00:03:33,466 Speaker 1: What's like... What's the scope of it right now? Like, 58 00:03:33,506 --> 00:03:37,426 Speaker 1: if I just pick some random A-list celebrity, will there, like, 100% 59 00:03:37,426 --> 00:03:42,266 Speaker 1: be deepfakes of them scamming on the internet right now? 60 00:03:42,726 --> 00:03:48,756 Speaker 2: Yeah, we track about 30,000 or so celebrity accounts, people, personas. 61 00:03:49,336 --> 00:03:51,416 Speaker 1: And those are your clients, or that's just you kind 62 00:03:51,436 --> 00:03:52,656 Speaker 1: of understanding the landscape? 63 00:03:53,916 --> 00:03:55,966 Speaker 2: Two-thirds, a little bit more, a little over 20,000 are clients. 64 00:03:57,136 --> 00:04:00,216 Speaker 1: So that's way below the A-list. That's way more famous 65 00:04:00,236 --> 00:04:02,066 Speaker 1: people than anybody could name. Oh, yeah. 66 00:04:02,126 --> 00:04:04,646 Speaker 2: No, it is. It's pretty much every famous person that 67 00:04:04,686 --> 00:04:06,766 Speaker 2: you can think of, even if you don't remember their name. 68 00:04:07,066 --> 00:04:09,006 Speaker 2: And so, yes, for those folks, you're going to get 69 00:04:09,146 --> 00:04:13,026 Speaker 2: nearly 100% coverage on impersonations and other scams. 70 00:04:13,626 --> 00:04:16,436 Speaker 1: When you say nearly 100% coverage, you mean for all 71 00:04:16,476 --> 00:04:18,356 Speaker 1: of those people, there are going to be deep fakes. 72 00:04:18,176 --> 00:04:21,696 Speaker 2: Of them on the Internet? Yes, and impersonations. So it's 73 00:04:21,716 --> 00:04:23,956 Speaker 2: like a real picture of them on an Instagram account, 74 00:04:24,236 --> 00:04:25,256 Speaker 2: but not their real account. 75 00:04:25,756 --> 00:04:28,156 Speaker 1: Right, so that's not a deepfake in the AI sense. 76 00:04:28,316 --> 00:04:30,656 Speaker 1: You can just take the picture of the person and 77 00:04:30,696 --> 00:04:33,236 Speaker 1: be like, look, this person loves our soap, and that's 78 00:04:33,556 --> 00:04:34,596 Speaker 1: a scam and a lie. 79 00:04:34,916 --> 00:04:35,336 Speaker 2: Absolutely. 80 00:04:35,376 --> 00:04:38,026 Speaker 1: And just to be clear, like, that is illegal because 81 00:04:38,066 --> 00:04:40,826 Speaker 1: people have a right to their name, image, and likeness? Like, 82 00:04:40,846 --> 00:04:44,666 Speaker 1: that's the fundamental business and kind of intellectual property thing 83 00:04:44,706 --> 00:04:45,526 Speaker 1: that we're talking about? 84 00:04:45,926 --> 00:04:48,426 Speaker 2: Yeah, it's called a right of publicity. And ironically, it 85 00:04:48,466 --> 00:04:51,936 Speaker 2: became a bigger conversation when the camera came around. Uh-huh. 86 00:04:52,296 --> 00:04:54,756 Speaker 2: And so, you know, 100 plus years ago when people 87 00:04:54,796 --> 00:04:57,196 Speaker 2: started taking pictures of anyone, someone was like, well, I'll 88 00:04:57,226 --> 00:04:59,206 Speaker 2: put this on the front of my cereal box. This 89 00:04:59,246 --> 00:05:01,046 Speaker 2: is just like an average person. They're like, no, that 90 00:05:01,086 --> 00:05:04,166 Speaker 2: person has, all right, you can't just make, it really 91 00:05:04,186 --> 00:05:06,686 Speaker 2: is the commercial component. You can't make money off their 92 00:05:06,726 --> 00:05:09,786 Speaker 2: likeness without their approval or their endorsement. 93 00:05:11,426 --> 00:05:15,686 Speaker 1: Okay, so that's the lay of the land. Your company 94 00:05:15,706 --> 00:05:19,326 Speaker 1: sort of walks into this universe as it is growing 95 00:05:19,356 --> 00:05:23,186 Speaker 1: a few years ago. Like, Basically, what's your company do? 96 00:05:23,486 --> 00:05:25,226 Speaker 2: What we do is we download the internet every day, 97 00:05:25,946 --> 00:05:28,566 Speaker 2: and then we run face recognition over it looking for 98 00:05:28,606 --> 00:05:30,886 Speaker 2: our clients. And when we find them in a spot 99 00:05:30,906 --> 00:05:34,306 Speaker 2: that they don't belong, whether it be a scam, impersonation, NCII, 100 00:05:34,526 --> 00:05:37,246 Speaker 2: or any other type of thing, we figure out how 101 00:05:37,286 --> 00:05:40,506 Speaker 2: we get this thing removed. So the removal part is 102 00:05:40,856 --> 00:05:43,676 Speaker 2: another cool piece of tech because it's fully automated. But 103 00:05:43,696 --> 00:05:46,576 Speaker 2: we can reach out to adult sites and tell them 104 00:05:46,636 --> 00:05:50,316 Speaker 2: they are creating liability for their merchant accounts or, you know, 105 00:05:50,336 --> 00:05:53,166 Speaker 2: they're domain provider, or a number of other ways. And 106 00:05:53,186 --> 00:05:55,326 Speaker 2: then on social media, we kind of have this great 107 00:05:55,486 --> 00:05:59,946 Speaker 2: access and great partnership across everyone but X. And they 108 00:05:59,986 --> 00:06:02,096 Speaker 2: just give us a kind of API access to issue 109 00:06:02,126 --> 00:06:02,766 Speaker 2: those takedowns. 110 00:06:03,836 --> 00:06:07,676 Speaker 1: And so does that mean that it's fully automated, like 111 00:06:07,696 --> 00:06:15,036 Speaker 1: your AI, essentially your facial recognition software, spots these unauthorized 112 00:06:15,136 --> 00:06:19,726 Speaker 1: uses of people's likeness and then pings the software or 113 00:06:19,746 --> 00:06:23,526 Speaker 1: the AI of whatever, YouTube or TikTok, and says, hey, 114 00:06:23,586 --> 00:06:25,926 Speaker 1: take this down, and then the machine just takes it down? 115 00:06:25,986 --> 00:06:27,886 Speaker 1: Like there's no human needs to. 116 00:06:28,026 --> 00:06:29,246 Speaker 2: There's no human needed for this. 117 00:06:29,766 --> 00:06:34,066 Speaker 1: And presumably that's because the volume is too high. Like 118 00:06:34,666 --> 00:06:37,796 Speaker 1: what order of magnitude, how many, you know, how many 119 00:06:37,816 --> 00:06:40,856 Speaker 1: things has your company removed, whatever, today or this week? 120 00:06:42,096 --> 00:06:44,696 Speaker 2: We do tens of thousands of things a day. 121 00:06:45,016 --> 00:06:46,416 Speaker 1: Tens of thousands a day. 122 00:06:46,816 --> 00:06:47,816 Speaker 2: Yeah, it's pretty insane. 123 00:06:48,266 --> 00:06:52,626 Speaker 1: That is insane. So, like, at every moment, at every second, 124 00:06:52,726 --> 00:06:56,646 Speaker 1: someone is posting a scam likeness. I guess presumably at 125 00:06:56,666 --> 00:06:59,766 Speaker 1: this point it's machines. Like, you don't need a person. Like, 126 00:06:59,826 --> 00:07:03,246 Speaker 1: I could just build a bot to create a million 127 00:07:03,346 --> 00:07:04,286 Speaker 1: scam likenesses. 128 00:07:04,626 --> 00:07:07,426 Speaker 2: Correct. We called it an industrial game of whack-a-mole. 129 00:07:07,646 --> 00:07:12,186 Speaker 1: Yeah. Like, what's the frontier? What are you struggling with 130 00:07:12,226 --> 00:07:12,646 Speaker 1: right now? 131 00:07:12,666 --> 00:07:16,136 Speaker 2: Hmm. I think a lot of people don't understand that 132 00:07:16,296 --> 00:07:19,346 Speaker 2: you can have delete for the internet. And so we 133 00:07:19,386 --> 00:07:23,846 Speaker 2: think about how we position the company into a greater 134 00:07:23,906 --> 00:07:26,846 Speaker 2: context around like, okay, there's a thing of me online. 135 00:07:27,026 --> 00:07:28,606 Speaker 2: It's going to be online forever. And we kind of 136 00:07:28,626 --> 00:07:30,966 Speaker 2: have to interject ourselves and say, well, this won't be 137 00:07:30,986 --> 00:07:33,686 Speaker 2: online forever. I know everyone was told that for 30 years, 138 00:07:34,216 --> 00:07:36,056 Speaker 2: but now there's a really efficient way for us to 139 00:07:36,096 --> 00:07:36,616 Speaker 2: manage that. 140 00:07:37,296 --> 00:07:38,176 Speaker 1: What's the business model? 141 00:07:38,656 --> 00:07:41,796 Speaker 2: So for celebrities, we charge a monthly fee to go 142 00:07:41,856 --> 00:07:45,766 Speaker 2: scraping and removing this content. And for consumers, it'll be 143 00:07:45,806 --> 00:07:49,446 Speaker 2: more freemium. They can find everything of themselves online, and 144 00:07:49,466 --> 00:07:51,286 Speaker 2: then they get a certain number of takedowns if they 145 00:07:51,326 --> 00:07:53,606 Speaker 2: upgrade to Pro. Or we actually teach them how they 146 00:07:53,646 --> 00:07:56,886 Speaker 2: can go do the takedowns themselves, so we're not trying 147 00:07:56,926 --> 00:07:58,366 Speaker 2: to gatekeep their privacy. 148 00:07:58,826 --> 00:08:00,446 Speaker 1: When's it going to be available for consumers? 149 00:08:00,866 --> 00:08:01,806 Speaker 2: Probably early October. 150 00:08:02,246 --> 00:08:04,406 Speaker 1: Technically, how is it going? 151 00:08:05,826 --> 00:08:10,976 Speaker 2: Oh, it's both a technical nightmare and a marvel. It's insane. 152 00:08:12,236 --> 00:08:15,376 Speaker 2: because the problem set of downloading like 600 million assets 153 00:08:15,436 --> 00:08:18,836 Speaker 2: a day, you have database problems and disk problems. 154 00:08:18,896 --> 00:08:20,536 Speaker 1: Why do you need to download it? 155 00:08:21,176 --> 00:08:23,946 Speaker 2: Because that's how you run the face recognition on it. Okay. 156 00:08:24,146 --> 00:08:25,146 Speaker 1: Interesting. Yeah. 157 00:08:25,466 --> 00:08:28,206 Speaker 2: And so you download it, you run your face recognition 158 00:08:28,266 --> 00:08:31,546 Speaker 2: on it, and then that gives you the capability of saying, 159 00:08:31,586 --> 00:08:33,846 Speaker 2: this person's coming to me. Can you find everything in 160 00:08:33,886 --> 00:08:35,866 Speaker 2: my online? We look into our database. We say, oh, 161 00:08:36,136 --> 00:08:38,476 Speaker 2: we can now identify these rows as you. The other 162 00:08:38,496 --> 00:08:39,256 Speaker 2: things we toss out. 163 00:08:39,496 --> 00:08:44,456 Speaker 1: Yeah. And are the Bad people doing things to make 164 00:08:44,496 --> 00:08:45,156 Speaker 1: it harder for you? 165 00:08:46,456 --> 00:08:48,736 Speaker 2: No, they're making it easier. They're trying to make their 166 00:08:48,976 --> 00:08:53,416 Speaker 2: AI-generated faces more like the person, which actually is an 167 00:08:53,476 --> 00:08:57,016 Speaker 2: alignment to us. It's like the more their deepfakes look realistic, 168 00:08:57,276 --> 00:08:58,816 Speaker 2: the easier it is for us to catch them. 169 00:08:59,196 --> 00:09:03,076 Speaker 1: Oh, that's interesting. And so presumably, I mean, the next 170 00:09:03,156 --> 00:09:06,756 Speaker 1: step is figuring out what are the legitimate uses, right? Correct. 171 00:09:06,776 --> 00:09:08,566 Speaker 1: In an automated way. How do you do that part? 172 00:09:09,086 --> 00:09:11,446 Speaker 2: So that's more of the context understanding of what the 173 00:09:11,506 --> 00:09:14,786 Speaker 2: image is. We actually do that discovery as well. We 174 00:09:14,826 --> 00:09:17,626 Speaker 2: have this thing that we call a semantic understanding of 175 00:09:17,686 --> 00:09:20,146 Speaker 2: every image and video where it comes through and we say, 176 00:09:20,226 --> 00:09:22,206 Speaker 2: what's going on in this? Is someone running in this park? 177 00:09:22,566 --> 00:09:25,626 Speaker 2: Are they walking in the park? Are they doing drugs? 178 00:09:25,746 --> 00:09:27,316 Speaker 2: Are they doing something they shouldn't be doing in this 179 00:09:27,366 --> 00:09:30,216 Speaker 2: picture or image? And then we have this massive collection of, 180 00:09:30,336 --> 00:09:33,936 Speaker 2: we know that it's this celebrity doing this thing. And 181 00:09:33,976 --> 00:09:38,316 Speaker 2: was there like an expectation of privacy here? Or More importantly, 182 00:09:38,416 --> 00:09:40,616 Speaker 2: is this authorized? And they can give us their rules 183 00:09:40,636 --> 00:09:42,196 Speaker 2: around what should or shouldn't be authorized. 184 00:09:43,436 --> 00:09:46,356 Speaker 1: And presumably the giving the rules of what should and 185 00:09:46,376 --> 00:09:49,536 Speaker 1: shouldn't be authorized also needs to be automated, right? Like 186 00:09:49,576 --> 00:09:52,656 Speaker 1: presumably the clients need to say, these are the things 187 00:09:52,736 --> 00:09:54,696 Speaker 1: I am endorsing. It's sort of a white list. Is 188 00:09:54,736 --> 00:09:55,576 Speaker 1: that the system? 189 00:09:55,836 --> 00:09:56,956 Speaker 2: Yes. Absolutely. 190 00:10:00,356 --> 00:10:02,756 Speaker 1: Are there any technical problems you're trying to solve or 191 00:10:02,796 --> 00:10:04,626 Speaker 1: is it technically a sort of solved issue for you? 192 00:10:04,906 --> 00:10:07,126 Speaker 2: Every couple of months we we trip on something that 193 00:10:07,226 --> 00:10:10,216 Speaker 2: is kind of insane. We're like, that's oh, I didn't 194 00:10:10,296 --> 00:10:11,916 Speaker 2: realize we'd even have to solve that problem. 195 00:10:12,276 --> 00:10:13,896 Speaker 1: What's an example? What's an example? 196 00:10:14,336 --> 00:10:16,356 Speaker 2: Well, when we went to process video, we found out 197 00:10:16,376 --> 00:10:18,296 Speaker 2: the reason why no one had done what we do 198 00:10:18,356 --> 00:10:21,136 Speaker 2: before us, like no one did this in 2005 was 199 00:10:21,176 --> 00:10:24,116 Speaker 2: because it's expensive to process video. So like to take 200 00:10:24,156 --> 00:10:26,256 Speaker 2: a video in and actually like look at all the frames. 201 00:10:27,176 --> 00:10:29,736 Speaker 2: even without great machine learning or AI at that time, 202 00:10:29,856 --> 00:10:32,536 Speaker 2: it was just really hard to do that. And so 203 00:10:32,576 --> 00:10:35,076 Speaker 2: we were broke and we couldn't afford to just spool 204 00:10:35,116 --> 00:10:37,336 Speaker 2: up more servers like some really large AI companies are 205 00:10:37,376 --> 00:10:40,376 Speaker 2: able to do. And so necessity was the mother of invention. 206 00:10:40,396 --> 00:10:43,756 Speaker 2: And we invented a method of processing videos that were 207 00:10:43,936 --> 00:10:46,756 Speaker 2: like a thousand times faster than anyone else. And we 208 00:10:46,776 --> 00:10:48,446 Speaker 2: just did this thinking like, well, this is what you 209 00:10:48,486 --> 00:10:50,376 Speaker 2: have to do, right? Like this is just whatever the 210 00:10:50,416 --> 00:10:53,196 Speaker 2: next step is. And it came to find out that we, uh, 211 00:10:53,576 --> 00:10:55,236 Speaker 2: We had to build something that no one else had 212 00:10:55,276 --> 00:10:57,916 Speaker 2: done that others had tried to do and failed. Now 213 00:10:57,976 --> 00:11:01,996 Speaker 2: it's like, it's one of the most guarded, protected algorithms 214 00:11:02,036 --> 00:11:04,636 Speaker 2: at our company because before it was just kind of like, 215 00:11:04,656 --> 00:11:05,536 Speaker 2: it was like floating everywhere. 216 00:11:05,596 --> 00:11:07,716 Speaker 1: Did you patent it or is it a trade secret? 217 00:11:07,736 --> 00:11:09,296 Speaker 2: Oh, interesting. It's a trade secret. 218 00:11:09,516 --> 00:11:11,396 Speaker 1: So you can do a thing and you're not telling 219 00:11:11,466 --> 00:11:12,436 Speaker 1: anybody how to do it. 220 00:11:12,916 --> 00:11:14,886 Speaker 2: Correct. And people know that we do it. They still 221 00:11:14,926 --> 00:11:15,566 Speaker 2: can't figure it out. 222 00:11:16,046 --> 00:11:16,846 Speaker 1: Oh, that's interesting. 223 00:11:16,866 --> 00:11:19,346 Speaker 2: We can prove it. And that's the only part that matters. 224 00:11:19,626 --> 00:11:22,566 Speaker 1: And what it allows you to do is to essentially... 225 00:11:22,976 --> 00:11:26,996 Speaker 1: have AI look at a video much more efficiently and 226 00:11:27,056 --> 00:11:30,656 Speaker 1: cheaply than anybody else can or than anybody else could before? Interesting. 227 00:11:30,676 --> 00:11:31,996 Speaker 2: That's right. 228 00:11:33,336 --> 00:11:35,876 Speaker 1: And presumably you chose not to patent it because when 229 00:11:35,916 --> 00:11:37,716 Speaker 1: you patent something, you have to tell the world how 230 00:11:37,756 --> 00:11:39,586 Speaker 1: you did it and then people can copy it. 231 00:11:39,816 --> 00:11:42,156 Speaker 2: And then they can tweak one part of the mathematics 232 00:11:42,396 --> 00:11:45,286 Speaker 2: and essentially have something like that. And it's hard to 233 00:11:45,566 --> 00:11:47,286 Speaker 2: patent like a pure math solution. 234 00:11:47,546 --> 00:11:50,326 Speaker 1: Yeah. Has anybody figured it out? Can you tell? 235 00:11:50,346 --> 00:11:54,116 Speaker 2: No, it's been two or three years and No competitor 236 00:11:54,136 --> 00:11:57,856 Speaker 2: we've had is able to— they all stumble on video. Like, 237 00:11:57,876 --> 00:12:01,076 Speaker 2: the volume that we could demonstrate is much higher. 238 00:12:01,816 --> 00:12:06,276 Speaker 1: So we've been talking about what you have called elsewhere defense, right? 239 00:12:06,446 --> 00:12:12,446 Speaker 1: Protecting yourself, protecting ourselves from online fakes and scams, people 240 00:12:12,626 --> 00:12:16,886 Speaker 1: using our likeness without our permission. I've also heard you 241 00:12:16,926 --> 00:12:22,136 Speaker 1: talk about offense, which is— you know, choosing under certain 242 00:12:22,176 --> 00:12:27,616 Speaker 1: circumstances to license our image and likeness for use on 243 00:12:27,656 --> 00:12:29,956 Speaker 1: the internet. I like, I've heard you talk about that 244 00:12:30,136 --> 00:12:33,966 Speaker 1: as this big, you know, underrated asset class that is 245 00:12:33,986 --> 00:12:37,226 Speaker 1: going to blossom. Tell me about that. 246 00:12:37,866 --> 00:12:40,186 Speaker 2: All right. So if you are a big gen AI 247 00:12:40,246 --> 00:12:42,826 Speaker 2: company right now, you either have the choice of trying 248 00:12:42,846 --> 00:12:45,206 Speaker 2: to do a bunch of one-off deals with celebrities, which 249 00:12:45,286 --> 00:12:49,046 Speaker 2: is kind of painful. It's time-consuming. And there's a lot 250 00:12:49,066 --> 00:12:51,576 Speaker 2: of red lines. And then most of the time they 251 00:12:51,596 --> 00:12:53,636 Speaker 2: get stuck in the fact that like, well, who's watching 252 00:12:53,676 --> 00:12:55,476 Speaker 2: you to make sure that bad stuff isn't being made? 253 00:12:55,936 --> 00:12:58,296 Speaker 2: And the tech companies end up saying, well, I'll watch myself. 254 00:12:58,676 --> 00:13:01,416 Speaker 1: There is an example where OpenAI for a while had 255 00:13:01,436 --> 00:13:06,066 Speaker 1: this video generation app called Sora, right? They made this 256 00:13:06,166 --> 00:13:11,706 Speaker 1: deal with Disney where people using Sora could use Disney characters, 257 00:13:11,766 --> 00:13:14,426 Speaker 1: could generate videos with Disney characters. This is the kind 258 00:13:14,466 --> 00:13:15,726 Speaker 1: of thing you're talking about, right? 259 00:13:15,786 --> 00:13:18,066 Speaker 2: Yes. And this is actually a great example because that 260 00:13:18,146 --> 00:13:20,336 Speaker 2: particular deal never really got off the ground because it 261 00:13:20,376 --> 00:13:22,476 Speaker 2: took them months to figure out, well, how do the 262 00:13:22,536 --> 00:13:23,556 Speaker 2: guardrails work for that? 263 00:13:23,816 --> 00:13:25,976 Speaker 1: I thought it never got off the ground because OpenAI 264 00:13:25,996 --> 00:13:28,516 Speaker 1: decided to shut down Sora and focus on other things. 265 00:13:28,556 --> 00:13:31,836 Speaker 2: Well, they shut it down three months after. Right. So 266 00:13:32,156 --> 00:13:34,256 Speaker 2: within the three-month period, they still didn't get a single 267 00:13:34,296 --> 00:13:35,106 Speaker 2: character out the door. 268 00:13:35,286 --> 00:13:35,586 Speaker 1: Uh-huh. 269 00:13:35,986 --> 00:13:37,446 Speaker 2: And so a lot of that had to do with 270 00:13:37,486 --> 00:13:40,146 Speaker 2: the guardrail conversations around, like, well, who protects what? 271 00:13:40,426 --> 00:13:42,646 Speaker 1: Why is this a deal everybody wants to make? Like, 272 00:13:42,686 --> 00:13:44,486 Speaker 1: what's the context here? 273 00:13:45,026 --> 00:13:48,756 Speaker 2: Oh, yeah. Because people want to generate the real world, right? 274 00:13:49,276 --> 00:13:51,626 Speaker 2: So I think that we are going to use this 275 00:13:51,686 --> 00:13:55,386 Speaker 2: as a way to collaboratively dream. Like, let's build this 276 00:13:55,546 --> 00:13:58,526 Speaker 2: story together. Yeah. Or let's build a product together or 277 00:13:58,546 --> 00:14:00,766 Speaker 2: a company or something. And you'll be using Gen AI 278 00:14:00,826 --> 00:14:04,076 Speaker 2: tools to dream those things with others. It'll be important 279 00:14:04,096 --> 00:14:07,316 Speaker 2: in that context to use what the real world looks like. Otherwise, 280 00:14:07,396 --> 00:14:10,536 Speaker 2: the fake version won't quite be the dream that you 281 00:14:10,576 --> 00:14:11,056 Speaker 2: had in mind. 282 00:14:11,756 --> 00:14:15,746 Speaker 1: So you'll... you'll want to make a little movie with 283 00:14:15,826 --> 00:14:18,986 Speaker 1: a friend using generative AI, and you maybe will want 284 00:14:19,006 --> 00:14:21,216 Speaker 1: to have a famous person in it, or you'll want 285 00:14:21,256 --> 00:14:25,196 Speaker 1: it to be at whatever, Starbucks, or at some particular 286 00:14:25,356 --> 00:14:28,516 Speaker 1: store that exists in the world. And if you're actually 287 00:14:28,536 --> 00:14:31,236 Speaker 1: going to sell that or post it online and have, 288 00:14:31,656 --> 00:14:34,526 Speaker 1: you know, have ads run in the middle of it, 289 00:14:34,896 --> 00:14:36,786 Speaker 1: you're not allowed to do it. You're not allowed to 290 00:14:36,826 --> 00:14:38,906 Speaker 1: have it be at an actual Starbucks unless you get 291 00:14:38,966 --> 00:14:40,046 Speaker 1: clearance from Starbucks. 292 00:14:40,566 --> 00:14:41,506 Speaker 2: Correct. Yes. 293 00:14:42,836 --> 00:14:47,576 Speaker 1: So how, like, right now are your celebrity clients selling 294 00:14:47,616 --> 00:14:50,556 Speaker 1: their name, image, and likeness, like, broadly? And are you 295 00:14:50,776 --> 00:14:51,636 Speaker 1: working on that with them? 296 00:14:52,296 --> 00:14:57,436 Speaker 2: Yeah. So you have kind of a bifurcation, which is 297 00:14:57,456 --> 00:15:00,036 Speaker 2: half are like, no, nothing with AI. And the other 298 00:15:00,056 --> 00:15:01,796 Speaker 2: half are like, it's coming and it's here and this 299 00:15:01,836 --> 00:15:03,916 Speaker 2: is something we have to get ahead of. So let's 300 00:15:03,976 --> 00:15:06,536 Speaker 2: make rules and set pricing that makes sense for me. 301 00:15:07,196 --> 00:15:09,916 Speaker 2: And so the interchange system, the clearinghouse system works for both. 302 00:15:10,726 --> 00:15:12,486 Speaker 2: One just blocks everything and the other one says, yes, 303 00:15:12,506 --> 00:15:13,526 Speaker 2: you can make it with these rules. 304 00:15:14,466 --> 00:15:17,426 Speaker 1: So are there examples that are public of people who 305 00:15:17,486 --> 00:15:20,646 Speaker 1: have allowed their likeness to be used in generative AI? 306 00:15:20,886 --> 00:15:23,726 Speaker 2: So like right now we're basically onboarding everyone. We have 307 00:15:23,746 --> 00:15:26,286 Speaker 2: two different gen AI providers that are going to be 308 00:15:26,346 --> 00:15:29,756 Speaker 2: using it. So the announcements for who you can generate 309 00:15:29,796 --> 00:15:31,456 Speaker 2: where will have to come at a different point. 310 00:15:32,236 --> 00:15:33,996 Speaker 1: Give me a sense of what that's going to look like. 311 00:15:34,096 --> 00:15:36,176 Speaker 1: Will there be like lots of movie stars I have 312 00:15:36,236 --> 00:15:39,256 Speaker 1: heard of who are saying it's okay to use my likeness? 313 00:15:40,886 --> 00:15:43,746 Speaker 1: Is Gen AI subject to certain restrictions? 314 00:15:44,266 --> 00:15:47,096 Speaker 2: I think you'll see a certain tier, for lack of 315 00:15:47,166 --> 00:15:51,836 Speaker 2: a better term, we'll say like B celebrities, a large 316 00:15:51,896 --> 00:15:54,156 Speaker 2: group of them. And you'll have kind of a sprinkling 317 00:15:54,176 --> 00:15:56,456 Speaker 2: of a few A stars. And then you'll be able 318 00:15:56,496 --> 00:15:58,236 Speaker 2: to just go to the Gen AI platforms and just 319 00:15:58,276 --> 00:16:01,586 Speaker 2: type in this person's name instead of it getting immediately rejected. 320 00:16:02,066 --> 00:16:03,566 Speaker 2: It says, okay, you can do that. But instead of 321 00:16:03,586 --> 00:16:04,266 Speaker 2: it being $ 0. 30, it's $ 0. 35 online. 322 00:16:05,806 --> 00:16:10,566 Speaker 1: Uh-huh. So basically you're saying that this platform charges you 323 00:16:10,846 --> 00:16:14,946 Speaker 1: extra to use a real famous person in the video 324 00:16:14,966 --> 00:16:17,446 Speaker 1: that you're generating and the real famous person gets some 325 00:16:17,486 --> 00:16:19,096 Speaker 1: of that money and you get a little tiny bit 326 00:16:19,116 --> 00:16:20,256 Speaker 1: of that money, presumably. 327 00:16:20,796 --> 00:16:22,936 Speaker 2: Um, well, I mean, we, we actually don't charge our clients. 328 00:16:22,956 --> 00:16:24,356 Speaker 2: We charge the gen AI platforms. 329 00:16:24,616 --> 00:16:24,856 Speaker 1: Uh-huh. 330 00:16:25,676 --> 00:16:28,576 Speaker 2: So if it's 5 cents, that, that goes to the client. 331 00:16:28,896 --> 00:16:30,576 Speaker 2: And then we have a different mechanism of how we 332 00:16:30,616 --> 00:16:31,756 Speaker 2: charge the gen AI companies. 333 00:16:32,286 --> 00:16:35,236 Speaker 1: And what kind of restrictions to your clients will there 334 00:16:35,296 --> 00:16:38,596 Speaker 1: be on the use of their likeness in generative videos? 335 00:16:38,936 --> 00:16:41,696 Speaker 2: There's kind of like a belief system. So you have 336 00:16:41,716 --> 00:16:47,376 Speaker 2: things like religion. And then more tangibly, it's things around sex, drugs, 337 00:16:47,396 --> 00:16:49,956 Speaker 2: and alcohol. And then they're all natural language. So you 338 00:16:49,976 --> 00:16:53,186 Speaker 2: can turn on drugs and say cannabis is just fine. 339 00:16:53,826 --> 00:16:58,566 Speaker 2: Like I'm okay with this. And there's financial products. Like 340 00:16:58,606 --> 00:17:01,026 Speaker 2: some people, they're consistently made to make kind of like 341 00:17:01,066 --> 00:17:03,586 Speaker 2: Bitcoin scams. And so you can have that turned off 342 00:17:03,646 --> 00:17:06,046 Speaker 2: entirely or say, yes, but it's only for this one 343 00:17:06,266 --> 00:17:09,236 Speaker 2: app that I'm being sponsored for. And then the last 344 00:17:09,266 --> 00:17:12,996 Speaker 2: part is literally the sponsorships, right? So you have, uh, 345 00:17:13,236 --> 00:17:15,356 Speaker 2: you know, Nike is sponsoring me. So that always goes 346 00:17:15,436 --> 00:17:18,676 Speaker 2: on a, on a sneaker or, you know, I did 347 00:17:18,716 --> 00:17:20,836 Speaker 2: a deal with hot pockets a few years ago. So 348 00:17:20,856 --> 00:17:23,016 Speaker 2: you can't ever make some kind of claim of me 349 00:17:23,076 --> 00:17:24,836 Speaker 2: saying that hot pockets burn my mouth for the next 350 00:17:24,856 --> 00:17:25,356 Speaker 2: year and a half. 351 00:17:25,536 --> 00:17:27,516 Speaker 1: Yes. What if I want to make a video of 352 00:17:27,556 --> 00:17:29,986 Speaker 1: a famous person saying that they love my shampoo? Uh, 353 00:17:31,876 --> 00:17:33,536 Speaker 1: that I'm selling. Can I do that? 354 00:17:34,536 --> 00:17:37,856 Speaker 2: No, we do understand the difference between whether it's commercialized 355 00:17:37,916 --> 00:17:40,696 Speaker 2: or not. And this is where the defense part comes in, 356 00:17:40,736 --> 00:17:43,756 Speaker 2: which is, let's say that you have them say, happy birthday, 357 00:17:43,996 --> 00:17:45,736 Speaker 2: and you go start a car wash called the happy 358 00:17:45,776 --> 00:17:49,056 Speaker 2: birthday car wash. There's no point when you generate it 359 00:17:49,076 --> 00:17:51,476 Speaker 2: that I would know that that was what the intent was. 360 00:17:51,946 --> 00:17:54,526 Speaker 2: But when it actually gets used out on Instagram, I 361 00:17:54,556 --> 00:17:56,586 Speaker 2: will see, oh, this is used in a commercialized way. 362 00:17:56,626 --> 00:17:59,406 Speaker 2: I'll look up your license and say, oh, that's That 363 00:17:59,466 --> 00:18:01,726 Speaker 2: wasn't how it was purchased. This was a trickery thing. 364 00:18:01,906 --> 00:18:04,186 Speaker 2: So then we go ahead and delete that. So they 365 00:18:04,226 --> 00:18:06,506 Speaker 2: end up having to buy the correct version, but maybe 366 00:18:06,526 --> 00:18:07,966 Speaker 2: they get away with it for an hour or two. 367 00:18:12,006 --> 00:18:25,336 Speaker 1: We'll be back in just a minute. Are there any 368 00:18:25,416 --> 00:18:29,036 Speaker 1: instances where deepfakes are legal? Or the First Amendment might 369 00:18:29,536 --> 00:18:30,096 Speaker 1: be relevant? 370 00:18:31,706 --> 00:18:36,386 Speaker 2: So this is tricky. I would say that no fakes 371 00:18:36,446 --> 00:18:39,346 Speaker 2: is coming down the pipeline, which is a federal act. 372 00:18:39,806 --> 00:18:42,166 Speaker 2: And that is going to require authorization from the person 373 00:18:42,186 --> 00:18:44,266 Speaker 2: that you're deep faking. I think this is a really 374 00:18:44,476 --> 00:18:47,516 Speaker 2: solid place for us to operate from. So between now 375 00:18:47,536 --> 00:18:49,636 Speaker 2: and when that passes, and it looks like it's going 376 00:18:49,656 --> 00:18:53,076 Speaker 2: to pass, there's really strong support on both sides of 377 00:18:53,096 --> 00:18:55,986 Speaker 2: the aisles from tech companies and entertainment. It's kind of 378 00:18:56,026 --> 00:18:59,066 Speaker 2: a massive collaborative multi-year effort, and it looks like it'll 379 00:18:59,086 --> 00:19:02,306 Speaker 2: probably pass early next year. It's going to force tech 380 00:19:02,346 --> 00:19:07,056 Speaker 2: companies to have to ask for authorizations from my neighbor 381 00:19:07,396 --> 00:19:09,976 Speaker 2: or a celebrity down the street. Everyone is going to 382 00:19:09,996 --> 00:19:13,356 Speaker 2: need some kind of authorization system for that. And between 383 00:19:13,396 --> 00:19:15,296 Speaker 2: now and then, you can make a case that it's 384 00:19:15,336 --> 00:19:18,226 Speaker 2: illegal for right of publicity reasons. If you generate a 385 00:19:18,846 --> 00:19:22,026 Speaker 2: deepfake endorsing someone, that's like a Lanham Act violation because 386 00:19:22,046 --> 00:19:24,476 Speaker 2: it's false endorsement. So there's the ways in which it's 387 00:19:24,576 --> 00:19:25,436 Speaker 2: not about the tech. 388 00:19:25,476 --> 00:19:27,716 Speaker 1: Well, the endorsement is the easy case. I mean, the 389 00:19:27,796 --> 00:19:30,726 Speaker 1: one I'm asking you, like, say you do a politician, 390 00:19:30,886 --> 00:19:32,966 Speaker 1: a satirical video of a politician. 391 00:19:33,046 --> 00:19:33,246 Speaker 2: Right. 392 00:19:33,346 --> 00:19:36,766 Speaker 1: Like, yes, we know that satire tends to have more 393 00:19:36,846 --> 00:19:41,566 Speaker 1: First Amendment protection. We know that discussion of political figures 394 00:19:41,626 --> 00:19:43,366 Speaker 1: tends to have more First Amendment protection. 395 00:19:43,386 --> 00:19:43,846 Speaker 2: Yes. Right. 396 00:19:43,926 --> 00:19:47,826 Speaker 1: Like. I don't know. I can understand why you could 397 00:19:47,846 --> 00:19:49,636 Speaker 1: make a case that a deepfake is not like a 398 00:19:49,676 --> 00:19:53,156 Speaker 1: political cartoon, but you could make a case that it is. 399 00:19:53,176 --> 00:19:56,136 Speaker 2: I think I would probably more make the case that 400 00:19:56,196 --> 00:20:00,076 Speaker 2: it is. I think there will be protections there. I 401 00:20:00,376 --> 00:20:01,816 Speaker 2: know that when we talk to both sides of the 402 00:20:01,926 --> 00:20:04,586 Speaker 2: aisle on who we protect, making sure that we get 403 00:20:04,626 --> 00:20:07,606 Speaker 2: First Amendment right is a big concern for everyone. And 404 00:20:07,666 --> 00:20:08,486 Speaker 2: us included. 405 00:20:08,546 --> 00:20:10,716 Speaker 1: I mean, it's a hard one. It's a hard one because. 406 00:20:12,276 --> 00:20:15,286 Speaker 1: As a matter of sort of public interest, you want 407 00:20:15,306 --> 00:20:17,526 Speaker 1: people to understand what is real and what is not real. 408 00:20:17,966 --> 00:20:18,386 Speaker 2: Yes. Right? 409 00:20:18,666 --> 00:20:20,966 Speaker 1: Like, I like satire. I'm a reporter. I like the 410 00:20:20,986 --> 00:20:24,206 Speaker 1: First Amendment. But, like, people being confused about whether the 411 00:20:24,246 --> 00:20:27,116 Speaker 1: president is really saying something or it's a deepfake of 412 00:20:27,136 --> 00:20:28,716 Speaker 1: the president, that seems quite bad. 413 00:20:29,556 --> 00:20:31,456 Speaker 2: I do think that it's going to be. So when 414 00:20:31,536 --> 00:20:33,716 Speaker 2: you did a political cartoon before, it was in the 415 00:20:33,736 --> 00:20:37,136 Speaker 2: political cartoon section. of the newspaper or The Economist or whatever. 416 00:20:37,436 --> 00:20:39,756 Speaker 1: And it's a drawing. Like, no one is going to 417 00:20:39,776 --> 00:20:40,776 Speaker 1: be confused. Yeah. 418 00:20:40,796 --> 00:20:42,406 Speaker 2: Well, I think that there'll just be kind of a 419 00:20:42,426 --> 00:20:44,546 Speaker 2: time and place in the internet for that, which is 420 00:20:44,566 --> 00:20:46,986 Speaker 2: there'll either be a certain part of the site or 421 00:20:47,226 --> 00:20:49,746 Speaker 2: it will be labeled clearly on the video. Right? There'll 422 00:20:49,766 --> 00:20:50,306 Speaker 2: be a restriction. 423 00:20:50,386 --> 00:20:52,356 Speaker 1: Yeah, like a label. A label is good. 424 00:20:52,636 --> 00:20:54,436 Speaker 2: If you labeled it, it would be just fine. 425 00:20:54,936 --> 00:20:57,176 Speaker 1: Is that functional? Like, will that work? 426 00:20:57,676 --> 00:21:00,256 Speaker 2: Well, this is This is where, you know, and not 427 00:21:00,296 --> 00:21:03,256 Speaker 2: to harp on us or to focus on what we do, 428 00:21:03,316 --> 00:21:05,956 Speaker 2: but anyone really that wants to build a business like ours, 429 00:21:06,456 --> 00:21:08,296 Speaker 2: what you would say is what the time it goes 430 00:21:08,336 --> 00:21:10,116 Speaker 2: to generate it, you'd say, oh, I see that you're 431 00:21:10,136 --> 00:21:12,356 Speaker 2: actually making a satirical thing. That doesn't cost you anything 432 00:21:12,376 --> 00:21:14,296 Speaker 2: because that's your first amendment, right? So we're not going 433 00:21:14,316 --> 00:21:16,256 Speaker 2: to charge you. You get a bypass on it. But 434 00:21:16,276 --> 00:21:17,416 Speaker 2: then when we see it out in the wild, if 435 00:21:17,456 --> 00:21:20,096 Speaker 2: you don't have labeling on it, or if you're not 436 00:21:20,236 --> 00:21:22,466 Speaker 2: using it in a satirical way, you're using it to like, say, 437 00:21:22,516 --> 00:21:25,226 Speaker 2: look at what the president said, right? You're trying to 438 00:21:25,246 --> 00:21:27,126 Speaker 2: like make a stance on it and trying to, you know, 439 00:21:27,226 --> 00:21:29,236 Speaker 2: push it off as being real. then we'll do a 440 00:21:29,256 --> 00:21:31,336 Speaker 2: takedown on it. So there's a way that you kind 441 00:21:31,376 --> 00:21:33,126 Speaker 2: of need both systems to work well. 442 00:21:33,796 --> 00:21:38,086 Speaker 1: That makes sense. You mentioned the federal legislation. I've heard 443 00:21:38,126 --> 00:21:41,486 Speaker 1: you talk elsewhere about state laws as well. Is the 444 00:21:41,546 --> 00:21:45,646 Speaker 1: federal legislation going to sort of preempt or render irrelevant 445 00:21:45,686 --> 00:21:47,846 Speaker 1: the state laws? Are there state laws that are important 446 00:21:47,866 --> 00:21:50,376 Speaker 1: that are doing other things that are worth talking about? 447 00:21:50,496 --> 00:21:52,416 Speaker 1: I mean, I seem to recall California being kind of 448 00:21:52,476 --> 00:21:55,216 Speaker 1: at the vanguard of name, image, and likeness stuff in 449 00:21:55,236 --> 00:21:58,076 Speaker 1: the 20th century because of Hollywood, basically, right? 450 00:21:58,326 --> 00:22:00,656 Speaker 2: Yeah. And then, oh, Tennessee is right now with the 451 00:22:00,736 --> 00:22:01,536 Speaker 2: Elvis Act. 452 00:22:01,756 --> 00:22:01,936 Speaker 1: Huh. 453 00:22:02,336 --> 00:22:04,676 Speaker 2: They've made the No Fakes Act essentially for the state. 454 00:22:05,176 --> 00:22:05,496 Speaker 1: I see. 455 00:22:05,936 --> 00:22:08,436 Speaker 2: So if I have a client in Nashville, I have 456 00:22:08,716 --> 00:22:11,396 Speaker 2: a lot of power for them to go and get 457 00:22:11,416 --> 00:22:14,696 Speaker 2: things taken down. If someone made just a benign someone 458 00:22:14,716 --> 00:22:17,336 Speaker 2: walking down the street in AI, if they live in Tennessee, 459 00:22:17,386 --> 00:22:20,286 Speaker 2: it gets taken down. If they don't, I need to 460 00:22:20,326 --> 00:22:22,426 Speaker 2: make a pretty good case about why this harms that 461 00:22:22,466 --> 00:22:24,306 Speaker 2: person and not use AI rules to do it. 462 00:22:25,446 --> 00:22:28,996 Speaker 1: Oh, interesting. So just to be clear- Outside of Tennessee, 463 00:22:29,036 --> 00:22:31,326 Speaker 1: which is to say in most of the country at 464 00:22:31,366 --> 00:22:34,766 Speaker 1: this point, if I make a deep fake of whatever, 465 00:22:34,926 --> 00:22:42,756 Speaker 1: George Clooney, and he's just like, whatever, reciting Shakespeare or something. And, well, 466 00:22:42,976 --> 00:22:43,656 Speaker 1: is that okay? 467 00:22:45,596 --> 00:22:47,756 Speaker 2: Yeah, I mean, it would be allowed. Like, we wouldn't 468 00:22:47,776 --> 00:22:48,956 Speaker 2: be able to necessarily take that down. 469 00:22:49,526 --> 00:22:51,936 Speaker 1: If I sell ads against it, if they're just like 470 00:22:52,076 --> 00:22:54,046 Speaker 1: programmatic ads that show up in the middle of it, 471 00:22:54,316 --> 00:22:55,436 Speaker 1: now can you take it down? 472 00:22:55,536 --> 00:22:57,316 Speaker 2: That's right. It's a violation of right of publicity. 473 00:22:57,656 --> 00:22:58,956 Speaker 1: Because I'm making money off it. 474 00:22:59,196 --> 00:23:00,236 Speaker 2: Right, off of their likeness. 475 00:23:00,516 --> 00:23:03,336 Speaker 1: But if this bill becomes law, then it won't be 476 00:23:03,386 --> 00:23:03,946 Speaker 1: okay anymore. 477 00:23:04,256 --> 00:23:05,846 Speaker 2: Right. Those people will be able to pick what gets 478 00:23:05,906 --> 00:23:08,826 Speaker 2: made of them. With a 1A carve-out, right? 479 00:23:09,366 --> 00:23:10,906 Speaker 1: The First Amendment carve-out, potentially. 480 00:23:10,946 --> 00:23:11,386 Speaker 2: Yeah. 481 00:23:11,566 --> 00:23:13,846 Speaker 1: It's interesting that it's okay to do it now. I 482 00:23:13,886 --> 00:23:16,046 Speaker 1: think that is not intuitive. Yeah. 483 00:23:16,176 --> 00:23:18,716 Speaker 2: Well, I don't want to encourage people to do it. Um, 484 00:23:18,856 --> 00:23:20,776 Speaker 2: you literally, like these people are upset that you were 485 00:23:20,796 --> 00:23:24,076 Speaker 2: stealing part of their private nature, right? Because the people 486 00:23:24,096 --> 00:23:26,516 Speaker 2: are also doing it for people that aren't celebrities. Um, 487 00:23:26,616 --> 00:23:29,176 Speaker 2: you know, doing it, there's deep fakes made of, uh, 488 00:23:29,276 --> 00:23:31,996 Speaker 2: teachers saying crazy things and then them losing their jobs. 489 00:23:32,016 --> 00:23:35,356 Speaker 2: Like there's still harm, right? That comes from it. So 490 00:23:35,376 --> 00:23:37,856 Speaker 2: I don't want to really tell people go on the 491 00:23:37,906 --> 00:23:40,506 Speaker 2: boom rush now and go make it while it's legal. Um, 492 00:23:40,566 --> 00:23:42,426 Speaker 2: you still are creating harm, even if it may not 493 00:23:42,466 --> 00:23:42,786 Speaker 2: be a law. 494 00:23:43,526 --> 00:23:46,416 Speaker 1: Is there a universe where the big platforms just do 495 00:23:46,476 --> 00:23:50,576 Speaker 1: this themselves? Like, why when I sign up for whatever, Instagram, 496 00:23:50,716 --> 00:23:53,536 Speaker 1: can I not just say, no deep fakes of me? 497 00:23:53,636 --> 00:23:54,916 Speaker 1: Why do we need an extra company? 498 00:23:55,576 --> 00:23:58,406 Speaker 2: So every platform, most platforms, and we're good partners with them, 499 00:23:58,466 --> 00:24:00,086 Speaker 2: so I don't want to be disparaging about this, but 500 00:24:00,186 --> 00:24:02,166 Speaker 2: they've tried over the last few years to do this, 501 00:24:02,686 --> 00:24:06,486 Speaker 2: and there's issues with doing it. One is, it's hard 502 00:24:06,526 --> 00:24:09,776 Speaker 2: for them to set up similar rules across all systems. Two, 503 00:24:10,076 --> 00:24:12,626 Speaker 2: it's hard for anyone to believe that they will properly 504 00:24:12,666 --> 00:24:13,546 Speaker 2: gatekeep themselves. 505 00:24:14,066 --> 00:24:17,346 Speaker 1: Because the platforms want more content. They're not incentivized to 506 00:24:17,386 --> 00:24:18,746 Speaker 1: take things down. Right. 507 00:24:18,766 --> 00:24:20,646 Speaker 2: And so you're like, hey, you as the gatekeeper, can 508 00:24:20,666 --> 00:24:22,466 Speaker 2: you generate things? And then if you're a celebrity, it's 509 00:24:22,486 --> 00:24:25,706 Speaker 2: even more complicated because you're saying a biometric release is 510 00:24:25,766 --> 00:24:28,046 Speaker 2: needed for that. And most humans do not want to 511 00:24:28,066 --> 00:24:30,326 Speaker 2: have Google give a full biometric release because they know 512 00:24:30,396 --> 00:24:32,476 Speaker 2: everything about you, right? You also don't need to know 513 00:24:32,506 --> 00:24:35,536 Speaker 2: what you sound like and look like. And these large 514 00:24:35,576 --> 00:24:37,776 Speaker 2: tech companies have been sued for face recognition usage in 515 00:24:37,796 --> 00:24:41,156 Speaker 2: the past because of how they have this weird interstitial 516 00:24:41,196 --> 00:24:43,516 Speaker 2: data set of like everything you've ever gone to online 517 00:24:43,646 --> 00:24:45,886 Speaker 2: and what you look like in real life. Places like 518 00:24:45,946 --> 00:24:48,926 Speaker 2: Illinois and Texas have sued them. So platforms aren't in 519 00:24:48,966 --> 00:24:51,046 Speaker 2: a great place to say, I'll be the person that 520 00:24:51,066 --> 00:24:53,906 Speaker 2: steps up and does this because the technology is largely 521 00:24:54,046 --> 00:24:57,486 Speaker 2: handicapped for them because of this past litigation. And then 522 00:24:57,506 --> 00:24:59,226 Speaker 2: the other part is people don't want to trust them 523 00:24:59,266 --> 00:25:03,096 Speaker 2: with the biometrics. Those releases are pretty insane to think about. 524 00:25:03,576 --> 00:25:05,796 Speaker 2: For us, like this is our only job, right? It's 525 00:25:05,916 --> 00:25:08,416 Speaker 2: easy to give Lodi those biometric releases and Because all 526 00:25:08,456 --> 00:25:10,036 Speaker 2: we're doing is protecting your biometrics. 527 00:25:10,656 --> 00:25:16,016 Speaker 1: I've heard you talk about user agreements as weirdly important 528 00:25:16,156 --> 00:25:19,836 Speaker 1: in this broader conversation. Why? What's going on there? 529 00:25:20,756 --> 00:25:22,886 Speaker 2: There's a land grab for AI rights right now, for 530 00:25:22,926 --> 00:25:26,466 Speaker 2: likeness rights. So if you're a tech company, it doesn't matter. 531 00:25:26,486 --> 00:25:28,746 Speaker 2: It's not even the big ones, even small startups. They're 532 00:25:28,766 --> 00:25:30,646 Speaker 2: trying to figure out how to get as many people 533 00:25:30,686 --> 00:25:33,366 Speaker 2: as possible to do releases on their data, their likeness, 534 00:25:33,386 --> 00:25:36,436 Speaker 2: their biometrics, because that data is really valuable. 535 00:25:36,826 --> 00:25:39,416 Speaker 1: And like if I'm just a normal guy with an iPhone, 536 00:25:39,456 --> 00:25:42,416 Speaker 1: like when I do unlock my phone with my face, 537 00:25:42,516 --> 00:25:44,726 Speaker 1: am I participating in this without knowing it? 538 00:25:45,116 --> 00:25:48,176 Speaker 2: No, but actually Apple, if you use an iPhone, they 539 00:25:48,256 --> 00:25:49,996 Speaker 2: are aware of the same kind of problem sets that 540 00:25:50,016 --> 00:25:51,556 Speaker 2: we are, which is they make sure that that's all 541 00:25:51,616 --> 00:25:52,316 Speaker 2: local to the phone. 542 00:25:52,716 --> 00:25:55,536 Speaker 1: Oh, interesting. So like Apple doesn't know what my face 543 00:25:55,576 --> 00:25:57,796 Speaker 1: looks like. It's just my phone that knows what my 544 00:25:57,836 --> 00:25:59,136 Speaker 1: face looks like. And that's good. 545 00:25:59,596 --> 00:26:01,466 Speaker 2: There's a world that's coming in which all that's going 546 00:26:01,486 --> 00:26:02,326 Speaker 2: to get legislated. 547 00:26:02,666 --> 00:26:03,006 Speaker 1: I see. 548 00:26:03,166 --> 00:26:05,286 Speaker 2: I think that's the next thing after no fakes. Now 549 00:26:05,306 --> 00:26:08,046 Speaker 2: we're all going to decide what constitutes the data that 550 00:26:08,086 --> 00:26:11,066 Speaker 2: goes into making a fake, right? I think it's coming 551 00:26:11,766 --> 00:26:14,726 Speaker 2: and it's probably going to hurt a lot of smaller companies, 552 00:26:14,746 --> 00:26:16,446 Speaker 2: but the bigger ones now are starting to see that 553 00:26:16,526 --> 00:26:18,086 Speaker 2: is what's on the horizon and they might as well 554 00:26:18,106 --> 00:26:19,526 Speaker 2: just design their program accordingly. 555 00:26:20,226 --> 00:26:23,696 Speaker 1: Yeah, I mean, often regulation... ends up being a moat 556 00:26:23,796 --> 00:26:26,836 Speaker 1: for big companies, right? It is essentially a barrier to entry. 557 00:26:26,896 --> 00:26:30,056 Speaker 1: It makes it more expensive to do things. And so ultimately, 558 00:26:30,096 --> 00:26:32,316 Speaker 1: it's good for the big incumbents. 559 00:26:32,396 --> 00:26:36,596 Speaker 2: That's absolutely true. You have that plus regulatory capture. And 560 00:26:36,936 --> 00:26:39,816 Speaker 2: it's definitely something that we'll see in the AI industry soon. 561 00:26:40,376 --> 00:26:47,506 Speaker 1: Interesting. Let's talk more about the future. We've talked about legislation. 562 00:26:47,546 --> 00:26:51,466 Speaker 1: We've talked about technology. We've talked about rights. But like, 563 00:26:52,706 --> 00:26:58,476 Speaker 1: When you think, uh, you know, about whatever amount of 564 00:26:58,536 --> 00:27:00,356 Speaker 1: time it makes sense to think about five years out, 565 00:27:00,376 --> 00:27:01,976 Speaker 1: I don't know, 10 years, who knows what's going to 566 00:27:01,996 --> 00:27:03,996 Speaker 1: go on in 10 years, right? It's so crazy. It's 567 00:27:04,016 --> 00:27:07,896 Speaker 1: so fast. Um, first of all, what do you worry about? 568 00:27:10,706 --> 00:27:17,046 Speaker 2: Man. Um, I worry about us being stuck in 2024 569 00:27:17,046 --> 00:27:20,186 Speaker 2: art and culture for a long time because of how 570 00:27:20,246 --> 00:27:23,776 Speaker 2: machine learning is trained that, uh, You know, you basically 571 00:27:23,816 --> 00:27:26,366 Speaker 2: have the corpus of human creativity that built models a 572 00:27:26,406 --> 00:27:28,926 Speaker 2: couple of years ago that we all started using. And 573 00:27:28,946 --> 00:27:31,106 Speaker 2: then everything that comes out of those models will be 574 00:27:31,186 --> 00:27:34,346 Speaker 2: like what the past thousands of years of creative look like. 575 00:27:35,066 --> 00:27:36,766 Speaker 2: And what we won't have is something that kind of 576 00:27:36,806 --> 00:27:39,726 Speaker 2: jumps out. A lot of machine learning is based on 577 00:27:39,806 --> 00:27:42,326 Speaker 2: training on the top of the bell curve of items, 578 00:27:42,346 --> 00:27:45,646 Speaker 2: like what looks like the most, like 50th to 75th percentile. 579 00:27:45,886 --> 00:27:48,206 Speaker 1: Sort of the average, the average. Yeah. 580 00:27:48,536 --> 00:27:50,886 Speaker 2: So you have the average of 2024 that's coming out 581 00:27:50,906 --> 00:27:52,886 Speaker 2: of models now. And you have to have something that 582 00:27:52,906 --> 00:27:55,306 Speaker 2: looks truly unique. And AI doesn't want to make that. 583 00:27:55,446 --> 00:27:57,566 Speaker 2: It doesn't know how to. And that's where the crap 584 00:27:57,626 --> 00:27:59,526 Speaker 2: lives on one side. And that's where like the really 585 00:27:59,586 --> 00:28:00,986 Speaker 2: amazing stuff lives on the other. 586 00:28:02,346 --> 00:28:03,906 Speaker 1: Of the bell curve. You're talking about sort of the 587 00:28:03,986 --> 00:28:06,486 Speaker 1: tails of the distribution, things that are really bad and 588 00:28:06,506 --> 00:28:07,506 Speaker 1: things that are really good. 589 00:28:07,586 --> 00:28:11,176 Speaker 2: Correct. Right. So you have like this mathematical preventative thing 590 00:28:11,216 --> 00:28:13,596 Speaker 2: that says, hey, I don't know how to do math 591 00:28:13,616 --> 00:28:16,826 Speaker 2: that allows the really cool stuff in without letting the 592 00:28:16,876 --> 00:28:18,576 Speaker 2: bad stuff in. So I'm just going to go ahead 593 00:28:18,596 --> 00:28:20,016 Speaker 2: and say all this big stuff in the middle is 594 00:28:20,036 --> 00:28:22,236 Speaker 2: what I can produce. So the fear I have is 595 00:28:22,256 --> 00:28:25,376 Speaker 2: that we've kind of walked down this path not really 596 00:28:25,476 --> 00:28:28,896 Speaker 2: understanding that this is a feature of the mathematics. It's 597 00:28:28,916 --> 00:28:31,886 Speaker 2: not like someone at OpenAI is making the decision. It's 598 00:28:31,906 --> 00:28:34,446 Speaker 2: that they are using math that is forcing this decision. 599 00:28:34,866 --> 00:28:36,546 Speaker 2: And then it's in 10 or 20 years and we're 600 00:28:36,626 --> 00:28:40,046 Speaker 2: basically generating the same kind of content because nothing can 601 00:28:40,086 --> 00:28:43,086 Speaker 2: really like pop out without getting hammered down by either 602 00:28:43,226 --> 00:28:45,676 Speaker 2: algorithms or or by the generative models themselves. 603 00:28:45,876 --> 00:28:50,536 Speaker 1: Well, if that's true, there's a weird, happy, a silver 604 00:28:50,596 --> 00:28:53,576 Speaker 1: lining to that story, which is that's an opportunity for 605 00:28:53,646 --> 00:28:57,486 Speaker 1: talented humans, right? If it's true that AI can only 606 00:28:57,546 --> 00:29:03,526 Speaker 1: generate kind of mid-creative work... That's good for people who 607 00:29:03,626 --> 00:29:07,226 Speaker 1: are weirdos and make things that are maybe really bad 608 00:29:07,286 --> 00:29:09,506 Speaker 1: or maybe really great or a few people love even 609 00:29:09,526 --> 00:29:12,686 Speaker 1: though most people don't. If that's the case, that's actually 610 00:29:12,726 --> 00:29:14,496 Speaker 1: a happier outcome than I'm worried about. 611 00:29:14,656 --> 00:29:17,056 Speaker 2: And I have a 15-year-old, a 13-year-old, and a 9-year-old, 612 00:29:17,076 --> 00:29:19,556 Speaker 2: and that's what I tell them, is that the world 613 00:29:19,596 --> 00:29:22,496 Speaker 2: is going to get really average quickly, which is going 614 00:29:22,516 --> 00:29:25,156 Speaker 2: to be good for you because you're all very weird. 615 00:29:25,916 --> 00:29:27,836 Speaker 2: You kind of have to figure out what your voice is. 616 00:29:28,356 --> 00:29:30,216 Speaker 2: And in a world that wants to collapse you down, 617 00:29:30,556 --> 00:29:32,676 Speaker 2: that wants to hammer the nail that sticks out, just 618 00:29:32,716 --> 00:29:34,816 Speaker 2: know that your competitive advantage is that you won't. 619 00:29:35,756 --> 00:29:39,206 Speaker 1: I mean, I feel like there's a universe where you 620 00:29:39,266 --> 00:29:42,446 Speaker 1: can play with the AI and get it to make 621 00:29:42,486 --> 00:29:45,286 Speaker 1: stuff out in the tails, right? You can use AI 622 00:29:45,366 --> 00:29:49,686 Speaker 1: to make weirder stuff by kind of iterating, right? And 623 00:29:51,306 --> 00:29:54,156 Speaker 1: to what extent you need a person doing that, you know, 624 00:29:54,256 --> 00:29:56,296 Speaker 1: like to what extent it's just one shot where it's 625 00:29:56,336 --> 00:29:59,696 Speaker 1: like make something really weird and different in this way 626 00:29:59,716 --> 00:30:02,436 Speaker 1: versus to what extent it's like, well, I write something 627 00:30:02,456 --> 00:30:03,876 Speaker 1: and then I ask the AI about it and it 628 00:30:03,896 --> 00:30:06,076 Speaker 1: gives me some ideas and I write some more. is 629 00:30:06,276 --> 00:30:08,806 Speaker 1: to me kind of the open question and a meaningful 630 00:30:08,866 --> 00:30:12,846 Speaker 1: open question because the less you need of me, maybe 631 00:30:12,906 --> 00:30:13,926 Speaker 1: the worse at some level. 632 00:30:14,506 --> 00:30:16,886 Speaker 2: So this is where I get optimistic about AI is 633 00:30:16,906 --> 00:30:19,606 Speaker 2: because I do think that humans can push AI to 634 00:30:19,646 --> 00:30:22,546 Speaker 2: make something extraordinary, but it's the difference between being the 635 00:30:22,586 --> 00:30:25,756 Speaker 2: painter or the paintbrush, right? Like if AI is the paintbrush, 636 00:30:25,996 --> 00:30:28,256 Speaker 2: you're good, right? But as long as you, as soon 637 00:30:28,276 --> 00:30:30,516 Speaker 2: as you say, well, actually AI write a script for 638 00:30:30,556 --> 00:30:33,766 Speaker 2: me or paint the painting for me, Um, and it 639 00:30:33,806 --> 00:30:36,186 Speaker 2: becomes more of the painter. You kind of are screwed. 640 00:30:36,206 --> 00:30:37,546 Speaker 2: You're in that spot of like, it's going to make 641 00:30:37,566 --> 00:30:39,866 Speaker 2: something that's average, but you're right. If you're a human 642 00:30:39,886 --> 00:30:42,006 Speaker 2: that's involved in the process and you kind of push 643 00:30:42,066 --> 00:30:44,966 Speaker 2: it beyond its boundaries, you are then taking it out 644 00:30:45,006 --> 00:30:47,476 Speaker 2: of like the architect role and more into the, like 645 00:30:47,606 --> 00:30:49,676 Speaker 2: the engineer, like just do this thing. I know it 646 00:30:49,696 --> 00:30:52,116 Speaker 2: looks weird or sounds weird, but do it anyway. Um, 647 00:30:52,136 --> 00:30:55,656 Speaker 2: that's like taste, right? That that's human creativity getting imbued 648 00:30:55,696 --> 00:30:58,336 Speaker 2: into the AI product are very bullish on that. I 649 00:30:58,356 --> 00:31:01,796 Speaker 2: think that that's where AI will actually help our species. 650 00:31:02,656 --> 00:31:04,656 Speaker 2: I just don't like the idea that some people will say, well, 651 00:31:04,676 --> 00:31:07,716 Speaker 2: they'll be writing movies soon. Like, hopefully not. Or they'll 652 00:31:07,736 --> 00:31:10,436 Speaker 2: be very mid. Or they'll be writing music soon. Hopefully 653 00:31:10,456 --> 00:31:12,676 Speaker 2: a human that wants to make a movie or wants 654 00:31:12,716 --> 00:31:15,016 Speaker 2: to make a song says, like, there's something that is 655 00:31:15,076 --> 00:31:17,226 Speaker 2: so out of this world I have to use AI 656 00:31:17,266 --> 00:31:20,066 Speaker 2: for that clip of it. That's cool. I'm all for that. 657 00:31:20,506 --> 00:31:22,146 Speaker 1: Yeah, I mean, I hope so. You used the word 658 00:31:22,206 --> 00:31:24,566 Speaker 1: taste in there, which is a word people have been 659 00:31:24,646 --> 00:31:26,666 Speaker 1: using a lot, right? People have taste. I don't know. 660 00:31:29,246 --> 00:31:34,356 Speaker 1: That's like an interesting big... conversation about the future of 661 00:31:34,396 --> 00:31:37,186 Speaker 1: sort of human creativity and AI. If we think more 662 00:31:37,276 --> 00:31:41,766 Speaker 1: narrowly about the business that you are in, you know, 663 00:31:41,846 --> 00:31:48,066 Speaker 1: people's own likeness in this world of exploding AI, what 664 00:31:48,086 --> 00:31:50,906 Speaker 1: do you worry about in the, you know, next few years? 665 00:31:50,926 --> 00:31:53,046 Speaker 2: I would say the problems I want to solve for 666 00:31:53,346 --> 00:31:57,446 Speaker 2: are making sure that there's very low friction so people 667 00:31:57,486 --> 00:31:59,846 Speaker 2: can control their likeness, whether you are a celebrity or 668 00:31:59,886 --> 00:32:02,006 Speaker 2: you're like me and you're just some average guy walking 669 00:32:02,046 --> 00:32:05,156 Speaker 2: around I think that that's a difficult problem to solve. Like, 670 00:32:05,196 --> 00:32:06,856 Speaker 2: how do you do that at scale? Which friend of 671 00:32:06,896 --> 00:32:08,636 Speaker 2: mine is allowed to generate me as well as a 672 00:32:08,696 --> 00:32:09,396 Speaker 2: weird problem? 673 00:32:09,956 --> 00:32:11,076 Speaker 1: I'm going to just say none. 674 00:32:11,476 --> 00:32:14,396 Speaker 2: Yes. You might say none or, or what you might 675 00:32:14,496 --> 00:32:17,586 Speaker 2: actually find out is, um, you have a parent that 676 00:32:17,666 --> 00:32:21,536 Speaker 2: is ill and they're like, actually, um, I do want 677 00:32:21,556 --> 00:32:23,876 Speaker 2: to make sure that even after I'm gone, I'm reading 678 00:32:23,916 --> 00:32:27,036 Speaker 2: books to your kids. Right. And then you're like, okay, 679 00:32:27,076 --> 00:32:27,816 Speaker 2: that's an exception. 680 00:32:27,856 --> 00:32:28,016 Speaker 1: Right. 681 00:32:28,236 --> 00:32:29,936 Speaker 2: And then you'll end up finding these kinds of outliers 682 00:32:29,956 --> 00:32:31,776 Speaker 2: where you're like, can my partner make something? They had 683 00:32:31,816 --> 00:32:33,036 Speaker 2: a dream of me and they want to like show 684 00:32:33,056 --> 00:32:35,406 Speaker 2: me the exact dream they had of me. And it's 685 00:32:35,426 --> 00:32:37,566 Speaker 2: going to be cathartic for them. Yeah, sure. I don't 686 00:32:37,606 --> 00:32:39,406 Speaker 2: mind if my partner, I trust them with everything. Might 687 00:32:39,426 --> 00:32:41,026 Speaker 2: as well trust them with that. I think there'll be 688 00:32:41,066 --> 00:32:43,146 Speaker 2: slivers of things. You're going to have like high school 689 00:32:43,166 --> 00:32:45,646 Speaker 2: kids that'll be able to give their likeness because their 690 00:32:45,686 --> 00:32:47,846 Speaker 2: friends always love to create crazy stuff. 691 00:32:47,906 --> 00:32:48,086 Speaker 1: Right. 692 00:32:48,786 --> 00:32:51,796 Speaker 2: Um, And you need good guardrails around that too. And 693 00:32:51,836 --> 00:32:54,496 Speaker 2: so I think there'll be pockets of things that people 694 00:32:54,536 --> 00:32:56,926 Speaker 2: that even traditionally would be staunch, be like, I'm not 695 00:32:56,966 --> 00:32:59,066 Speaker 2: going to do anything with AI, will say, well, actually, 696 00:33:00,706 --> 00:33:02,766 Speaker 2: I think that that's cool. And then it'll just grow 697 00:33:02,786 --> 00:33:03,086 Speaker 2: from there. 698 00:33:05,466 --> 00:33:08,606 Speaker 1: You do seem optimistic. So give me your happy vision 699 00:33:08,646 --> 00:33:10,066 Speaker 1: of the world five or 10 years from now. 700 00:33:10,666 --> 00:33:15,366 Speaker 2: The happy vision is one in which we don't ever 701 00:33:15,506 --> 00:33:18,636 Speaker 2: use AI as the painter. It's humans that have figured 702 00:33:18,676 --> 00:33:21,396 Speaker 2: out how to push the boundary of creativity with this 703 00:33:21,456 --> 00:33:23,886 Speaker 2: new kind of paintbrush. That's the first part. And then 704 00:33:23,906 --> 00:33:27,186 Speaker 2: the second part is people don't see the internet as 705 00:33:27,226 --> 00:33:29,866 Speaker 2: this big wild west of anything can get uploaded of 706 00:33:29,886 --> 00:33:32,966 Speaker 2: you whenever, wherever, and can destroy your life overnight. And 707 00:33:33,006 --> 00:33:35,106 Speaker 2: so I think both of these things, like the privacy 708 00:33:35,146 --> 00:33:38,336 Speaker 2: component and the control and autonomy and agency, hopefully will 709 00:33:38,376 --> 00:33:40,696 Speaker 2: come to the internet. And then the second part, of course, 710 00:33:40,736 --> 00:33:45,376 Speaker 2: being creativity, getting solved for, but solved in my mind 711 00:33:45,636 --> 00:33:47,116 Speaker 2: is more paintbrush and less painter. 712 00:33:50,996 --> 00:34:03,626 Speaker 1: We'll be back in a minute with the lightning round. Okay, 713 00:34:03,926 --> 00:34:06,276 Speaker 1: let's do a lightning round. What's your favorite deepfake? 714 00:34:10,596 --> 00:34:13,756 Speaker 2: The bunny rabbits on the trampoline. Totally harmless. It got 715 00:34:14,076 --> 00:34:16,446 Speaker 2: a million people laughing and enjoying and seeing that AI 716 00:34:16,466 --> 00:34:19,186 Speaker 2: could make some funny, cool stuff and tricked a lot 717 00:34:19,206 --> 00:34:21,686 Speaker 2: of people into thinking they were real as well. So 718 00:34:21,706 --> 00:34:22,626 Speaker 2: I actually did like that. 719 00:34:23,026 --> 00:34:24,376 Speaker 1: I'll have to check it out. I haven't seen it. 720 00:34:26,706 --> 00:34:30,946 Speaker 1: What celebrity likeness would you license to advertise for Lodi, 721 00:34:31,066 --> 00:34:32,286 Speaker 1: to be a spokesman for Lodi? 722 00:34:33,086 --> 00:34:36,586 Speaker 2: Cannot say. Yeah. It's coming out actually in the next 723 00:34:36,646 --> 00:34:37,986 Speaker 2: few weeks, so cannot say. 724 00:34:38,146 --> 00:34:38,746 Speaker 1: Okay. 725 00:34:38,966 --> 00:34:40,806 Speaker 2: Yeah. 726 00:34:41,546 --> 00:34:45,256 Speaker 1: I'm very interested to see. Is there one or will 727 00:34:45,296 --> 00:34:45,976 Speaker 1: there be a bunch? 728 00:34:46,536 --> 00:34:49,516 Speaker 2: There's one that's on the forefront of all of this. Okay. 729 00:34:50,756 --> 00:34:52,456 Speaker 2: Someone that we deeply respect in this space. 730 00:34:54,036 --> 00:34:57,036 Speaker 1: I'm very interested. How do you want your kids to 731 00:34:57,216 --> 00:34:57,816 Speaker 1: use AI? 732 00:35:00,186 --> 00:35:02,886 Speaker 2: I told my 15-year-old that he could probably cheat on 733 00:35:02,966 --> 00:35:05,866 Speaker 2: everything with AI. And that there's probably no way that 734 00:35:05,966 --> 00:35:08,226 Speaker 2: me or a teacher could figure that out. Because I 735 00:35:08,246 --> 00:35:10,446 Speaker 2: know that kids are really crafty. So that he has 736 00:35:10,486 --> 00:35:12,186 Speaker 2: to make the active choice that if he wants to 737 00:35:12,226 --> 00:35:14,856 Speaker 2: be successful, he has to do hard work. And that's 738 00:35:14,916 --> 00:35:19,936 Speaker 2: a completely different skill set. So it really is coming 739 00:35:19,956 --> 00:35:21,796 Speaker 2: down to just, especially if you have a smart kid, 740 00:35:21,856 --> 00:35:23,356 Speaker 2: just tell them, look, I know you can do it. 741 00:35:23,836 --> 00:35:25,356 Speaker 2: But here, let me make the case for why you 742 00:35:25,376 --> 00:35:27,816 Speaker 2: should work harder instead. Is if it's in your head, 743 00:35:28,316 --> 00:35:30,206 Speaker 2: you could then leverage it to do things that other 744 00:35:30,226 --> 00:35:33,936 Speaker 2: people cannot. So it's a little competitive, but It's essentially 745 00:35:33,976 --> 00:35:36,316 Speaker 2: to tell my kids, don't shy away from it. Like 746 00:35:36,356 --> 00:35:39,476 Speaker 2: my 15-year-old's on the $ 200 a month cloud plan. Like 747 00:35:39,496 --> 00:35:42,336 Speaker 2: he uses AI a ton, right? He's like building apps 748 00:35:42,376 --> 00:35:45,006 Speaker 2: and stuff. It's crazy. I'm really proud of him. But 749 00:35:45,566 --> 00:35:47,606 Speaker 2: at the same time, he knows that you can't use 750 00:35:47,666 --> 00:35:49,786 Speaker 2: it to fill out all your homework without you robbing 751 00:35:49,826 --> 00:35:50,666 Speaker 2: yourself of something. 752 00:35:51,406 --> 00:35:57,046 Speaker 1: Yes, right. The point of homework is actually not the whatever, 753 00:35:57,086 --> 00:35:59,426 Speaker 1: the math you're learning. It's learning to just sit there 754 00:35:59,646 --> 00:36:02,066 Speaker 1: and do it, right? That's actually the work. 755 00:36:02,166 --> 00:36:03,896 Speaker 2: Yeah, those are the... So a lot of our discussions 756 00:36:03,916 --> 00:36:06,196 Speaker 2: around AI are actually the reasons why you wouldn't want 757 00:36:06,216 --> 00:36:09,126 Speaker 2: to use it in a home that deeply uses it. Yeah. 758 00:36:09,246 --> 00:36:12,186 Speaker 2: So it's, but I think that that's like the, the 759 00:36:12,226 --> 00:36:14,146 Speaker 2: brake and gas pedal kind of thing. Like you need 760 00:36:14,186 --> 00:36:15,026 Speaker 2: both to run the car. 761 00:36:15,386 --> 00:36:15,526 Speaker 1: Yeah. 762 00:36:15,546 --> 00:36:17,726 Speaker 2: And so a lot of the conversation is around the 763 00:36:17,766 --> 00:36:21,046 Speaker 2: more complicated use of a brake system. What? 764 00:36:21,166 --> 00:36:23,606 Speaker 1: I was trying to think, frankly, of a good question 765 00:36:23,666 --> 00:36:26,926 Speaker 1: about like AI and the future of movie stars. Like, 766 00:36:27,066 --> 00:36:28,966 Speaker 1: I don't know the first movie with an AI genre, 767 00:36:28,986 --> 00:36:33,316 Speaker 1: but I couldn't quite get the right question. Like, make 768 00:36:33,336 --> 00:36:36,056 Speaker 1: a prediction about AI and the future of movie stars. 769 00:36:37,136 --> 00:36:40,876 Speaker 2: I think that there's going to be, there's going to 770 00:36:40,896 --> 00:36:44,546 Speaker 2: be a demand for like authentic performances. That's what they'll 771 00:36:44,566 --> 00:36:47,006 Speaker 2: be called. Something to that effect, some euphemism for it's 772 00:36:47,046 --> 00:36:49,586 Speaker 2: really them in front of film. And I think they're 773 00:36:49,606 --> 00:36:52,306 Speaker 2: going to make movies that portray that a little bit better. 774 00:36:52,446 --> 00:36:54,466 Speaker 2: Like it'll be like human expressions that AI is not 775 00:36:54,526 --> 00:36:57,456 Speaker 2: able to get or more complicated scripts that land in 776 00:36:57,496 --> 00:37:00,936 Speaker 2: a way that's not like uncanny valley. This is a 777 00:37:00,956 --> 00:37:05,656 Speaker 2: good example. There's the, The Silver Springs, Fleetwood Mac, where 778 00:37:05,856 --> 00:37:07,876 Speaker 2: the lead singer and the guitarist basically are in this 779 00:37:07,936 --> 00:37:12,326 Speaker 2: long-term situationship, and they're essentially singing about themselves on stage, 780 00:37:12,366 --> 00:37:14,026 Speaker 2: and everyone can kind of feel that tension. It's a 781 00:37:14,046 --> 00:37:15,986 Speaker 2: human thing. And if you were to swap them out 782 00:37:16,026 --> 00:37:19,546 Speaker 2: for robots, you probably would not feel the same way 783 00:37:19,586 --> 00:37:22,726 Speaker 2: as a human. I think there's something about knowing that 784 00:37:22,746 --> 00:37:25,666 Speaker 2: that person suffered in the way that you've suffered that 785 00:37:25,746 --> 00:37:27,946 Speaker 2: makes you identify with their art that is going to 786 00:37:27,966 --> 00:37:30,546 Speaker 2: be much bigger, and we talk about Taste, I know, 787 00:37:30,606 --> 00:37:32,876 Speaker 2: is a cheesy word, but still is something real here. 788 00:37:32,936 --> 00:37:35,656 Speaker 2: I think taste for that will go up because it'll 789 00:37:35,676 --> 00:37:37,476 Speaker 2: be more rare. And then I think there's going to 790 00:37:37,496 --> 00:37:40,796 Speaker 2: be a bunch of superhero action movies that are made 791 00:37:40,836 --> 00:37:41,896 Speaker 2: in AI and no one's going to care. 792 00:37:42,896 --> 00:37:46,836 Speaker 1: I mean, superhero action movies are already largely made in AI. 793 00:37:46,856 --> 00:37:48,876 Speaker 2: Correct. That's it. Yes. Yes. 794 00:37:49,356 --> 00:37:55,426 Speaker 1: Like, it is interestingly not as binary as as we say, right? 795 00:37:55,906 --> 00:38:00,166 Speaker 2: If your art is making really generic romance songs for radio, 796 00:38:00,406 --> 00:38:03,596 Speaker 2: you probably are in for some competition with AI. But 797 00:38:03,656 --> 00:38:07,016 Speaker 2: if you have a unique sound, unique bass, whatever it 798 00:38:07,096 --> 00:38:09,756 Speaker 2: is that you bring to the table that is truly 799 00:38:09,816 --> 00:38:12,616 Speaker 2: not like something else we hear everywhere else, the demand 800 00:38:12,636 --> 00:38:13,716 Speaker 2: for that's going to go through the roof. 801 00:38:14,696 --> 00:38:17,386 Speaker 1: That's encouraging. I appreciate your time. 802 00:38:18,046 --> 00:38:18,426 Speaker 2: Of course. 803 00:38:25,836 --> 00:38:30,056 Speaker 1: Luke Garagoni is co-founder and CEO of Lodi. Our show 804 00:38:30,116 --> 00:38:33,556 Speaker 1: is produced by Gabriel Hunter Chang and Trina Menino. Our 805 00:38:33,696 --> 00:38:36,916 Speaker 1: editor is Lydia Jean Cott, and our engineer is Sarah Bruguier. 806 00:38:37,456 --> 00:38:39,726 Speaker 1: I'm Jacob Goldstein, and we'll be back next week with 807 00:38:39,806 --> 00:38:41,006 Speaker 1: another episode of What's Your Problem?