WEBVTT - Can We Defeat Deepfakes?

0:00:15.396 --> 0:00:32.836
<v Speaker 1>Pushkin. I'm Jacob Goldstein, and this is What's Your Problem?

0:00:33.016 --> 0:00:36.236
<v Speaker 1>My guest today is Luke Aragoni. He is the co-founder

0:00:36.256 --> 0:00:39.506
<v Speaker 1>and CEO of a company called Lodi. And Lodi is

0:00:39.526 --> 0:00:42.846
<v Speaker 1>an acronym, and it stands for Likeness on the Internet.

0:00:43.456 --> 0:00:46.846
<v Speaker 1>Tens of thousands of famous people pay Lodi to scour

0:00:46.886 --> 0:00:51.726
<v Speaker 1>the internet and take down deepfakes and unapproved endorsements. If

0:00:51.746 --> 0:00:53.726
<v Speaker 1>you've heard of a person, if a person is a

0:00:53.766 --> 0:00:56.226
<v Speaker 1>little bit famous, there is a very good chance they

0:00:56.266 --> 0:00:59.376
<v Speaker 1>are one of Lodi's clients. But in Luke's telling, this

0:00:59.516 --> 0:01:02.736
<v Speaker 1>is just the beginning. The company will soon start offering

0:01:02.796 --> 0:01:06.456
<v Speaker 1>a service for celebrities to license the use of their image.

0:01:06.976 --> 0:01:09.716
<v Speaker 1>In other words, ordinary people will be able to pay

0:01:09.736 --> 0:01:12.666
<v Speaker 1>Lodi to create stuff with AI using the faces and

0:01:12.736 --> 0:01:16.946
<v Speaker 1>voices of real celebrities. Lodi also plans to start offering

0:01:16.986 --> 0:01:20.426
<v Speaker 1>its services to ordinary people. The plan is to include

0:01:20.446 --> 0:01:23.406
<v Speaker 1>a free tier for people who are not famous, but

0:01:23.446 --> 0:01:26.096
<v Speaker 1>who are worried about what's happening with their likeness on

0:01:26.136 --> 0:01:29.896
<v Speaker 1>the internet. At a macro level, the biggest level, Luke's

0:01:29.936 --> 0:01:33.656
<v Speaker 1>problem is this. How do you turn the internet into

0:01:33.716 --> 0:01:37.046
<v Speaker 1>a place where people have more control over their name, image,

0:01:37.186 --> 0:01:40.936
<v Speaker 1>and likeness? In our conversation, Luke and I talked about

0:01:41.056 --> 0:01:45.376
<v Speaker 1>the technology behind Lodi. We talked about federal legislation that

0:01:45.436 --> 0:01:48.746
<v Speaker 1>may soon make it easier for people to fight deep fakes.

0:01:49.426 --> 0:01:52.146
<v Speaker 1>And we also talk about AI and the future of creativity.

0:01:52.846 --> 0:01:54.666
<v Speaker 1>To start, I asked Luke to give me the state

0:01:54.686 --> 0:01:58.476
<v Speaker 1>of play for what's happening right now with unauthorized use

0:01:58.536 --> 0:02:00.556
<v Speaker 1>of people's likeness on the internet.

0:02:01.416 --> 0:02:03.636
<v Speaker 2>So you have kind of two different categories of people.

0:02:03.796 --> 0:02:06.776
<v Speaker 2>If you're really famous and your likeness is going to

0:02:06.816 --> 0:02:10.126
<v Speaker 2>get used to sell random things. It could be to

0:02:10.146 --> 0:02:13.086
<v Speaker 2>sell a scam. It's like an impersonation, a fake Instagram

0:02:13.126 --> 0:02:15.786
<v Speaker 2>of a celebrity, and they're selling you anything from VIP

0:02:15.906 --> 0:02:19.106
<v Speaker 2>access to the next tour to a romance scam that

0:02:19.126 --> 0:02:21.766
<v Speaker 2>they're falling in love with you, right? And all of

0:02:21.806 --> 0:02:25.506
<v Speaker 2>them have some kind of economic extraction from it. Or

0:02:25.566 --> 0:02:28.086
<v Speaker 2>it's kind of more straightforward. It's like, hey, I've come

0:02:28.126 --> 0:02:30.516
<v Speaker 2>up with a brand new shampoo, and this famous celebrity

0:02:30.596 --> 0:02:33.596
<v Speaker 2>uses the shampoo. Here's a deepfake video talking about how

0:02:33.636 --> 0:02:37.556
<v Speaker 2>they love it. So that's probably the most common way

0:02:37.596 --> 0:02:39.356
<v Speaker 2>in which it's used. And then you have a kind

0:02:39.376 --> 0:02:44.676
<v Speaker 2>of more incendiary thing where everyone else, like me, for example,

0:02:44.956 --> 0:02:47.036
<v Speaker 2>what we have to be concerned about is someone deepfaking

0:02:47.116 --> 0:02:51.286
<v Speaker 2>us for any kind of authorization reason. And then there's

0:02:51.306 --> 0:02:53.526
<v Speaker 2>an even darker way in which it's used that we

0:02:53.586 --> 0:02:56.686
<v Speaker 2>saw at some high schools this year, which is a

0:02:56.726 --> 0:02:59.286
<v Speaker 2>lot of folks are getting deepfaked to make images of

0:02:59.326 --> 0:03:02.906
<v Speaker 2>them that were never taken, but have a very harmful

0:03:02.986 --> 0:03:04.426
<v Speaker 2>nature to them and their reputation.

0:03:04.996 --> 0:03:08.416
<v Speaker 1>Do you mean revenge porn? You're talking around revenge porn there.

0:03:08.436 --> 0:03:12.236
<v Speaker 2>Yes. We like to call it non-consensual intimate images, which

0:03:12.316 --> 0:03:14.396
<v Speaker 2>is a lot better of a term than revenge porn.

0:03:14.456 --> 0:03:17.256
<v Speaker 2>But colloquially, we all call it revenge porn, and so

0:03:17.276 --> 0:03:19.176
<v Speaker 2>we even have to kind of mix the terms together.

0:03:19.196 --> 0:03:21.196
<v Speaker 2>But just to kind of re-educate a lot of people,

0:03:21.716 --> 0:03:23.456
<v Speaker 2>NCII is a much better term for it.

0:03:23.876 --> 0:03:28.686
<v Speaker 1>Okay, fair enough. So this is a potential problem for everybody.

0:03:29.426 --> 0:03:33.466
<v Speaker 1>What's like... What's the scope of it right now? Like,

0:03:33.506 --> 0:03:37.426
<v Speaker 1>if I just pick some random A-list celebrity, will there, like, 100%

0:03:37.426 --> 0:03:42.266
<v Speaker 1>be deepfakes of them scamming on the internet right now?

0:03:42.726 --> 0:03:48.756
<v Speaker 2>Yeah, we track about 30,000 or so celebrity accounts, people, personas.

0:03:49.336 --> 0:03:51.416
<v Speaker 1>And those are your clients, or that's just you kind

0:03:51.436 --> 0:03:52.656
<v Speaker 1>of understanding the landscape?

0:03:53.916 --> 0:03:55.966
<v Speaker 2>Two-thirds, a little bit more, a little over 20,000 are clients.

0:03:57.136 --> 0:04:00.216
<v Speaker 1>So that's way below the A-list. That's way more famous

0:04:00.236 --> 0:04:02.066
<v Speaker 1>people than anybody could name. Oh, yeah.

0:04:02.126 --> 0:04:04.646
<v Speaker 2>No, it is. It's pretty much every famous person that

0:04:04.686 --> 0:04:06.766
<v Speaker 2>you can think of, even if you don't remember their name.

0:04:07.066 --> 0:04:09.006
<v Speaker 2>And so, yes, for those folks, you're going to get

0:04:09.146 --> 0:04:13.026
<v Speaker 2>nearly 100% coverage on impersonations and other scams.

0:04:13.626 --> 0:04:16.436
<v Speaker 1>When you say nearly 100% coverage, you mean for all

0:04:16.476 --> 0:04:18.356
<v Speaker 1>of those people, there are going to be deep fakes.

0:04:18.176 --> 0:04:21.696
<v Speaker 2>Of them on the Internet? Yes, and impersonations. So it's

0:04:21.716 --> 0:04:23.956
<v Speaker 2>like a real picture of them on an Instagram account,

0:04:24.236 --> 0:04:25.256
<v Speaker 2>but not their real account.

0:04:25.756 --> 0:04:28.156
<v Speaker 1>Right, so that's not a deepfake in the AI sense.

0:04:28.316 --> 0:04:30.656
<v Speaker 1>You can just take the picture of the person and

0:04:30.696 --> 0:04:33.236
<v Speaker 1>be like, look, this person loves our soap, and that's

0:04:33.556 --> 0:04:34.596
<v Speaker 1>a scam and a lie.

0:04:34.916 --> 0:04:35.336
<v Speaker 2>Absolutely.

0:04:35.376 --> 0:04:38.026
<v Speaker 1>And just to be clear, like, that is illegal because

0:04:38.066 --> 0:04:40.826
<v Speaker 1>people have a right to their name, image, and likeness? Like,

0:04:40.846 --> 0:04:44.666
<v Speaker 1>that's the fundamental business and kind of intellectual property thing

0:04:44.706 --> 0:04:45.526
<v Speaker 1>that we're talking about?

0:04:45.926 --> 0:04:48.426
<v Speaker 2>Yeah, it's called a right of publicity. And ironically, it

0:04:48.466 --> 0:04:51.936
<v Speaker 2>became a bigger conversation when the camera came around. Uh-huh.

0:04:52.296 --> 0:04:54.756
<v Speaker 2>And so, you know, 100 plus years ago when people

0:04:54.796 --> 0:04:57.196
<v Speaker 2>started taking pictures of anyone, someone was like, well, I'll

0:04:57.226 --> 0:04:59.206
<v Speaker 2>put this on the front of my cereal box. This

0:04:59.246 --> 0:05:01.046
<v Speaker 2>is just like an average person. They're like, no, that

0:05:01.086 --> 0:05:04.166
<v Speaker 2>person has, all right, you can't just make, it really

0:05:04.186 --> 0:05:06.686
<v Speaker 2>is the commercial component. You can't make money off their

0:05:06.726 --> 0:05:09.786
<v Speaker 2>likeness without their approval or their endorsement.

0:05:11.426 --> 0:05:15.686
<v Speaker 1>Okay, so that's the lay of the land. Your company

0:05:15.706 --> 0:05:19.326
<v Speaker 1>sort of walks into this universe as it is growing

0:05:19.356 --> 0:05:23.186
<v Speaker 1>a few years ago. Like, Basically, what's your company do?

0:05:23.486 --> 0:05:25.226
<v Speaker 2>What we do is we download the internet every day,

0:05:25.946 --> 0:05:28.566
<v Speaker 2>and then we run face recognition over it looking for

0:05:28.606 --> 0:05:30.886
<v Speaker 2>our clients. And when we find them in a spot

0:05:30.906 --> 0:05:34.306
<v Speaker 2>that they don't belong, whether it be a scam, impersonation, NCII,

0:05:34.526 --> 0:05:37.246
<v Speaker 2>or any other type of thing, we figure out how

0:05:37.286 --> 0:05:40.506
<v Speaker 2>we get this thing removed. So the removal part is

0:05:40.856 --> 0:05:43.676
<v Speaker 2>another cool piece of tech because it's fully automated. But

0:05:43.696 --> 0:05:46.576
<v Speaker 2>we can reach out to adult sites and tell them

0:05:46.636 --> 0:05:50.316
<v Speaker 2>they are creating liability for their merchant accounts or, you know,

0:05:50.336 --> 0:05:53.166
<v Speaker 2>they're domain provider, or a number of other ways. And

0:05:53.186 --> 0:05:55.326
<v Speaker 2>then on social media, we kind of have this great

0:05:55.486 --> 0:05:59.946
<v Speaker 2>access and great partnership across everyone but X. And they

0:05:59.986 --> 0:06:02.096
<v Speaker 2>just give us a kind of API access to issue

0:06:02.126 --> 0:06:02.766
<v Speaker 2>those takedowns.

0:06:03.836 --> 0:06:07.676
<v Speaker 1>And so does that mean that it's fully automated, like

0:06:07.696 --> 0:06:15.036
<v Speaker 1>your AI, essentially your facial recognition software, spots these unauthorized

0:06:15.136 --> 0:06:19.726
<v Speaker 1>uses of people's likeness and then pings the software or

0:06:19.746 --> 0:06:23.526
<v Speaker 1>the AI of whatever, YouTube or TikTok, and says, hey,

0:06:23.586 --> 0:06:25.926
<v Speaker 1>take this down, and then the machine just takes it down?

0:06:25.986 --> 0:06:27.886
<v Speaker 1>Like there's no human needs to.

0:06:28.026 --> 0:06:29.246
<v Speaker 2>There's no human needed for this.

0:06:29.766 --> 0:06:34.066
<v Speaker 1>And presumably that's because the volume is too high. Like

0:06:34.666 --> 0:06:37.796
<v Speaker 1>what order of magnitude, how many, you know, how many

0:06:37.816 --> 0:06:40.856
<v Speaker 1>things has your company removed, whatever, today or this week?

0:06:42.096 --> 0:06:44.696
<v Speaker 2>We do tens of thousands of things a day.

0:06:45.016 --> 0:06:46.416
<v Speaker 1>Tens of thousands a day.

0:06:46.816 --> 0:06:47.816
<v Speaker 2>Yeah, it's pretty insane.

0:06:48.266 --> 0:06:52.626
<v Speaker 1>That is insane. So, like, at every moment, at every second,

0:06:52.726 --> 0:06:56.646
<v Speaker 1>someone is posting a scam likeness. I guess presumably at

0:06:56.666 --> 0:06:59.766
<v Speaker 1>this point it's machines. Like, you don't need a person. Like,

0:06:59.826 --> 0:07:03.246
<v Speaker 1>I could just build a bot to create a million

0:07:03.346 --> 0:07:04.286
<v Speaker 1>scam likenesses.

0:07:04.626 --> 0:07:07.426
<v Speaker 2>Correct. We called it an industrial game of whack-a-mole.

0:07:07.646 --> 0:07:12.186
<v Speaker 1>Yeah. Like, what's the frontier? What are you struggling with

0:07:12.226 --> 0:07:12.646
<v Speaker 1>right now?

0:07:12.666 --> 0:07:16.136
<v Speaker 2>Hmm. I think a lot of people don't understand that

0:07:16.296 --> 0:07:19.346
<v Speaker 2>you can have delete for the internet. And so we

0:07:19.386 --> 0:07:23.846
<v Speaker 2>think about how we position the company into a greater

0:07:23.906 --> 0:07:26.846
<v Speaker 2>context around like, okay, there's a thing of me online.

0:07:27.026 --> 0:07:28.606
<v Speaker 2>It's going to be online forever. And we kind of

0:07:28.626 --> 0:07:30.966
<v Speaker 2>have to interject ourselves and say, well, this won't be

0:07:30.986 --> 0:07:33.686
<v Speaker 2>online forever. I know everyone was told that for 30 years,

0:07:34.216 --> 0:07:36.056
<v Speaker 2>but now there's a really efficient way for us to

0:07:36.096 --> 0:07:36.616
<v Speaker 2>manage that.

0:07:37.296 --> 0:07:38.176
<v Speaker 1>What's the business model?

0:07:38.656 --> 0:07:41.796
<v Speaker 2>So for celebrities, we charge a monthly fee to go

0:07:41.856 --> 0:07:45.766
<v Speaker 2>scraping and removing this content. And for consumers, it'll be

0:07:45.806 --> 0:07:49.446
<v Speaker 2>more freemium. They can find everything of themselves online, and

0:07:49.466 --> 0:07:51.286
<v Speaker 2>then they get a certain number of takedowns if they

0:07:51.326 --> 0:07:53.606
<v Speaker 2>upgrade to Pro. Or we actually teach them how they

0:07:53.646 --> 0:07:56.886
<v Speaker 2>can go do the takedowns themselves, so we're not trying

0:07:56.926 --> 0:07:58.366
<v Speaker 2>to gatekeep their privacy.

0:07:58.826 --> 0:08:00.446
<v Speaker 1>When's it going to be available for consumers?

0:08:00.866 --> 0:08:01.806
<v Speaker 2>Probably early October.

0:08:02.246 --> 0:08:04.406
<v Speaker 1>Technically, how is it going?

0:08:05.826 --> 0:08:10.976
<v Speaker 2>Oh, it's both a technical nightmare and a marvel. It's insane.

0:08:12.236 --> 0:08:15.376
<v Speaker 2>because the problem set of downloading like 600 million assets

0:08:15.436 --> 0:08:18.836
<v Speaker 2>a day, you have database problems and disk problems.

0:08:18.896 --> 0:08:20.536
<v Speaker 1>Why do you need to download it?

0:08:21.176 --> 0:08:23.946
<v Speaker 2>Because that's how you run the face recognition on it. Okay.

0:08:24.146 --> 0:08:25.146
<v Speaker 1>Interesting. Yeah.

0:08:25.466 --> 0:08:28.206
<v Speaker 2>And so you download it, you run your face recognition

0:08:28.266 --> 0:08:31.546
<v Speaker 2>on it, and then that gives you the capability of saying,

0:08:31.586 --> 0:08:33.846
<v Speaker 2>this person's coming to me. Can you find everything in

0:08:33.886 --> 0:08:35.866
<v Speaker 2>my online? We look into our database. We say, oh,

0:08:36.136 --> 0:08:38.476
<v Speaker 2>we can now identify these rows as you. The other

0:08:38.496 --> 0:08:39.256
<v Speaker 2>things we toss out.

0:08:39.496 --> 0:08:44.456
<v Speaker 1>Yeah. And are the Bad people doing things to make

0:08:44.496 --> 0:08:45.156
<v Speaker 1>it harder for you?

0:08:46.456 --> 0:08:48.736
<v Speaker 2>No, they're making it easier. They're trying to make their

0:08:48.976 --> 0:08:53.416
<v Speaker 2>AI-generated faces more like the person, which actually is an

0:08:53.476 --> 0:08:57.016
<v Speaker 2>alignment to us. It's like the more their deepfakes look realistic,

0:08:57.276 --> 0:08:58.816
<v Speaker 2>the easier it is for us to catch them.

0:08:59.196 --> 0:09:03.076
<v Speaker 1>Oh, that's interesting. And so presumably, I mean, the next

0:09:03.156 --> 0:09:06.756
<v Speaker 1>step is figuring out what are the legitimate uses, right? Correct.

0:09:06.776 --> 0:09:08.566
<v Speaker 1>In an automated way. How do you do that part?

0:09:09.086 --> 0:09:11.446
<v Speaker 2>So that's more of the context understanding of what the

0:09:11.506 --> 0:09:14.786
<v Speaker 2>image is. We actually do that discovery as well. We

0:09:14.826 --> 0:09:17.626
<v Speaker 2>have this thing that we call a semantic understanding of

0:09:17.686 --> 0:09:20.146
<v Speaker 2>every image and video where it comes through and we say,

0:09:20.226 --> 0:09:22.206
<v Speaker 2>what's going on in this? Is someone running in this park?

0:09:22.566 --> 0:09:25.626
<v Speaker 2>Are they walking in the park? Are they doing drugs?

0:09:25.746 --> 0:09:27.316
<v Speaker 2>Are they doing something they shouldn't be doing in this

0:09:27.366 --> 0:09:30.216
<v Speaker 2>picture or image? And then we have this massive collection of,

0:09:30.336 --> 0:09:33.936
<v Speaker 2>we know that it's this celebrity doing this thing. And

0:09:33.976 --> 0:09:38.316
<v Speaker 2>was there like an expectation of privacy here? Or More importantly,

0:09:38.416 --> 0:09:40.616
<v Speaker 2>is this authorized? And they can give us their rules

0:09:40.636 --> 0:09:42.196
<v Speaker 2>around what should or shouldn't be authorized.

0:09:43.436 --> 0:09:46.356
<v Speaker 1>And presumably the giving the rules of what should and

0:09:46.376 --> 0:09:49.536
<v Speaker 1>shouldn't be authorized also needs to be automated, right? Like

0:09:49.576 --> 0:09:52.656
<v Speaker 1>presumably the clients need to say, these are the things

0:09:52.736 --> 0:09:54.696
<v Speaker 1>I am endorsing. It's sort of a white list. Is

0:09:54.736 --> 0:09:55.576
<v Speaker 1>that the system?

0:09:55.836 --> 0:09:56.956
<v Speaker 2>Yes. Absolutely.

0:10:00.356 --> 0:10:02.756
<v Speaker 1>Are there any technical problems you're trying to solve or

0:10:02.796 --> 0:10:04.626
<v Speaker 1>is it technically a sort of solved issue for you?

0:10:04.906 --> 0:10:07.126
<v Speaker 2>Every couple of months we we trip on something that

0:10:07.226 --> 0:10:10.216
<v Speaker 2>is kind of insane. We're like, that's oh, I didn't

0:10:10.296 --> 0:10:11.916
<v Speaker 2>realize we'd even have to solve that problem.

0:10:12.276 --> 0:10:13.896
<v Speaker 1>What's an example? What's an example?

0:10:14.336 --> 0:10:16.356
<v Speaker 2>Well, when we went to process video, we found out

0:10:16.376 --> 0:10:18.296
<v Speaker 2>the reason why no one had done what we do

0:10:18.356 --> 0:10:21.136
<v Speaker 2>before us, like no one did this in 2005 was

0:10:21.176 --> 0:10:24.116
<v Speaker 2>because it's expensive to process video. So like to take

0:10:24.156 --> 0:10:26.256
<v Speaker 2>a video in and actually like look at all the frames.

0:10:27.176 --> 0:10:29.736
<v Speaker 2>even without great machine learning or AI at that time,

0:10:29.856 --> 0:10:32.536
<v Speaker 2>it was just really hard to do that. And so

0:10:32.576 --> 0:10:35.076
<v Speaker 2>we were broke and we couldn't afford to just spool

0:10:35.116 --> 0:10:37.336
<v Speaker 2>up more servers like some really large AI companies are

0:10:37.376 --> 0:10:40.376
<v Speaker 2>able to do. And so necessity was the mother of invention.

0:10:40.396 --> 0:10:43.756
<v Speaker 2>And we invented a method of processing videos that were

0:10:43.936 --> 0:10:46.756
<v Speaker 2>like a thousand times faster than anyone else. And we

0:10:46.776 --> 0:10:48.446
<v Speaker 2>just did this thinking like, well, this is what you

0:10:48.486 --> 0:10:50.376
<v Speaker 2>have to do, right? Like this is just whatever the

0:10:50.416 --> 0:10:53.196
<v Speaker 2>next step is. And it came to find out that we, uh,

0:10:53.576 --> 0:10:55.236
<v Speaker 2>We had to build something that no one else had

0:10:55.276 --> 0:10:57.916
<v Speaker 2>done that others had tried to do and failed. Now

0:10:57.976 --> 0:11:01.996
<v Speaker 2>it's like, it's one of the most guarded, protected algorithms

0:11:02.036 --> 0:11:04.636
<v Speaker 2>at our company because before it was just kind of like,

0:11:04.656 --> 0:11:05.536
<v Speaker 2>it was like floating everywhere.

0:11:05.596 --> 0:11:07.716
<v Speaker 1>Did you patent it or is it a trade secret?

0:11:07.736 --> 0:11:09.296
<v Speaker 2>Oh, interesting. It's a trade secret.

0:11:09.516 --> 0:11:11.396
<v Speaker 1>So you can do a thing and you're not telling

0:11:11.466 --> 0:11:12.436
<v Speaker 1>anybody how to do it.

0:11:12.916 --> 0:11:14.886
<v Speaker 2>Correct. And people know that we do it. They still

0:11:14.926 --> 0:11:15.566
<v Speaker 2>can't figure it out.

0:11:16.046 --> 0:11:16.846
<v Speaker 1>Oh, that's interesting.

0:11:16.866 --> 0:11:19.346
<v Speaker 2>We can prove it. And that's the only part that matters.

0:11:19.626 --> 0:11:22.566
<v Speaker 1>And what it allows you to do is to essentially...

0:11:22.976 --> 0:11:26.996
<v Speaker 1>have AI look at a video much more efficiently and

0:11:27.056 --> 0:11:30.656
<v Speaker 1>cheaply than anybody else can or than anybody else could before? Interesting.

0:11:30.676 --> 0:11:31.996
<v Speaker 2>That's right.

0:11:33.336 --> 0:11:35.876
<v Speaker 1>And presumably you chose not to patent it because when

0:11:35.916 --> 0:11:37.716
<v Speaker 1>you patent something, you have to tell the world how

0:11:37.756 --> 0:11:39.586
<v Speaker 1>you did it and then people can copy it.

0:11:39.816 --> 0:11:42.156
<v Speaker 2>And then they can tweak one part of the mathematics

0:11:42.396 --> 0:11:45.286
<v Speaker 2>and essentially have something like that. And it's hard to

0:11:45.566 --> 0:11:47.286
<v Speaker 2>patent like a pure math solution.

0:11:47.546 --> 0:11:50.326
<v Speaker 1>Yeah. Has anybody figured it out? Can you tell?

0:11:50.346 --> 0:11:54.116
<v Speaker 2>No, it's been two or three years and No competitor

0:11:54.136 --> 0:11:57.856
<v Speaker 2>we've had is able to— they all stumble on video. Like,

0:11:57.876 --> 0:12:01.076
<v Speaker 2>the volume that we could demonstrate is much higher.

0:12:01.816 --> 0:12:06.276
<v Speaker 1>So we've been talking about what you have called elsewhere defense, right?

0:12:06.446 --> 0:12:12.446
<v Speaker 1>Protecting yourself, protecting ourselves from online fakes and scams, people

0:12:12.626 --> 0:12:16.886
<v Speaker 1>using our likeness without our permission. I've also heard you

0:12:16.926 --> 0:12:22.136
<v Speaker 1>talk about offense, which is— you know, choosing under certain

0:12:22.176 --> 0:12:27.616
<v Speaker 1>circumstances to license our image and likeness for use on

0:12:27.656 --> 0:12:29.956
<v Speaker 1>the internet. I like, I've heard you talk about that

0:12:30.136 --> 0:12:33.966
<v Speaker 1>as this big, you know, underrated asset class that is

0:12:33.986 --> 0:12:37.226
<v Speaker 1>going to blossom. Tell me about that.

0:12:37.866 --> 0:12:40.186
<v Speaker 2>All right. So if you are a big gen AI

0:12:40.246 --> 0:12:42.826
<v Speaker 2>company right now, you either have the choice of trying

0:12:42.846 --> 0:12:45.206
<v Speaker 2>to do a bunch of one-off deals with celebrities, which

0:12:45.286 --> 0:12:49.046
<v Speaker 2>is kind of painful. It's time-consuming. And there's a lot

0:12:49.066 --> 0:12:51.576
<v Speaker 2>of red lines. And then most of the time they

0:12:51.596 --> 0:12:53.636
<v Speaker 2>get stuck in the fact that like, well, who's watching

0:12:53.676 --> 0:12:55.476
<v Speaker 2>you to make sure that bad stuff isn't being made?

0:12:55.936 --> 0:12:58.296
<v Speaker 2>And the tech companies end up saying, well, I'll watch myself.

0:12:58.676 --> 0:13:01.416
<v Speaker 1>There is an example where OpenAI for a while had

0:13:01.436 --> 0:13:06.066
<v Speaker 1>this video generation app called Sora, right? They made this

0:13:06.166 --> 0:13:11.706
<v Speaker 1>deal with Disney where people using Sora could use Disney characters,

0:13:11.766 --> 0:13:14.426
<v Speaker 1>could generate videos with Disney characters. This is the kind

0:13:14.466 --> 0:13:15.726
<v Speaker 1>of thing you're talking about, right?

0:13:15.786 --> 0:13:18.066
<v Speaker 2>Yes. And this is actually a great example because that

0:13:18.146 --> 0:13:20.336
<v Speaker 2>particular deal never really got off the ground because it

0:13:20.376 --> 0:13:22.476
<v Speaker 2>took them months to figure out, well, how do the

0:13:22.536 --> 0:13:23.556
<v Speaker 2>guardrails work for that?

0:13:23.816 --> 0:13:25.976
<v Speaker 1>I thought it never got off the ground because OpenAI

0:13:25.996 --> 0:13:28.516
<v Speaker 1>decided to shut down Sora and focus on other things.

0:13:28.556 --> 0:13:31.836
<v Speaker 2>Well, they shut it down three months after. Right. So

0:13:32.156 --> 0:13:34.256
<v Speaker 2>within the three-month period, they still didn't get a single

0:13:34.296 --> 0:13:35.106
<v Speaker 2>character out the door.

0:13:35.286 --> 0:13:35.586
<v Speaker 1>Uh-huh.

0:13:35.986 --> 0:13:37.446
<v Speaker 2>And so a lot of that had to do with

0:13:37.486 --> 0:13:40.146
<v Speaker 2>the guardrail conversations around, like, well, who protects what?

0:13:40.426 --> 0:13:42.646
<v Speaker 1>Why is this a deal everybody wants to make? Like,

0:13:42.686 --> 0:13:44.486
<v Speaker 1>what's the context here?

0:13:45.026 --> 0:13:48.756
<v Speaker 2>Oh, yeah. Because people want to generate the real world, right?

0:13:49.276 --> 0:13:51.626
<v Speaker 2>So I think that we are going to use this

0:13:51.686 --> 0:13:55.386
<v Speaker 2>as a way to collaboratively dream. Like, let's build this

0:13:55.546 --> 0:13:58.526
<v Speaker 2>story together. Yeah. Or let's build a product together or

0:13:58.546 --> 0:14:00.766
<v Speaker 2>a company or something. And you'll be using Gen AI

0:14:00.826 --> 0:14:04.076
<v Speaker 2>tools to dream those things with others. It'll be important

0:14:04.096 --> 0:14:07.316
<v Speaker 2>in that context to use what the real world looks like. Otherwise,

0:14:07.396 --> 0:14:10.536
<v Speaker 2>the fake version won't quite be the dream that you

0:14:10.576 --> 0:14:11.056
<v Speaker 2>had in mind.

0:14:11.756 --> 0:14:15.746
<v Speaker 1>So you'll... you'll want to make a little movie with

0:14:15.826 --> 0:14:18.986
<v Speaker 1>a friend using generative AI, and you maybe will want

0:14:19.006 --> 0:14:21.216
<v Speaker 1>to have a famous person in it, or you'll want

0:14:21.256 --> 0:14:25.196
<v Speaker 1>it to be at whatever, Starbucks, or at some particular

0:14:25.356 --> 0:14:28.516
<v Speaker 1>store that exists in the world. And if you're actually

0:14:28.536 --> 0:14:31.236
<v Speaker 1>going to sell that or post it online and have,

0:14:31.656 --> 0:14:34.526
<v Speaker 1>you know, have ads run in the middle of it,

0:14:34.896 --> 0:14:36.786
<v Speaker 1>you're not allowed to do it. You're not allowed to

0:14:36.826 --> 0:14:38.906
<v Speaker 1>have it be at an actual Starbucks unless you get

0:14:38.966 --> 0:14:40.046
<v Speaker 1>clearance from Starbucks.

0:14:40.566 --> 0:14:41.506
<v Speaker 2>Correct. Yes.

0:14:42.836 --> 0:14:47.576
<v Speaker 1>So how, like, right now are your celebrity clients selling

0:14:47.616 --> 0:14:50.556
<v Speaker 1>their name, image, and likeness, like, broadly? And are you

0:14:50.776 --> 0:14:51.636
<v Speaker 1>working on that with them?

0:14:52.296 --> 0:14:57.436
<v Speaker 2>Yeah. So you have kind of a bifurcation, which is

0:14:57.456 --> 0:15:00.036
<v Speaker 2>half are like, no, nothing with AI. And the other

0:15:00.056 --> 0:15:01.796
<v Speaker 2>half are like, it's coming and it's here and this

0:15:01.836 --> 0:15:03.916
<v Speaker 2>is something we have to get ahead of. So let's

0:15:03.976 --> 0:15:06.536
<v Speaker 2>make rules and set pricing that makes sense for me.

0:15:07.196 --> 0:15:09.916
<v Speaker 2>And so the interchange system, the clearinghouse system works for both.

0:15:10.726 --> 0:15:12.486
<v Speaker 2>One just blocks everything and the other one says, yes,

0:15:12.506 --> 0:15:13.526
<v Speaker 2>you can make it with these rules.

0:15:14.466 --> 0:15:17.426
<v Speaker 1>So are there examples that are public of people who

0:15:17.486 --> 0:15:20.646
<v Speaker 1>have allowed their likeness to be used in generative AI?

0:15:20.886 --> 0:15:23.726
<v Speaker 2>So like right now we're basically onboarding everyone. We have

0:15:23.746 --> 0:15:26.286
<v Speaker 2>two different gen AI providers that are going to be

0:15:26.346 --> 0:15:29.756
<v Speaker 2>using it. So the announcements for who you can generate

0:15:29.796 --> 0:15:31.456
<v Speaker 2>where will have to come at a different point.

0:15:32.236 --> 0:15:33.996
<v Speaker 1>Give me a sense of what that's going to look like.

0:15:34.096 --> 0:15:36.176
<v Speaker 1>Will there be like lots of movie stars I have

0:15:36.236 --> 0:15:39.256
<v Speaker 1>heard of who are saying it's okay to use my likeness?

0:15:40.886 --> 0:15:43.746
<v Speaker 1>Is Gen AI subject to certain restrictions?

0:15:44.266 --> 0:15:47.096
<v Speaker 2>I think you'll see a certain tier, for lack of

0:15:47.166 --> 0:15:51.836
<v Speaker 2>a better term, we'll say like B celebrities, a large

0:15:51.896 --> 0:15:54.156
<v Speaker 2>group of them. And you'll have kind of a sprinkling

0:15:54.176 --> 0:15:56.456
<v Speaker 2>of a few A stars. And then you'll be able

0:15:56.496 --> 0:15:58.236
<v Speaker 2>to just go to the Gen AI platforms and just

0:15:58.276 --> 0:16:01.586
<v Speaker 2>type in this person's name instead of it getting immediately rejected.

0:16:02.066 --> 0:16:03.566
<v Speaker 2>It says, okay, you can do that. But instead of

0:16:03.586 --> 0:16:04.266
<v Speaker 2>it being $ 0. 30, it's $ 0. 35 online.

0:16:05.806 --> 0:16:10.566
<v Speaker 1>Uh-huh. So basically you're saying that this platform charges you

0:16:10.846 --> 0:16:14.946
<v Speaker 1>extra to use a real famous person in the video

0:16:14.966 --> 0:16:17.446
<v Speaker 1>that you're generating and the real famous person gets some

0:16:17.486 --> 0:16:19.096
<v Speaker 1>of that money and you get a little tiny bit

0:16:19.116 --> 0:16:20.256
<v Speaker 1>of that money, presumably.

0:16:20.796 --> 0:16:22.936
<v Speaker 2>Um, well, I mean, we, we actually don't charge our clients.

0:16:22.956 --> 0:16:24.356
<v Speaker 2>We charge the gen AI platforms.

0:16:24.616 --> 0:16:24.856
<v Speaker 1>Uh-huh.

0:16:25.676 --> 0:16:28.576
<v Speaker 2>So if it's 5 cents, that, that goes to the client.

0:16:28.896 --> 0:16:30.576
<v Speaker 2>And then we have a different mechanism of how we

0:16:30.616 --> 0:16:31.756
<v Speaker 2>charge the gen AI companies.

0:16:32.286 --> 0:16:35.236
<v Speaker 1>And what kind of restrictions to your clients will there

0:16:35.296 --> 0:16:38.596
<v Speaker 1>be on the use of their likeness in generative videos?

0:16:38.936 --> 0:16:41.696
<v Speaker 2>There's kind of like a belief system. So you have

0:16:41.716 --> 0:16:47.376
<v Speaker 2>things like religion. And then more tangibly, it's things around sex, drugs,

0:16:47.396 --> 0:16:49.956
<v Speaker 2>and alcohol. And then they're all natural language. So you

0:16:49.976 --> 0:16:53.186
<v Speaker 2>can turn on drugs and say cannabis is just fine.

0:16:53.826 --> 0:16:58.566
<v Speaker 2>Like I'm okay with this. And there's financial products. Like

0:16:58.606 --> 0:17:01.026
<v Speaker 2>some people, they're consistently made to make kind of like

0:17:01.066 --> 0:17:03.586
<v Speaker 2>Bitcoin scams. And so you can have that turned off

0:17:03.646 --> 0:17:06.046
<v Speaker 2>entirely or say, yes, but it's only for this one

0:17:06.266 --> 0:17:09.236
<v Speaker 2>app that I'm being sponsored for. And then the last

0:17:09.266 --> 0:17:12.996
<v Speaker 2>part is literally the sponsorships, right? So you have, uh,

0:17:13.236 --> 0:17:15.356
<v Speaker 2>you know, Nike is sponsoring me. So that always goes

0:17:15.436 --> 0:17:18.676
<v Speaker 2>on a, on a sneaker or, you know, I did

0:17:18.716 --> 0:17:20.836
<v Speaker 2>a deal with hot pockets a few years ago. So

0:17:20.856 --> 0:17:23.016
<v Speaker 2>you can't ever make some kind of claim of me

0:17:23.076 --> 0:17:24.836
<v Speaker 2>saying that hot pockets burn my mouth for the next

0:17:24.856 --> 0:17:25.356
<v Speaker 2>year and a half.

0:17:25.536 --> 0:17:27.516
<v Speaker 1>Yes. What if I want to make a video of

0:17:27.556 --> 0:17:29.986
<v Speaker 1>a famous person saying that they love my shampoo? Uh,

0:17:31.876 --> 0:17:33.536
<v Speaker 1>that I'm selling. Can I do that?

0:17:34.536 --> 0:17:37.856
<v Speaker 2>No, we do understand the difference between whether it's commercialized

0:17:37.916 --> 0:17:40.696
<v Speaker 2>or not. And this is where the defense part comes in,

0:17:40.736 --> 0:17:43.756
<v Speaker 2>which is, let's say that you have them say, happy birthday,

0:17:43.996 --> 0:17:45.736
<v Speaker 2>and you go start a car wash called the happy

0:17:45.776 --> 0:17:49.056
<v Speaker 2>birthday car wash. There's no point when you generate it

0:17:49.076 --> 0:17:51.476
<v Speaker 2>that I would know that that was what the intent was.

0:17:51.946 --> 0:17:54.526
<v Speaker 2>But when it actually gets used out on Instagram, I

0:17:54.556 --> 0:17:56.586
<v Speaker 2>will see, oh, this is used in a commercialized way.

0:17:56.626 --> 0:17:59.406
<v Speaker 2>I'll look up your license and say, oh, that's That

0:17:59.466 --> 0:18:01.726
<v Speaker 2>wasn't how it was purchased. This was a trickery thing.

0:18:01.906 --> 0:18:04.186
<v Speaker 2>So then we go ahead and delete that. So they

0:18:04.226 --> 0:18:06.506
<v Speaker 2>end up having to buy the correct version, but maybe

0:18:06.526 --> 0:18:07.966
<v Speaker 2>they get away with it for an hour or two.

0:18:12.006 --> 0:18:25.336
<v Speaker 1>We'll be back in just a minute. Are there any

0:18:25.416 --> 0:18:29.036
<v Speaker 1>instances where deepfakes are legal? Or the First Amendment might

0:18:29.536 --> 0:18:30.096
<v Speaker 1>be relevant?

0:18:31.706 --> 0:18:36.386
<v Speaker 2>So this is tricky. I would say that no fakes

0:18:36.446 --> 0:18:39.346
<v Speaker 2>is coming down the pipeline, which is a federal act.

0:18:39.806 --> 0:18:42.166
<v Speaker 2>And that is going to require authorization from the person

0:18:42.186 --> 0:18:44.266
<v Speaker 2>that you're deep faking. I think this is a really

0:18:44.476 --> 0:18:47.516
<v Speaker 2>solid place for us to operate from. So between now

0:18:47.536 --> 0:18:49.636
<v Speaker 2>and when that passes, and it looks like it's going

0:18:49.656 --> 0:18:53.076
<v Speaker 2>to pass, there's really strong support on both sides of

0:18:53.096 --> 0:18:55.986
<v Speaker 2>the aisles from tech companies and entertainment. It's kind of

0:18:56.026 --> 0:18:59.066
<v Speaker 2>a massive collaborative multi-year effort, and it looks like it'll

0:18:59.086 --> 0:19:02.306
<v Speaker 2>probably pass early next year. It's going to force tech

0:19:02.346 --> 0:19:07.056
<v Speaker 2>companies to have to ask for authorizations from my neighbor

0:19:07.396 --> 0:19:09.976
<v Speaker 2>or a celebrity down the street. Everyone is going to

0:19:09.996 --> 0:19:13.356
<v Speaker 2>need some kind of authorization system for that. And between

0:19:13.396 --> 0:19:15.296
<v Speaker 2>now and then, you can make a case that it's

0:19:15.336 --> 0:19:18.226
<v Speaker 2>illegal for right of publicity reasons. If you generate a

0:19:18.846 --> 0:19:22.026
<v Speaker 2>deepfake endorsing someone, that's like a Lanham Act violation because

0:19:22.046 --> 0:19:24.476
<v Speaker 2>it's false endorsement. So there's the ways in which it's

0:19:24.576 --> 0:19:25.436
<v Speaker 2>not about the tech.

0:19:25.476 --> 0:19:27.716
<v Speaker 1>Well, the endorsement is the easy case. I mean, the

0:19:27.796 --> 0:19:30.726
<v Speaker 1>one I'm asking you, like, say you do a politician,

0:19:30.886 --> 0:19:32.966
<v Speaker 1>a satirical video of a politician.

0:19:33.046 --> 0:19:33.246
<v Speaker 2>Right.

0:19:33.346 --> 0:19:36.766
<v Speaker 1>Like, yes, we know that satire tends to have more

0:19:36.846 --> 0:19:41.566
<v Speaker 1>First Amendment protection. We know that discussion of political figures

0:19:41.626 --> 0:19:43.366
<v Speaker 1>tends to have more First Amendment protection.

0:19:43.386 --> 0:19:43.846
<v Speaker 2>Yes. Right.

0:19:43.926 --> 0:19:47.826
<v Speaker 1>Like. I don't know. I can understand why you could

0:19:47.846 --> 0:19:49.636
<v Speaker 1>make a case that a deepfake is not like a

0:19:49.676 --> 0:19:53.156
<v Speaker 1>political cartoon, but you could make a case that it is.

0:19:53.176 --> 0:19:56.136
<v Speaker 2>I think I would probably more make the case that

0:19:56.196 --> 0:20:00.076
<v Speaker 2>it is. I think there will be protections there. I

0:20:00.376 --> 0:20:01.816
<v Speaker 2>know that when we talk to both sides of the

0:20:01.926 --> 0:20:04.586
<v Speaker 2>aisle on who we protect, making sure that we get

0:20:04.626 --> 0:20:07.606
<v Speaker 2>First Amendment right is a big concern for everyone. And

0:20:07.666 --> 0:20:08.486
<v Speaker 2>us included.

0:20:08.546 --> 0:20:10.716
<v Speaker 1>I mean, it's a hard one. It's a hard one because.

0:20:12.276 --> 0:20:15.286
<v Speaker 1>As a matter of sort of public interest, you want

0:20:15.306 --> 0:20:17.526
<v Speaker 1>people to understand what is real and what is not real.

0:20:17.966 --> 0:20:18.386
<v Speaker 2>Yes. Right?

0:20:18.666 --> 0:20:20.966
<v Speaker 1>Like, I like satire. I'm a reporter. I like the

0:20:20.986 --> 0:20:24.206
<v Speaker 1>First Amendment. But, like, people being confused about whether the

0:20:24.246 --> 0:20:27.116
<v Speaker 1>president is really saying something or it's a deepfake of

0:20:27.136 --> 0:20:28.716
<v Speaker 1>the president, that seems quite bad.

0:20:29.556 --> 0:20:31.456
<v Speaker 2>I do think that it's going to be. So when

0:20:31.536 --> 0:20:33.716
<v Speaker 2>you did a political cartoon before, it was in the

0:20:33.736 --> 0:20:37.136
<v Speaker 2>political cartoon section. of the newspaper or The Economist or whatever.

0:20:37.436 --> 0:20:39.756
<v Speaker 1>And it's a drawing. Like, no one is going to

0:20:39.776 --> 0:20:40.776
<v Speaker 1>be confused. Yeah.

0:20:40.796 --> 0:20:42.406
<v Speaker 2>Well, I think that there'll just be kind of a

0:20:42.426 --> 0:20:44.546
<v Speaker 2>time and place in the internet for that, which is

0:20:44.566 --> 0:20:46.986
<v Speaker 2>there'll either be a certain part of the site or

0:20:47.226 --> 0:20:49.746
<v Speaker 2>it will be labeled clearly on the video. Right? There'll

0:20:49.766 --> 0:20:50.306
<v Speaker 2>be a restriction.

0:20:50.386 --> 0:20:52.356
<v Speaker 1>Yeah, like a label. A label is good.

0:20:52.636 --> 0:20:54.436
<v Speaker 2>If you labeled it, it would be just fine.

0:20:54.936 --> 0:20:57.176
<v Speaker 1>Is that functional? Like, will that work?

0:20:57.676 --> 0:21:00.256
<v Speaker 2>Well, this is This is where, you know, and not

0:21:00.296 --> 0:21:03.256
<v Speaker 2>to harp on us or to focus on what we do,

0:21:03.316 --> 0:21:05.956
<v Speaker 2>but anyone really that wants to build a business like ours,

0:21:06.456 --> 0:21:08.296
<v Speaker 2>what you would say is what the time it goes

0:21:08.336 --> 0:21:10.116
<v Speaker 2>to generate it, you'd say, oh, I see that you're

0:21:10.136 --> 0:21:12.356
<v Speaker 2>actually making a satirical thing. That doesn't cost you anything

0:21:12.376 --> 0:21:14.296
<v Speaker 2>because that's your first amendment, right? So we're not going

0:21:14.316 --> 0:21:16.256
<v Speaker 2>to charge you. You get a bypass on it. But

0:21:16.276 --> 0:21:17.416
<v Speaker 2>then when we see it out in the wild, if

0:21:17.456 --> 0:21:20.096
<v Speaker 2>you don't have labeling on it, or if you're not

0:21:20.236 --> 0:21:22.466
<v Speaker 2>using it in a satirical way, you're using it to like, say,

0:21:22.516 --> 0:21:25.226
<v Speaker 2>look at what the president said, right? You're trying to

0:21:25.246 --> 0:21:27.126
<v Speaker 2>like make a stance on it and trying to, you know,

0:21:27.226 --> 0:21:29.236
<v Speaker 2>push it off as being real. then we'll do a

0:21:29.256 --> 0:21:31.336
<v Speaker 2>takedown on it. So there's a way that you kind

0:21:31.376 --> 0:21:33.126
<v Speaker 2>of need both systems to work well.

0:21:33.796 --> 0:21:38.086
<v Speaker 1>That makes sense. You mentioned the federal legislation. I've heard

0:21:38.126 --> 0:21:41.486
<v Speaker 1>you talk elsewhere about state laws as well. Is the

0:21:41.546 --> 0:21:45.646
<v Speaker 1>federal legislation going to sort of preempt or render irrelevant

0:21:45.686 --> 0:21:47.846
<v Speaker 1>the state laws? Are there state laws that are important

0:21:47.866 --> 0:21:50.376
<v Speaker 1>that are doing other things that are worth talking about?

0:21:50.496 --> 0:21:52.416
<v Speaker 1>I mean, I seem to recall California being kind of

0:21:52.476 --> 0:21:55.216
<v Speaker 1>at the vanguard of name, image, and likeness stuff in

0:21:55.236 --> 0:21:58.076
<v Speaker 1>the 20th century because of Hollywood, basically, right?

0:21:58.326 --> 0:22:00.656
<v Speaker 2>Yeah. And then, oh, Tennessee is right now with the

0:22:00.736 --> 0:22:01.536
<v Speaker 2>Elvis Act.

0:22:01.756 --> 0:22:01.936
<v Speaker 1>Huh.

0:22:02.336 --> 0:22:04.676
<v Speaker 2>They've made the No Fakes Act essentially for the state.

0:22:05.176 --> 0:22:05.496
<v Speaker 1>I see.

0:22:05.936 --> 0:22:08.436
<v Speaker 2>So if I have a client in Nashville, I have

0:22:08.716 --> 0:22:11.396
<v Speaker 2>a lot of power for them to go and get

0:22:11.416 --> 0:22:14.696
<v Speaker 2>things taken down. If someone made just a benign someone

0:22:14.716 --> 0:22:17.336
<v Speaker 2>walking down the street in AI, if they live in Tennessee,

0:22:17.386 --> 0:22:20.286
<v Speaker 2>it gets taken down. If they don't, I need to

0:22:20.326 --> 0:22:22.426
<v Speaker 2>make a pretty good case about why this harms that

0:22:22.466 --> 0:22:24.306
<v Speaker 2>person and not use AI rules to do it.

0:22:25.446 --> 0:22:28.996
<v Speaker 1>Oh, interesting. So just to be clear- Outside of Tennessee,

0:22:29.036 --> 0:22:31.326
<v Speaker 1>which is to say in most of the country at

0:22:31.366 --> 0:22:34.766
<v Speaker 1>this point, if I make a deep fake of whatever,

0:22:34.926 --> 0:22:42.756
<v Speaker 1>George Clooney, and he's just like, whatever, reciting Shakespeare or something. And, well,

0:22:42.976 --> 0:22:43.656
<v Speaker 1>is that okay?

0:22:45.596 --> 0:22:47.756
<v Speaker 2>Yeah, I mean, it would be allowed. Like, we wouldn't

0:22:47.776 --> 0:22:48.956
<v Speaker 2>be able to necessarily take that down.

0:22:49.526 --> 0:22:51.936
<v Speaker 1>If I sell ads against it, if they're just like

0:22:52.076 --> 0:22:54.046
<v Speaker 1>programmatic ads that show up in the middle of it,

0:22:54.316 --> 0:22:55.436
<v Speaker 1>now can you take it down?

0:22:55.536 --> 0:22:57.316
<v Speaker 2>That's right. It's a violation of right of publicity.

0:22:57.656 --> 0:22:58.956
<v Speaker 1>Because I'm making money off it.

0:22:59.196 --> 0:23:00.236
<v Speaker 2>Right, off of their likeness.

0:23:00.516 --> 0:23:03.336
<v Speaker 1>But if this bill becomes law, then it won't be

0:23:03.386 --> 0:23:03.946
<v Speaker 1>okay anymore.

0:23:04.256 --> 0:23:05.846
<v Speaker 2>Right. Those people will be able to pick what gets

0:23:05.906 --> 0:23:08.826
<v Speaker 2>made of them. With a 1A carve-out, right?

0:23:09.366 --> 0:23:10.906
<v Speaker 1>The First Amendment carve-out, potentially.

0:23:10.946 --> 0:23:11.386
<v Speaker 2>Yeah.

0:23:11.566 --> 0:23:13.846
<v Speaker 1>It's interesting that it's okay to do it now. I

0:23:13.886 --> 0:23:16.046
<v Speaker 1>think that is not intuitive. Yeah.

0:23:16.176 --> 0:23:18.716
<v Speaker 2>Well, I don't want to encourage people to do it. Um,

0:23:18.856 --> 0:23:20.776
<v Speaker 2>you literally, like these people are upset that you were

0:23:20.796 --> 0:23:24.076
<v Speaker 2>stealing part of their private nature, right? Because the people

0:23:24.096 --> 0:23:26.516
<v Speaker 2>are also doing it for people that aren't celebrities. Um,

0:23:26.616 --> 0:23:29.176
<v Speaker 2>you know, doing it, there's deep fakes made of, uh,

0:23:29.276 --> 0:23:31.996
<v Speaker 2>teachers saying crazy things and then them losing their jobs.

0:23:32.016 --> 0:23:35.356
<v Speaker 2>Like there's still harm, right? That comes from it. So

0:23:35.376 --> 0:23:37.856
<v Speaker 2>I don't want to really tell people go on the

0:23:37.906 --> 0:23:40.506
<v Speaker 2>boom rush now and go make it while it's legal. Um,

0:23:40.566 --> 0:23:42.426
<v Speaker 2>you still are creating harm, even if it may not

0:23:42.466 --> 0:23:42.786
<v Speaker 2>be a law.

0:23:43.526 --> 0:23:46.416
<v Speaker 1>Is there a universe where the big platforms just do

0:23:46.476 --> 0:23:50.576
<v Speaker 1>this themselves? Like, why when I sign up for whatever, Instagram,

0:23:50.716 --> 0:23:53.536
<v Speaker 1>can I not just say, no deep fakes of me?

0:23:53.636 --> 0:23:54.916
<v Speaker 1>Why do we need an extra company?

0:23:55.576 --> 0:23:58.406
<v Speaker 2>So every platform, most platforms, and we're good partners with them,

0:23:58.466 --> 0:24:00.086
<v Speaker 2>so I don't want to be disparaging about this, but

0:24:00.186 --> 0:24:02.166
<v Speaker 2>they've tried over the last few years to do this,

0:24:02.686 --> 0:24:06.486
<v Speaker 2>and there's issues with doing it. One is, it's hard

0:24:06.526 --> 0:24:09.776
<v Speaker 2>for them to set up similar rules across all systems. Two,

0:24:10.076 --> 0:24:12.626
<v Speaker 2>it's hard for anyone to believe that they will properly

0:24:12.666 --> 0:24:13.546
<v Speaker 2>gatekeep themselves.

0:24:14.066 --> 0:24:17.346
<v Speaker 1>Because the platforms want more content. They're not incentivized to

0:24:17.386 --> 0:24:18.746
<v Speaker 1>take things down. Right.

0:24:18.766 --> 0:24:20.646
<v Speaker 2>And so you're like, hey, you as the gatekeeper, can

0:24:20.666 --> 0:24:22.466
<v Speaker 2>you generate things? And then if you're a celebrity, it's

0:24:22.486 --> 0:24:25.706
<v Speaker 2>even more complicated because you're saying a biometric release is

0:24:25.766 --> 0:24:28.046
<v Speaker 2>needed for that. And most humans do not want to

0:24:28.066 --> 0:24:30.326
<v Speaker 2>have Google give a full biometric release because they know

0:24:30.396 --> 0:24:32.476
<v Speaker 2>everything about you, right? You also don't need to know

0:24:32.506 --> 0:24:35.536
<v Speaker 2>what you sound like and look like. And these large

0:24:35.576 --> 0:24:37.776
<v Speaker 2>tech companies have been sued for face recognition usage in

0:24:37.796 --> 0:24:41.156
<v Speaker 2>the past because of how they have this weird interstitial

0:24:41.196 --> 0:24:43.516
<v Speaker 2>data set of like everything you've ever gone to online

0:24:43.646 --> 0:24:45.886
<v Speaker 2>and what you look like in real life. Places like

0:24:45.946 --> 0:24:48.926
<v Speaker 2>Illinois and Texas have sued them. So platforms aren't in

0:24:48.966 --> 0:24:51.046
<v Speaker 2>a great place to say, I'll be the person that

0:24:51.066 --> 0:24:53.906
<v Speaker 2>steps up and does this because the technology is largely

0:24:54.046 --> 0:24:57.486
<v Speaker 2>handicapped for them because of this past litigation. And then

0:24:57.506 --> 0:24:59.226
<v Speaker 2>the other part is people don't want to trust them

0:24:59.266 --> 0:25:03.096
<v Speaker 2>with the biometrics. Those releases are pretty insane to think about.

0:25:03.576 --> 0:25:05.796
<v Speaker 2>For us, like this is our only job, right? It's

0:25:05.916 --> 0:25:08.416
<v Speaker 2>easy to give Lodi those biometric releases and Because all

0:25:08.456 --> 0:25:10.036
<v Speaker 2>we're doing is protecting your biometrics.

0:25:10.656 --> 0:25:16.016
<v Speaker 1>I've heard you talk about user agreements as weirdly important

0:25:16.156 --> 0:25:19.836
<v Speaker 1>in this broader conversation. Why? What's going on there?

0:25:20.756 --> 0:25:22.886
<v Speaker 2>There's a land grab for AI rights right now, for

0:25:22.926 --> 0:25:26.466
<v Speaker 2>likeness rights. So if you're a tech company, it doesn't matter.

0:25:26.486 --> 0:25:28.746
<v Speaker 2>It's not even the big ones, even small startups. They're

0:25:28.766 --> 0:25:30.646
<v Speaker 2>trying to figure out how to get as many people

0:25:30.686 --> 0:25:33.366
<v Speaker 2>as possible to do releases on their data, their likeness,

0:25:33.386 --> 0:25:36.436
<v Speaker 2>their biometrics, because that data is really valuable.

0:25:36.826 --> 0:25:39.416
<v Speaker 1>And like if I'm just a normal guy with an iPhone,

0:25:39.456 --> 0:25:42.416
<v Speaker 1>like when I do unlock my phone with my face,

0:25:42.516 --> 0:25:44.726
<v Speaker 1>am I participating in this without knowing it?

0:25:45.116 --> 0:25:48.176
<v Speaker 2>No, but actually Apple, if you use an iPhone, they

0:25:48.256 --> 0:25:49.996
<v Speaker 2>are aware of the same kind of problem sets that

0:25:50.016 --> 0:25:51.556
<v Speaker 2>we are, which is they make sure that that's all

0:25:51.616 --> 0:25:52.316
<v Speaker 2>local to the phone.

0:25:52.716 --> 0:25:55.536
<v Speaker 1>Oh, interesting. So like Apple doesn't know what my face

0:25:55.576 --> 0:25:57.796
<v Speaker 1>looks like. It's just my phone that knows what my

0:25:57.836 --> 0:25:59.136
<v Speaker 1>face looks like. And that's good.

0:25:59.596 --> 0:26:01.466
<v Speaker 2>There's a world that's coming in which all that's going

0:26:01.486 --> 0:26:02.326
<v Speaker 2>to get legislated.

0:26:02.666 --> 0:26:03.006
<v Speaker 1>I see.

0:26:03.166 --> 0:26:05.286
<v Speaker 2>I think that's the next thing after no fakes. Now

0:26:05.306 --> 0:26:08.046
<v Speaker 2>we're all going to decide what constitutes the data that

0:26:08.086 --> 0:26:11.066
<v Speaker 2>goes into making a fake, right? I think it's coming

0:26:11.766 --> 0:26:14.726
<v Speaker 2>and it's probably going to hurt a lot of smaller companies,

0:26:14.746 --> 0:26:16.446
<v Speaker 2>but the bigger ones now are starting to see that

0:26:16.526 --> 0:26:18.086
<v Speaker 2>is what's on the horizon and they might as well

0:26:18.106 --> 0:26:19.526
<v Speaker 2>just design their program accordingly.

0:26:20.226 --> 0:26:23.696
<v Speaker 1>Yeah, I mean, often regulation... ends up being a moat

0:26:23.796 --> 0:26:26.836
<v Speaker 1>for big companies, right? It is essentially a barrier to entry.

0:26:26.896 --> 0:26:30.056
<v Speaker 1>It makes it more expensive to do things. And so ultimately,

0:26:30.096 --> 0:26:32.316
<v Speaker 1>it's good for the big incumbents.

0:26:32.396 --> 0:26:36.596
<v Speaker 2>That's absolutely true. You have that plus regulatory capture. And

0:26:36.936 --> 0:26:39.816
<v Speaker 2>it's definitely something that we'll see in the AI industry soon.

0:26:40.376 --> 0:26:47.506
<v Speaker 1>Interesting. Let's talk more about the future. We've talked about legislation.

0:26:47.546 --> 0:26:51.466
<v Speaker 1>We've talked about technology. We've talked about rights. But like,

0:26:52.706 --> 0:26:58.476
<v Speaker 1>When you think, uh, you know, about whatever amount of

0:26:58.536 --> 0:27:00.356
<v Speaker 1>time it makes sense to think about five years out,

0:27:00.376 --> 0:27:01.976
<v Speaker 1>I don't know, 10 years, who knows what's going to

0:27:01.996 --> 0:27:03.996
<v Speaker 1>go on in 10 years, right? It's so crazy. It's

0:27:04.016 --> 0:27:07.896
<v Speaker 1>so fast. Um, first of all, what do you worry about?

0:27:10.706 --> 0:27:17.046
<v Speaker 2>Man. Um, I worry about us being stuck in 2024

0:27:17.046 --> 0:27:20.186
<v Speaker 2>art and culture for a long time because of how

0:27:20.246 --> 0:27:23.776
<v Speaker 2>machine learning is trained that, uh, You know, you basically

0:27:23.816 --> 0:27:26.366
<v Speaker 2>have the corpus of human creativity that built models a

0:27:26.406 --> 0:27:28.926
<v Speaker 2>couple of years ago that we all started using. And

0:27:28.946 --> 0:27:31.106
<v Speaker 2>then everything that comes out of those models will be

0:27:31.186 --> 0:27:34.346
<v Speaker 2>like what the past thousands of years of creative look like.

0:27:35.066 --> 0:27:36.766
<v Speaker 2>And what we won't have is something that kind of

0:27:36.806 --> 0:27:39.726
<v Speaker 2>jumps out. A lot of machine learning is based on

0:27:39.806 --> 0:27:42.326
<v Speaker 2>training on the top of the bell curve of items,

0:27:42.346 --> 0:27:45.646
<v Speaker 2>like what looks like the most, like 50th to 75th percentile.

0:27:45.886 --> 0:27:48.206
<v Speaker 1>Sort of the average, the average. Yeah.

0:27:48.536 --> 0:27:50.886
<v Speaker 2>So you have the average of 2024 that's coming out

0:27:50.906 --> 0:27:52.886
<v Speaker 2>of models now. And you have to have something that

0:27:52.906 --> 0:27:55.306
<v Speaker 2>looks truly unique. And AI doesn't want to make that.

0:27:55.446 --> 0:27:57.566
<v Speaker 2>It doesn't know how to. And that's where the crap

0:27:57.626 --> 0:27:59.526
<v Speaker 2>lives on one side. And that's where like the really

0:27:59.586 --> 0:28:00.986
<v Speaker 2>amazing stuff lives on the other.

0:28:02.346 --> 0:28:03.906
<v Speaker 1>Of the bell curve. You're talking about sort of the

0:28:03.986 --> 0:28:06.486
<v Speaker 1>tails of the distribution, things that are really bad and

0:28:06.506 --> 0:28:07.506
<v Speaker 1>things that are really good.

0:28:07.586 --> 0:28:11.176
<v Speaker 2>Correct. Right. So you have like this mathematical preventative thing

0:28:11.216 --> 0:28:13.596
<v Speaker 2>that says, hey, I don't know how to do math

0:28:13.616 --> 0:28:16.826
<v Speaker 2>that allows the really cool stuff in without letting the

0:28:16.876 --> 0:28:18.576
<v Speaker 2>bad stuff in. So I'm just going to go ahead

0:28:18.596 --> 0:28:20.016
<v Speaker 2>and say all this big stuff in the middle is

0:28:20.036 --> 0:28:22.236
<v Speaker 2>what I can produce. So the fear I have is

0:28:22.256 --> 0:28:25.376
<v Speaker 2>that we've kind of walked down this path not really

0:28:25.476 --> 0:28:28.896
<v Speaker 2>understanding that this is a feature of the mathematics. It's

0:28:28.916 --> 0:28:31.886
<v Speaker 2>not like someone at OpenAI is making the decision. It's

0:28:31.906 --> 0:28:34.446
<v Speaker 2>that they are using math that is forcing this decision.

0:28:34.866 --> 0:28:36.546
<v Speaker 2>And then it's in 10 or 20 years and we're

0:28:36.626 --> 0:28:40.046
<v Speaker 2>basically generating the same kind of content because nothing can

0:28:40.086 --> 0:28:43.086
<v Speaker 2>really like pop out without getting hammered down by either

0:28:43.226 --> 0:28:45.676
<v Speaker 2>algorithms or or by the generative models themselves.

0:28:45.876 --> 0:28:50.536
<v Speaker 1>Well, if that's true, there's a weird, happy, a silver

0:28:50.596 --> 0:28:53.576
<v Speaker 1>lining to that story, which is that's an opportunity for

0:28:53.646 --> 0:28:57.486
<v Speaker 1>talented humans, right? If it's true that AI can only

0:28:57.546 --> 0:29:03.526
<v Speaker 1>generate kind of mid-creative work... That's good for people who

0:29:03.626 --> 0:29:07.226
<v Speaker 1>are weirdos and make things that are maybe really bad

0:29:07.286 --> 0:29:09.506
<v Speaker 1>or maybe really great or a few people love even

0:29:09.526 --> 0:29:12.686
<v Speaker 1>though most people don't. If that's the case, that's actually

0:29:12.726 --> 0:29:14.496
<v Speaker 1>a happier outcome than I'm worried about.

0:29:14.656 --> 0:29:17.056
<v Speaker 2>And I have a 15-year-old, a 13-year-old, and a 9-year-old,

0:29:17.076 --> 0:29:19.556
<v Speaker 2>and that's what I tell them, is that the world

0:29:19.596 --> 0:29:22.496
<v Speaker 2>is going to get really average quickly, which is going

0:29:22.516 --> 0:29:25.156
<v Speaker 2>to be good for you because you're all very weird.

0:29:25.916 --> 0:29:27.836
<v Speaker 2>You kind of have to figure out what your voice is.

0:29:28.356 --> 0:29:30.216
<v Speaker 2>And in a world that wants to collapse you down,

0:29:30.556 --> 0:29:32.676
<v Speaker 2>that wants to hammer the nail that sticks out, just

0:29:32.716 --> 0:29:34.816
<v Speaker 2>know that your competitive advantage is that you won't.

0:29:35.756 --> 0:29:39.206
<v Speaker 1>I mean, I feel like there's a universe where you

0:29:39.266 --> 0:29:42.446
<v Speaker 1>can play with the AI and get it to make

0:29:42.486 --> 0:29:45.286
<v Speaker 1>stuff out in the tails, right? You can use AI

0:29:45.366 --> 0:29:49.686
<v Speaker 1>to make weirder stuff by kind of iterating, right? And

0:29:51.306 --> 0:29:54.156
<v Speaker 1>to what extent you need a person doing that, you know,

0:29:54.256 --> 0:29:56.296
<v Speaker 1>like to what extent it's just one shot where it's

0:29:56.336 --> 0:29:59.696
<v Speaker 1>like make something really weird and different in this way

0:29:59.716 --> 0:30:02.436
<v Speaker 1>versus to what extent it's like, well, I write something

0:30:02.456 --> 0:30:03.876
<v Speaker 1>and then I ask the AI about it and it

0:30:03.896 --> 0:30:06.076
<v Speaker 1>gives me some ideas and I write some more. is

0:30:06.276 --> 0:30:08.806
<v Speaker 1>to me kind of the open question and a meaningful

0:30:08.866 --> 0:30:12.846
<v Speaker 1>open question because the less you need of me, maybe

0:30:12.906 --> 0:30:13.926
<v Speaker 1>the worse at some level.

0:30:14.506 --> 0:30:16.886
<v Speaker 2>So this is where I get optimistic about AI is

0:30:16.906 --> 0:30:19.606
<v Speaker 2>because I do think that humans can push AI to

0:30:19.646 --> 0:30:22.546
<v Speaker 2>make something extraordinary, but it's the difference between being the

0:30:22.586 --> 0:30:25.756
<v Speaker 2>painter or the paintbrush, right? Like if AI is the paintbrush,

0:30:25.996 --> 0:30:28.256
<v Speaker 2>you're good, right? But as long as you, as soon

0:30:28.276 --> 0:30:30.516
<v Speaker 2>as you say, well, actually AI write a script for

0:30:30.556 --> 0:30:33.766
<v Speaker 2>me or paint the painting for me, Um, and it

0:30:33.806 --> 0:30:36.186
<v Speaker 2>becomes more of the painter. You kind of are screwed.

0:30:36.206 --> 0:30:37.546
<v Speaker 2>You're in that spot of like, it's going to make

0:30:37.566 --> 0:30:39.866
<v Speaker 2>something that's average, but you're right. If you're a human

0:30:39.886 --> 0:30:42.006
<v Speaker 2>that's involved in the process and you kind of push

0:30:42.066 --> 0:30:44.966
<v Speaker 2>it beyond its boundaries, you are then taking it out

0:30:45.006 --> 0:30:47.476
<v Speaker 2>of like the architect role and more into the, like

0:30:47.606 --> 0:30:49.676
<v Speaker 2>the engineer, like just do this thing. I know it

0:30:49.696 --> 0:30:52.116
<v Speaker 2>looks weird or sounds weird, but do it anyway. Um,

0:30:52.136 --> 0:30:55.656
<v Speaker 2>that's like taste, right? That that's human creativity getting imbued

0:30:55.696 --> 0:30:58.336
<v Speaker 2>into the AI product are very bullish on that. I

0:30:58.356 --> 0:31:01.796
<v Speaker 2>think that that's where AI will actually help our species.

0:31:02.656 --> 0:31:04.656
<v Speaker 2>I just don't like the idea that some people will say, well,

0:31:04.676 --> 0:31:07.716
<v Speaker 2>they'll be writing movies soon. Like, hopefully not. Or they'll

0:31:07.736 --> 0:31:10.436
<v Speaker 2>be very mid. Or they'll be writing music soon. Hopefully

0:31:10.456 --> 0:31:12.676
<v Speaker 2>a human that wants to make a movie or wants

0:31:12.716 --> 0:31:15.016
<v Speaker 2>to make a song says, like, there's something that is

0:31:15.076 --> 0:31:17.226
<v Speaker 2>so out of this world I have to use AI

0:31:17.266 --> 0:31:20.066
<v Speaker 2>for that clip of it. That's cool. I'm all for that.

0:31:20.506 --> 0:31:22.146
<v Speaker 1>Yeah, I mean, I hope so. You used the word

0:31:22.206 --> 0:31:24.566
<v Speaker 1>taste in there, which is a word people have been

0:31:24.646 --> 0:31:26.666
<v Speaker 1>using a lot, right? People have taste. I don't know.

0:31:29.246 --> 0:31:34.356
<v Speaker 1>That's like an interesting big... conversation about the future of

0:31:34.396 --> 0:31:37.186
<v Speaker 1>sort of human creativity and AI. If we think more

0:31:37.276 --> 0:31:41.766
<v Speaker 1>narrowly about the business that you are in, you know,

0:31:41.846 --> 0:31:48.066
<v Speaker 1>people's own likeness in this world of exploding AI, what

0:31:48.086 --> 0:31:50.906
<v Speaker 1>do you worry about in the, you know, next few years?

0:31:50.926 --> 0:31:53.046
<v Speaker 2>I would say the problems I want to solve for

0:31:53.346 --> 0:31:57.446
<v Speaker 2>are making sure that there's very low friction so people

0:31:57.486 --> 0:31:59.846
<v Speaker 2>can control their likeness, whether you are a celebrity or

0:31:59.886 --> 0:32:02.006
<v Speaker 2>you're like me and you're just some average guy walking

0:32:02.046 --> 0:32:05.156
<v Speaker 2>around I think that that's a difficult problem to solve. Like,

0:32:05.196 --> 0:32:06.856
<v Speaker 2>how do you do that at scale? Which friend of

0:32:06.896 --> 0:32:08.636
<v Speaker 2>mine is allowed to generate me as well as a

0:32:08.696 --> 0:32:09.396
<v Speaker 2>weird problem?

0:32:09.956 --> 0:32:11.076
<v Speaker 1>I'm going to just say none.

0:32:11.476 --> 0:32:14.396
<v Speaker 2>Yes. You might say none or, or what you might

0:32:14.496 --> 0:32:17.586
<v Speaker 2>actually find out is, um, you have a parent that

0:32:17.666 --> 0:32:21.536
<v Speaker 2>is ill and they're like, actually, um, I do want

0:32:21.556 --> 0:32:23.876
<v Speaker 2>to make sure that even after I'm gone, I'm reading

0:32:23.916 --> 0:32:27.036
<v Speaker 2>books to your kids. Right. And then you're like, okay,

0:32:27.076 --> 0:32:27.816
<v Speaker 2>that's an exception.

0:32:27.856 --> 0:32:28.016
<v Speaker 1>Right.

0:32:28.236 --> 0:32:29.936
<v Speaker 2>And then you'll end up finding these kinds of outliers

0:32:29.956 --> 0:32:31.776
<v Speaker 2>where you're like, can my partner make something? They had

0:32:31.816 --> 0:32:33.036
<v Speaker 2>a dream of me and they want to like show

0:32:33.056 --> 0:32:35.406
<v Speaker 2>me the exact dream they had of me. And it's

0:32:35.426 --> 0:32:37.566
<v Speaker 2>going to be cathartic for them. Yeah, sure. I don't

0:32:37.606 --> 0:32:39.406
<v Speaker 2>mind if my partner, I trust them with everything. Might

0:32:39.426 --> 0:32:41.026
<v Speaker 2>as well trust them with that. I think there'll be

0:32:41.066 --> 0:32:43.146
<v Speaker 2>slivers of things. You're going to have like high school

0:32:43.166 --> 0:32:45.646
<v Speaker 2>kids that'll be able to give their likeness because their

0:32:45.686 --> 0:32:47.846
<v Speaker 2>friends always love to create crazy stuff.

0:32:47.906 --> 0:32:48.086
<v Speaker 1>Right.

0:32:48.786 --> 0:32:51.796
<v Speaker 2>Um, And you need good guardrails around that too. And

0:32:51.836 --> 0:32:54.496
<v Speaker 2>so I think there'll be pockets of things that people

0:32:54.536 --> 0:32:56.926
<v Speaker 2>that even traditionally would be staunch, be like, I'm not

0:32:56.966 --> 0:32:59.066
<v Speaker 2>going to do anything with AI, will say, well, actually,

0:33:00.706 --> 0:33:02.766
<v Speaker 2>I think that that's cool. And then it'll just grow

0:33:02.786 --> 0:33:03.086
<v Speaker 2>from there.

0:33:05.466 --> 0:33:08.606
<v Speaker 1>You do seem optimistic. So give me your happy vision

0:33:08.646 --> 0:33:10.066
<v Speaker 1>of the world five or 10 years from now.

0:33:10.666 --> 0:33:15.366
<v Speaker 2>The happy vision is one in which we don't ever

0:33:15.506 --> 0:33:18.636
<v Speaker 2>use AI as the painter. It's humans that have figured

0:33:18.676 --> 0:33:21.396
<v Speaker 2>out how to push the boundary of creativity with this

0:33:21.456 --> 0:33:23.886
<v Speaker 2>new kind of paintbrush. That's the first part. And then

0:33:23.906 --> 0:33:27.186
<v Speaker 2>the second part is people don't see the internet as

0:33:27.226 --> 0:33:29.866
<v Speaker 2>this big wild west of anything can get uploaded of

0:33:29.886 --> 0:33:32.966
<v Speaker 2>you whenever, wherever, and can destroy your life overnight. And

0:33:33.006 --> 0:33:35.106
<v Speaker 2>so I think both of these things, like the privacy

0:33:35.146 --> 0:33:38.336
<v Speaker 2>component and the control and autonomy and agency, hopefully will

0:33:38.376 --> 0:33:40.696
<v Speaker 2>come to the internet. And then the second part, of course,

0:33:40.736 --> 0:33:45.376
<v Speaker 2>being creativity, getting solved for, but solved in my mind

0:33:45.636 --> 0:33:47.116
<v Speaker 2>is more paintbrush and less painter.

0:33:50.996 --> 0:34:03.626
<v Speaker 1>We'll be back in a minute with the lightning round. Okay,

0:34:03.926 --> 0:34:06.276
<v Speaker 1>let's do a lightning round. What's your favorite deepfake?

0:34:10.596 --> 0:34:13.756
<v Speaker 2>The bunny rabbits on the trampoline. Totally harmless. It got

0:34:14.076 --> 0:34:16.446
<v Speaker 2>a million people laughing and enjoying and seeing that AI

0:34:16.466 --> 0:34:19.186
<v Speaker 2>could make some funny, cool stuff and tricked a lot

0:34:19.206 --> 0:34:21.686
<v Speaker 2>of people into thinking they were real as well. So

0:34:21.706 --> 0:34:22.626
<v Speaker 2>I actually did like that.

0:34:23.026 --> 0:34:24.376
<v Speaker 1>I'll have to check it out. I haven't seen it.

0:34:26.706 --> 0:34:30.946
<v Speaker 1>What celebrity likeness would you license to advertise for Lodi,

0:34:31.066 --> 0:34:32.286
<v Speaker 1>to be a spokesman for Lodi?

0:34:33.086 --> 0:34:36.586
<v Speaker 2>Cannot say. Yeah. It's coming out actually in the next

0:34:36.646 --> 0:34:37.986
<v Speaker 2>few weeks, so cannot say.

0:34:38.146 --> 0:34:38.746
<v Speaker 1>Okay.

0:34:38.966 --> 0:34:40.806
<v Speaker 2>Yeah.

0:34:41.546 --> 0:34:45.256
<v Speaker 1>I'm very interested to see. Is there one or will

0:34:45.296 --> 0:34:45.976
<v Speaker 1>there be a bunch?

0:34:46.536 --> 0:34:49.516
<v Speaker 2>There's one that's on the forefront of all of this. Okay.

0:34:50.756 --> 0:34:52.456
<v Speaker 2>Someone that we deeply respect in this space.

0:34:54.036 --> 0:34:57.036
<v Speaker 1>I'm very interested. How do you want your kids to

0:34:57.216 --> 0:34:57.816
<v Speaker 1>use AI?

0:35:00.186 --> 0:35:02.886
<v Speaker 2>I told my 15-year-old that he could probably cheat on

0:35:02.966 --> 0:35:05.866
<v Speaker 2>everything with AI. And that there's probably no way that

0:35:05.966 --> 0:35:08.226
<v Speaker 2>me or a teacher could figure that out. Because I

0:35:08.246 --> 0:35:10.446
<v Speaker 2>know that kids are really crafty. So that he has

0:35:10.486 --> 0:35:12.186
<v Speaker 2>to make the active choice that if he wants to

0:35:12.226 --> 0:35:14.856
<v Speaker 2>be successful, he has to do hard work. And that's

0:35:14.916 --> 0:35:19.936
<v Speaker 2>a completely different skill set. So it really is coming

0:35:19.956 --> 0:35:21.796
<v Speaker 2>down to just, especially if you have a smart kid,

0:35:21.856 --> 0:35:23.356
<v Speaker 2>just tell them, look, I know you can do it.

0:35:23.836 --> 0:35:25.356
<v Speaker 2>But here, let me make the case for why you

0:35:25.376 --> 0:35:27.816
<v Speaker 2>should work harder instead. Is if it's in your head,

0:35:28.316 --> 0:35:30.206
<v Speaker 2>you could then leverage it to do things that other

0:35:30.226 --> 0:35:33.936
<v Speaker 2>people cannot. So it's a little competitive, but It's essentially

0:35:33.976 --> 0:35:36.316
<v Speaker 2>to tell my kids, don't shy away from it. Like

0:35:36.356 --> 0:35:39.476
<v Speaker 2>my 15-year-old's on the $ 200 a month cloud plan. Like

0:35:39.496 --> 0:35:42.336
<v Speaker 2>he uses AI a ton, right? He's like building apps

0:35:42.376 --> 0:35:45.006
<v Speaker 2>and stuff. It's crazy. I'm really proud of him. But

0:35:45.566 --> 0:35:47.606
<v Speaker 2>at the same time, he knows that you can't use

0:35:47.666 --> 0:35:49.786
<v Speaker 2>it to fill out all your homework without you robbing

0:35:49.826 --> 0:35:50.666
<v Speaker 2>yourself of something.

0:35:51.406 --> 0:35:57.046
<v Speaker 1>Yes, right. The point of homework is actually not the whatever,

0:35:57.086 --> 0:35:59.426
<v Speaker 1>the math you're learning. It's learning to just sit there

0:35:59.646 --> 0:36:02.066
<v Speaker 1>and do it, right? That's actually the work.

0:36:02.166 --> 0:36:03.896
<v Speaker 2>Yeah, those are the... So a lot of our discussions

0:36:03.916 --> 0:36:06.196
<v Speaker 2>around AI are actually the reasons why you wouldn't want

0:36:06.216 --> 0:36:09.126
<v Speaker 2>to use it in a home that deeply uses it. Yeah.

0:36:09.246 --> 0:36:12.186
<v Speaker 2>So it's, but I think that that's like the, the

0:36:12.226 --> 0:36:14.146
<v Speaker 2>brake and gas pedal kind of thing. Like you need

0:36:14.186 --> 0:36:15.026
<v Speaker 2>both to run the car.

0:36:15.386 --> 0:36:15.526
<v Speaker 1>Yeah.

0:36:15.546 --> 0:36:17.726
<v Speaker 2>And so a lot of the conversation is around the

0:36:17.766 --> 0:36:21.046
<v Speaker 2>more complicated use of a brake system. What?

0:36:21.166 --> 0:36:23.606
<v Speaker 1>I was trying to think, frankly, of a good question

0:36:23.666 --> 0:36:26.926
<v Speaker 1>about like AI and the future of movie stars. Like,

0:36:27.066 --> 0:36:28.966
<v Speaker 1>I don't know the first movie with an AI genre,

0:36:28.986 --> 0:36:33.316
<v Speaker 1>but I couldn't quite get the right question. Like, make

0:36:33.336 --> 0:36:36.056
<v Speaker 1>a prediction about AI and the future of movie stars.

0:36:37.136 --> 0:36:40.876
<v Speaker 2>I think that there's going to be, there's going to

0:36:40.896 --> 0:36:44.546
<v Speaker 2>be a demand for like authentic performances. That's what they'll

0:36:44.566 --> 0:36:47.006
<v Speaker 2>be called. Something to that effect, some euphemism for it's

0:36:47.046 --> 0:36:49.586
<v Speaker 2>really them in front of film. And I think they're

0:36:49.606 --> 0:36:52.306
<v Speaker 2>going to make movies that portray that a little bit better.

0:36:52.446 --> 0:36:54.466
<v Speaker 2>Like it'll be like human expressions that AI is not

0:36:54.526 --> 0:36:57.456
<v Speaker 2>able to get or more complicated scripts that land in

0:36:57.496 --> 0:37:00.936
<v Speaker 2>a way that's not like uncanny valley. This is a

0:37:00.956 --> 0:37:05.656
<v Speaker 2>good example. There's the, The Silver Springs, Fleetwood Mac, where

0:37:05.856 --> 0:37:07.876
<v Speaker 2>the lead singer and the guitarist basically are in this

0:37:07.936 --> 0:37:12.326
<v Speaker 2>long-term situationship, and they're essentially singing about themselves on stage,

0:37:12.366 --> 0:37:14.026
<v Speaker 2>and everyone can kind of feel that tension. It's a

0:37:14.046 --> 0:37:15.986
<v Speaker 2>human thing. And if you were to swap them out

0:37:16.026 --> 0:37:19.546
<v Speaker 2>for robots, you probably would not feel the same way

0:37:19.586 --> 0:37:22.726
<v Speaker 2>as a human. I think there's something about knowing that

0:37:22.746 --> 0:37:25.666
<v Speaker 2>that person suffered in the way that you've suffered that

0:37:25.746 --> 0:37:27.946
<v Speaker 2>makes you identify with their art that is going to

0:37:27.966 --> 0:37:30.546
<v Speaker 2>be much bigger, and we talk about Taste, I know,

0:37:30.606 --> 0:37:32.876
<v Speaker 2>is a cheesy word, but still is something real here.

0:37:32.936 --> 0:37:35.656
<v Speaker 2>I think taste for that will go up because it'll

0:37:35.676 --> 0:37:37.476
<v Speaker 2>be more rare. And then I think there's going to

0:37:37.496 --> 0:37:40.796
<v Speaker 2>be a bunch of superhero action movies that are made

0:37:40.836 --> 0:37:41.896
<v Speaker 2>in AI and no one's going to care.

0:37:42.896 --> 0:37:46.836
<v Speaker 1>I mean, superhero action movies are already largely made in AI.

0:37:46.856 --> 0:37:48.876
<v Speaker 2>Correct. That's it. Yes. Yes.

0:37:49.356 --> 0:37:55.426
<v Speaker 1>Like, it is interestingly not as binary as as we say, right?

0:37:55.906 --> 0:38:00.166
<v Speaker 2>If your art is making really generic romance songs for radio,

0:38:00.406 --> 0:38:03.596
<v Speaker 2>you probably are in for some competition with AI. But

0:38:03.656 --> 0:38:07.016
<v Speaker 2>if you have a unique sound, unique bass, whatever it

0:38:07.096 --> 0:38:09.756
<v Speaker 2>is that you bring to the table that is truly

0:38:09.816 --> 0:38:12.616
<v Speaker 2>not like something else we hear everywhere else, the demand

0:38:12.636 --> 0:38:13.716
<v Speaker 2>for that's going to go through the roof.

0:38:14.696 --> 0:38:17.386
<v Speaker 1>That's encouraging. I appreciate your time.

0:38:18.046 --> 0:38:18.426
<v Speaker 2>Of course.

0:38:25.836 --> 0:38:30.056
<v Speaker 1>Luke Garagoni is co-founder and CEO of Lodi. Our show

0:38:30.116 --> 0:38:33.556
<v Speaker 1>is produced by Gabriel Hunter Chang and Trina Menino. Our

0:38:33.696 --> 0:38:36.916
<v Speaker 1>editor is Lydia Jean Cott, and our engineer is Sarah Bruguier.

0:38:37.456 --> 0:38:39.726
<v Speaker 1>I'm Jacob Goldstein, and we'll be back next week with

0:38:39.806 --> 0:38:41.006
<v Speaker 1>another episode of What's Your Problem?