Latest / What's Your Problem? / Can We Defeat Deepfakes?

Transcript

Can We Defeat Deepfakes? · What's Your Problem?

0:15
Speaker 1: Pushkin. I'm Jacob Goldstein, and this is What's Your Problem?
0:33
Speaker 1: My guest today is Luke Aragoni. He is the co-founder
0:36
Speaker 1: and CEO of a company called Lodi. And Lodi is
0:39
Speaker 1: an acronym, and it stands for Likeness on the Internet.
0:43
Speaker 1: Tens of thousands of famous people pay Lodi to scour
0:46
Speaker 1: the internet and take down deepfakes and unapproved endorsements. If
0:51
Speaker 1: you've heard of a person, if a person is a
0:53
Speaker 1: little bit famous, there is a very good chance they
0:56
Speaker 1: are one of Lodi's clients. But in Luke's telling, this
0:59
Speaker 1: is just the beginning. The company will soon start offering
1:02
Speaker 1: a service for celebrities to license the use of their image.
1:06
Speaker 1: In other words, ordinary people will be able to pay
1:09
Speaker 1: Lodi to create stuff with AI using the faces and
1:12
Speaker 1: voices of real celebrities. Lodi also plans to start offering
1:16
Speaker 1: its services to ordinary people. The plan is to include
1:20
Speaker 1: a free tier for people who are not famous, but
1:23
Speaker 1: who are worried about what's happening with their likeness on
1:26
Speaker 1: the internet. At a macro level, the biggest level, Luke's
1:29
Speaker 1: problem is this. How do you turn the internet into
1:33
Speaker 1: a place where people have more control over their name, image,
1:37
Speaker 1: and likeness? In our conversation, Luke and I talked about
1:41
Speaker 1: the technology behind Lodi. We talked about federal legislation that
1:45
Speaker 1: may soon make it easier for people to fight deep fakes.
1:49
Speaker 1: And we also talk about AI and the future of creativity.
1:52
Speaker 1: To start, I asked Luke to give me the state
1:54
Speaker 1: of play for what's happening right now with unauthorized use
1:58
Speaker 1: of people's likeness on the internet.
2:01
Speaker 2: So you have kind of two different categories of people.
2:03
Speaker 2: If you're really famous and your likeness is going to
2:06
Speaker 2: get used to sell random things. It could be to
2:10
Speaker 2: sell a scam. It's like an impersonation, a fake Instagram
2:13
Speaker 2: of a celebrity, and they're selling you anything from VIP
2:15
Speaker 2: access to the next tour to a romance scam that
2:19
Speaker 2: they're falling in love with you, right? And all of
2:21
Speaker 2: them have some kind of economic extraction from it. Or
2:25
Speaker 2: it's kind of more straightforward. It's like, hey, I've come
2:28
Speaker 2: up with a brand new shampoo, and this famous celebrity
2:30
Speaker 2: uses the shampoo. Here's a deepfake video talking about how
2:33
Speaker 2: they love it. So that's probably the most common way
2:37
Speaker 2: in which it's used. And then you have a kind
2:39
Speaker 2: of more incendiary thing where everyone else, like me, for example,
2:44
Speaker 2: what we have to be concerned about is someone deepfaking
2:47
Speaker 2: us for any kind of authorization reason. And then there's
2:51
Speaker 2: an even darker way in which it's used that we
2:53
Speaker 2: saw at some high schools this year, which is a
2:56
Speaker 2: lot of folks are getting deepfaked to make images of
2:59
Speaker 2: them that were never taken, but have a very harmful
3:02
Speaker 2: nature to them and their reputation.
3:04
Speaker 1: Do you mean revenge porn? You're talking around revenge porn there.
3:08
Speaker 2: Yes. We like to call it non-consensual intimate images, which
3:12
Speaker 2: is a lot better of a term than revenge porn.
3:14
Speaker 2: But colloquially, we all call it revenge porn, and so
3:17
Speaker 2: we even have to kind of mix the terms together.
3:19
Speaker 2: But just to kind of re-educate a lot of people,
3:21
Speaker 2: NCII is a much better term for it.
3:23
Speaker 1: Okay, fair enough. So this is a potential problem for everybody.
3:29
Speaker 1: What's like... What's the scope of it right now? Like,
3:33
Speaker 1: if I just pick some random A-list celebrity, will there, like, 100%
3:37
Speaker 1: be deepfakes of them scamming on the internet right now?
3:42
Speaker 2: Yeah, we track about 30,000 or so celebrity accounts, people, personas.
3:49
Speaker 1: And those are your clients, or that's just you kind
3:51
Speaker 1: of understanding the landscape?
3:53
Speaker 2: Two-thirds, a little bit more, a little over 20,000 are clients.
3:57
Speaker 1: So that's way below the A-list. That's way more famous
4:00
Speaker 1: people than anybody could name. Oh, yeah.
4:02
Speaker 2: No, it is. It's pretty much every famous person that
4:04
Speaker 2: you can think of, even if you don't remember their name.
4:07
Speaker 2: And so, yes, for those folks, you're going to get
4:09
Speaker 2: nearly 100% coverage on impersonations and other scams.
4:13
Speaker 1: When you say nearly 100% coverage, you mean for all
4:16
Speaker 1: of those people, there are going to be deep fakes.
4:18
Speaker 2: Of them on the Internet? Yes, and impersonations. So it's
4:21
Speaker 2: like a real picture of them on an Instagram account,
4:24
Speaker 2: but not their real account.
4:25
Speaker 1: Right, so that's not a deepfake in the AI sense.
4:28
Speaker 1: You can just take the picture of the person and
4:30
Speaker 1: be like, look, this person loves our soap, and that's
4:33
Speaker 1: a scam and a lie.
4:34
Speaker 2: Absolutely.
4:35
Speaker 1: And just to be clear, like, that is illegal because
4:38
Speaker 1: people have a right to their name, image, and likeness? Like,
4:40
Speaker 1: that's the fundamental business and kind of intellectual property thing
4:44
Speaker 1: that we're talking about?
4:45
Speaker 2: Yeah, it's called a right of publicity. And ironically, it
4:48
Speaker 2: became a bigger conversation when the camera came around. Uh-huh.
4:52
Speaker 2: And so, you know, 100 plus years ago when people
4:54
Speaker 2: started taking pictures of anyone, someone was like, well, I'll
4:57
Speaker 2: put this on the front of my cereal box. This
4:59
Speaker 2: is just like an average person. They're like, no, that
5:01
Speaker 2: person has, all right, you can't just make, it really
5:04
Speaker 2: is the commercial component. You can't make money off their
5:06
Speaker 2: likeness without their approval or their endorsement.
5:11
Speaker 1: Okay, so that's the lay of the land. Your company
5:15
Speaker 1: sort of walks into this universe as it is growing
5:19
Speaker 1: a few years ago. Like, Basically, what's your company do?
5:23
Speaker 2: What we do is we download the internet every day,
5:25
Speaker 2: and then we run face recognition over it looking for
5:28
Speaker 2: our clients. And when we find them in a spot
5:30
Speaker 2: that they don't belong, whether it be a scam, impersonation, NCII,
5:34
Speaker 2: or any other type of thing, we figure out how
5:37
Speaker 2: we get this thing removed. So the removal part is
5:40
Speaker 2: another cool piece of tech because it's fully automated. But
5:43
Speaker 2: we can reach out to adult sites and tell them
5:46
Speaker 2: they are creating liability for their merchant accounts or, you know,
5:50
Speaker 2: they're domain provider, or a number of other ways. And
5:53
Speaker 2: then on social media, we kind of have this great
5:55
Speaker 2: access and great partnership across everyone but X. And they
5:59
Speaker 2: just give us a kind of API access to issue
6:02
Speaker 2: those takedowns.
6:03
Speaker 1: And so does that mean that it's fully automated, like
6:07
Speaker 1: your AI, essentially your facial recognition software, spots these unauthorized
6:15
Speaker 1: uses of people's likeness and then pings the software or
6:19
Speaker 1: the AI of whatever, YouTube or TikTok, and says, hey,
6:23
Speaker 1: take this down, and then the machine just takes it down?
6:25
Speaker 1: Like there's no human needs to.
6:28
Speaker 2: There's no human needed for this.
6:29
Speaker 1: And presumably that's because the volume is too high. Like
6:34
Speaker 1: what order of magnitude, how many, you know, how many
6:37
Speaker 1: things has your company removed, whatever, today or this week?
6:42
Speaker 2: We do tens of thousands of things a day.
6:45
Speaker 1: Tens of thousands a day.
6:46
Speaker 2: Yeah, it's pretty insane.
6:48
Speaker 1: That is insane. So, like, at every moment, at every second,
6:52
Speaker 1: someone is posting a scam likeness. I guess presumably at
6:56
Speaker 1: this point it's machines. Like, you don't need a person. Like,
6:59
Speaker 1: I could just build a bot to create a million
7:03
Speaker 1: scam likenesses.
7:04
Speaker 2: Correct. We called it an industrial game of whack-a-mole.
7:07
Speaker 1: Yeah. Like, what's the frontier? What are you struggling with
7:12
Speaker 1: right now?
7:12
Speaker 2: Hmm. I think a lot of people don't understand that
7:16
Speaker 2: you can have delete for the internet. And so we
7:19
Speaker 2: think about how we position the company into a greater
7:23
Speaker 2: context around like, okay, there's a thing of me online.
7:27
Speaker 2: It's going to be online forever. And we kind of
7:28
Speaker 2: have to interject ourselves and say, well, this won't be
7:30
Speaker 2: online forever. I know everyone was told that for 30 years,
7:34
Speaker 2: but now there's a really efficient way for us to
7:36
Speaker 2: manage that.
7:37
Speaker 1: What's the business model?
7:38
Speaker 2: So for celebrities, we charge a monthly fee to go
7:41
Speaker 2: scraping and removing this content. And for consumers, it'll be
7:45
Speaker 2: more freemium. They can find everything of themselves online, and
7:49
Speaker 2: then they get a certain number of takedowns if they
7:51
Speaker 2: upgrade to Pro. Or we actually teach them how they
7:53
Speaker 2: can go do the takedowns themselves, so we're not trying
7:56
Speaker 2: to gatekeep their privacy.
7:58
Speaker 1: When's it going to be available for consumers?
8:00
Speaker 2: Probably early October.
8:02
Speaker 1: Technically, how is it going?
8:05
Speaker 2: Oh, it's both a technical nightmare and a marvel. It's insane.
8:12
Speaker 2: because the problem set of downloading like 600 million assets
8:15
Speaker 2: a day, you have database problems and disk problems.
8:18
Speaker 1: Why do you need to download it?
8:21
Speaker 2: Because that's how you run the face recognition on it. Okay.
8:24
Speaker 1: Interesting. Yeah.
8:25
Speaker 2: And so you download it, you run your face recognition
8:28
Speaker 2: on it, and then that gives you the capability of saying,
8:31
Speaker 2: this person's coming to me. Can you find everything in
8:33
Speaker 2: my online? We look into our database. We say, oh,
8:36
Speaker 2: we can now identify these rows as you. The other
8:38
Speaker 2: things we toss out.
8:39
Speaker 1: Yeah. And are the Bad people doing things to make
8:44
Speaker 1: it harder for you?
8:46
Speaker 2: No, they're making it easier. They're trying to make their
8:48
Speaker 2: AI-generated faces more like the person, which actually is an
8:53
Speaker 2: alignment to us. It's like the more their deepfakes look realistic,
8:57
Speaker 2: the easier it is for us to catch them.
8:59
Speaker 1: Oh, that's interesting. And so presumably, I mean, the next
9:03
Speaker 1: step is figuring out what are the legitimate uses, right? Correct.
9:06
Speaker 1: In an automated way. How do you do that part?
9:09
Speaker 2: So that's more of the context understanding of what the
9:11
Speaker 2: image is. We actually do that discovery as well. We
9:14
Speaker 2: have this thing that we call a semantic understanding of
9:17
Speaker 2: every image and video where it comes through and we say,
9:20
Speaker 2: what's going on in this? Is someone running in this park?
9:22
Speaker 2: Are they walking in the park? Are they doing drugs?
9:25
Speaker 2: Are they doing something they shouldn't be doing in this
9:27
Speaker 2: picture or image? And then we have this massive collection of,
9:30
Speaker 2: we know that it's this celebrity doing this thing. And
9:33
Speaker 2: was there like an expectation of privacy here? Or More importantly,
9:38
Speaker 2: is this authorized? And they can give us their rules
9:40
Speaker 2: around what should or shouldn't be authorized.
9:43
Speaker 1: And presumably the giving the rules of what should and
9:46
Speaker 1: shouldn't be authorized also needs to be automated, right? Like
9:49
Speaker 1: presumably the clients need to say, these are the things
9:52
Speaker 1: I am endorsing. It's sort of a white list. Is
9:54
Speaker 1: that the system?
9:55
Speaker 2: Yes. Absolutely.
10:00
Speaker 1: Are there any technical problems you're trying to solve or
10:02
Speaker 1: is it technically a sort of solved issue for you?
10:04
Speaker 2: Every couple of months we we trip on something that
10:07
Speaker 2: is kind of insane. We're like, that's oh, I didn't
10:10
Speaker 2: realize we'd even have to solve that problem.
10:12
Speaker 1: What's an example? What's an example?
10:14
Speaker 2: Well, when we went to process video, we found out
10:16
Speaker 2: the reason why no one had done what we do
10:18
Speaker 2: before us, like no one did this in 2005 was
10:21
Speaker 2: because it's expensive to process video. So like to take
10:24
Speaker 2: a video in and actually like look at all the frames.
10:27
Speaker 2: even without great machine learning or AI at that time,
10:29
Speaker 2: it was just really hard to do that. And so
10:32
Speaker 2: we were broke and we couldn't afford to just spool
10:35
Speaker 2: up more servers like some really large AI companies are
10:37
Speaker 2: able to do. And so necessity was the mother of invention.
10:40
Speaker 2: And we invented a method of processing videos that were
10:43
Speaker 2: like a thousand times faster than anyone else. And we
10:46
Speaker 2: just did this thinking like, well, this is what you
10:48
Speaker 2: have to do, right? Like this is just whatever the
10:50
Speaker 2: next step is. And it came to find out that we, uh,
10:53
Speaker 2: We had to build something that no one else had
10:55
Speaker 2: done that others had tried to do and failed. Now
10:57
Speaker 2: it's like, it's one of the most guarded, protected algorithms
11:02
Speaker 2: at our company because before it was just kind of like,
11:04
Speaker 2: it was like floating everywhere.
11:05
Speaker 1: Did you patent it or is it a trade secret?
11:07
Speaker 2: Oh, interesting. It's a trade secret.
11:09
Speaker 1: So you can do a thing and you're not telling
11:11
Speaker 1: anybody how to do it.
11:12
Speaker 2: Correct. And people know that we do it. They still
11:14
Speaker 2: can't figure it out.
11:16
Speaker 1: Oh, that's interesting.
11:16
Speaker 2: We can prove it. And that's the only part that matters.
11:19
Speaker 1: And what it allows you to do is to essentially...
11:22
Speaker 1: have AI look at a video much more efficiently and
11:27
Speaker 1: cheaply than anybody else can or than anybody else could before? Interesting.
11:30
Speaker 2: That's right.
11:33
Speaker 1: And presumably you chose not to patent it because when
11:35
Speaker 1: you patent something, you have to tell the world how
11:37
Speaker 1: you did it and then people can copy it.
11:39
Speaker 2: And then they can tweak one part of the mathematics
11:42
Speaker 2: and essentially have something like that. And it's hard to
11:45
Speaker 2: patent like a pure math solution.
11:47
Speaker 1: Yeah. Has anybody figured it out? Can you tell?
11:50
Speaker 2: No, it's been two or three years and No competitor
11:54
Speaker 2: we've had is able to— they all stumble on video. Like,
11:57
Speaker 2: the volume that we could demonstrate is much higher.
12:01
Speaker 1: So we've been talking about what you have called elsewhere defense, right?
12:06
Speaker 1: Protecting yourself, protecting ourselves from online fakes and scams, people
12:12
Speaker 1: using our likeness without our permission. I've also heard you
12:16
Speaker 1: talk about offense, which is— you know, choosing under certain
12:22
Speaker 1: circumstances to license our image and likeness for use on
12:27
Speaker 1: the internet. I like, I've heard you talk about that
12:30
Speaker 1: as this big, you know, underrated asset class that is
12:33
Speaker 1: going to blossom. Tell me about that.
12:37
Speaker 2: All right. So if you are a big gen AI
12:40
Speaker 2: company right now, you either have the choice of trying
12:42
Speaker 2: to do a bunch of one-off deals with celebrities, which
12:45
Speaker 2: is kind of painful. It's time-consuming. And there's a lot
12:49
Speaker 2: of red lines. And then most of the time they
12:51
Speaker 2: get stuck in the fact that like, well, who's watching
12:53
Speaker 2: you to make sure that bad stuff isn't being made?
12:55
Speaker 2: And the tech companies end up saying, well, I'll watch myself.
12:58
Speaker 1: There is an example where OpenAI for a while had
13:01
Speaker 1: this video generation app called Sora, right? They made this
13:06
Speaker 1: deal with Disney where people using Sora could use Disney characters,
13:11
Speaker 1: could generate videos with Disney characters. This is the kind
13:14
Speaker 1: of thing you're talking about, right?
13:15
Speaker 2: Yes. And this is actually a great example because that
13:18
Speaker 2: particular deal never really got off the ground because it
13:20
Speaker 2: took them months to figure out, well, how do the
13:22
Speaker 2: guardrails work for that?
13:23
Speaker 1: I thought it never got off the ground because OpenAI
13:25
Speaker 1: decided to shut down Sora and focus on other things.
13:28
Speaker 2: Well, they shut it down three months after. Right. So
13:32
Speaker 2: within the three-month period, they still didn't get a single
13:34
Speaker 2: character out the door.
13:35
Speaker 1: Uh-huh.
13:35
Speaker 2: And so a lot of that had to do with
13:37
Speaker 2: the guardrail conversations around, like, well, who protects what?
13:40
Speaker 1: Why is this a deal everybody wants to make? Like,
13:42
Speaker 1: what's the context here?
13:45
Speaker 2: Oh, yeah. Because people want to generate the real world, right?
13:49
Speaker 2: So I think that we are going to use this
13:51
Speaker 2: as a way to collaboratively dream. Like, let's build this
13:55
Speaker 2: story together. Yeah. Or let's build a product together or
13:58
Speaker 2: a company or something. And you'll be using Gen AI
14:00
Speaker 2: tools to dream those things with others. It'll be important
14:04
Speaker 2: in that context to use what the real world looks like. Otherwise,
14:07
Speaker 2: the fake version won't quite be the dream that you
14:10
Speaker 2: had in mind.
14:11
Speaker 1: So you'll... you'll want to make a little movie with
14:15
Speaker 1: a friend using generative AI, and you maybe will want
14:19
Speaker 1: to have a famous person in it, or you'll want
14:21
Speaker 1: it to be at whatever, Starbucks, or at some particular
14:25
Speaker 1: store that exists in the world. And if you're actually
14:28
Speaker 1: going to sell that or post it online and have,
14:31
Speaker 1: you know, have ads run in the middle of it,
14:34
Speaker 1: you're not allowed to do it. You're not allowed to
14:36
Speaker 1: have it be at an actual Starbucks unless you get
14:38
Speaker 1: clearance from Starbucks.
14:40
Speaker 2: Correct. Yes.
14:42
Speaker 1: So how, like, right now are your celebrity clients selling
14:47
Speaker 1: their name, image, and likeness, like, broadly? And are you
14:50
Speaker 1: working on that with them?
14:52
Speaker 2: Yeah. So you have kind of a bifurcation, which is
14:57
Speaker 2: half are like, no, nothing with AI. And the other
15:00
Speaker 2: half are like, it's coming and it's here and this
15:01
Speaker 2: is something we have to get ahead of. So let's
15:03
Speaker 2: make rules and set pricing that makes sense for me.
15:07
Speaker 2: And so the interchange system, the clearinghouse system works for both.
15:10
Speaker 2: One just blocks everything and the other one says, yes,
15:12
Speaker 2: you can make it with these rules.
15:14
Speaker 1: So are there examples that are public of people who
15:17
Speaker 1: have allowed their likeness to be used in generative AI?
15:20
Speaker 2: So like right now we're basically onboarding everyone. We have
15:23
Speaker 2: two different gen AI providers that are going to be
15:26
Speaker 2: using it. So the announcements for who you can generate
15:29
Speaker 2: where will have to come at a different point.
15:32
Speaker 1: Give me a sense of what that's going to look like.
15:34
Speaker 1: Will there be like lots of movie stars I have
15:36
Speaker 1: heard of who are saying it's okay to use my likeness?
15:40
Speaker 1: Is Gen AI subject to certain restrictions?
15:44
Speaker 2: I think you'll see a certain tier, for lack of
15:47
Speaker 2: a better term, we'll say like B celebrities, a large
15:51
Speaker 2: group of them. And you'll have kind of a sprinkling
15:54
Speaker 2: of a few A stars. And then you'll be able
15:56
Speaker 2: to just go to the Gen AI platforms and just
15:58
Speaker 2: type in this person's name instead of it getting immediately rejected.
16:02
Speaker 2: It says, okay, you can do that. But instead of
16:03
Speaker 2: it being $ 0. 30, it's $ 0. 35 online.
16:05
Speaker 1: Uh-huh. So basically you're saying that this platform charges you
16:10
Speaker 1: extra to use a real famous person in the video
16:14
Speaker 1: that you're generating and the real famous person gets some
16:17
Speaker 1: of that money and you get a little tiny bit
16:19
Speaker 1: of that money, presumably.
16:20
Speaker 2: Um, well, I mean, we, we actually don't charge our clients.
16:22
Speaker 2: We charge the gen AI platforms.
16:24
Speaker 1: Uh-huh.
16:25
Speaker 2: So if it's 5 cents, that, that goes to the client.
16:28
Speaker 2: And then we have a different mechanism of how we
16:30
Speaker 2: charge the gen AI companies.
16:32
Speaker 1: And what kind of restrictions to your clients will there
16:35
Speaker 1: be on the use of their likeness in generative videos?
16:38
Speaker 2: There's kind of like a belief system. So you have
16:41
Speaker 2: things like religion. And then more tangibly, it's things around sex, drugs,
16:47
Speaker 2: and alcohol. And then they're all natural language. So you
16:49
Speaker 2: can turn on drugs and say cannabis is just fine.
16:53
Speaker 2: Like I'm okay with this. And there's financial products. Like
16:58
Speaker 2: some people, they're consistently made to make kind of like
17:01
Speaker 2: Bitcoin scams. And so you can have that turned off
17:03
Speaker 2: entirely or say, yes, but it's only for this one
17:06
Speaker 2: app that I'm being sponsored for. And then the last
17:09
Speaker 2: part is literally the sponsorships, right? So you have, uh,
17:13
Speaker 2: you know, Nike is sponsoring me. So that always goes
17:15
Speaker 2: on a, on a sneaker or, you know, I did
17:18
Speaker 2: a deal with hot pockets a few years ago. So
17:20
Speaker 2: you can't ever make some kind of claim of me
17:23
Speaker 2: saying that hot pockets burn my mouth for the next
17:24
Speaker 2: year and a half.
17:25
Speaker 1: Yes. What if I want to make a video of
17:27
Speaker 1: a famous person saying that they love my shampoo? Uh,
17:31
Speaker 1: that I'm selling. Can I do that?
17:34
Speaker 2: No, we do understand the difference between whether it's commercialized
17:37
Speaker 2: or not. And this is where the defense part comes in,
17:40
Speaker 2: which is, let's say that you have them say, happy birthday,
17:43
Speaker 2: and you go start a car wash called the happy
17:45
Speaker 2: birthday car wash. There's no point when you generate it
17:49
Speaker 2: that I would know that that was what the intent was.
17:51
Speaker 2: But when it actually gets used out on Instagram, I
17:54
Speaker 2: will see, oh, this is used in a commercialized way.
17:56
Speaker 2: I'll look up your license and say, oh, that's That
17:59
Speaker 2: wasn't how it was purchased. This was a trickery thing.
18:01
Speaker 2: So then we go ahead and delete that. So they
18:04
Speaker 2: end up having to buy the correct version, but maybe
18:06
Speaker 2: they get away with it for an hour or two.
18:12
Speaker 1: We'll be back in just a minute. Are there any
18:25
Speaker 1: instances where deepfakes are legal? Or the First Amendment might
18:29
Speaker 1: be relevant?
18:31
Speaker 2: So this is tricky. I would say that no fakes
18:36
Speaker 2: is coming down the pipeline, which is a federal act.
18:39
Speaker 2: And that is going to require authorization from the person
18:42
Speaker 2: that you're deep faking. I think this is a really
18:44
Speaker 2: solid place for us to operate from. So between now
18:47
Speaker 2: and when that passes, and it looks like it's going
18:49
Speaker 2: to pass, there's really strong support on both sides of
18:53
Speaker 2: the aisles from tech companies and entertainment. It's kind of
18:56
Speaker 2: a massive collaborative multi-year effort, and it looks like it'll
18:59
Speaker 2: probably pass early next year. It's going to force tech
19:02
Speaker 2: companies to have to ask for authorizations from my neighbor
19:07
Speaker 2: or a celebrity down the street. Everyone is going to
19:09
Speaker 2: need some kind of authorization system for that. And between
19:13
Speaker 2: now and then, you can make a case that it's
19:15
Speaker 2: illegal for right of publicity reasons. If you generate a
19:18
Speaker 2: deepfake endorsing someone, that's like a Lanham Act violation because
19:22
Speaker 2: it's false endorsement. So there's the ways in which it's
19:24
Speaker 2: not about the tech.
19:25
Speaker 1: Well, the endorsement is the easy case. I mean, the
19:27
Speaker 1: one I'm asking you, like, say you do a politician,
19:30
Speaker 1: a satirical video of a politician.
19:33
Speaker 2: Right.
19:33
Speaker 1: Like, yes, we know that satire tends to have more
19:36
Speaker 1: First Amendment protection. We know that discussion of political figures
19:41
Speaker 1: tends to have more First Amendment protection.
19:43
Speaker 2: Yes. Right.
19:43
Speaker 1: Like. I don't know. I can understand why you could
19:47
Speaker 1: make a case that a deepfake is not like a
19:49
Speaker 1: political cartoon, but you could make a case that it is.
19:53
Speaker 2: I think I would probably more make the case that
19:56
Speaker 2: it is. I think there will be protections there. I
20:00
Speaker 2: know that when we talk to both sides of the
20:01
Speaker 2: aisle on who we protect, making sure that we get
20:04
Speaker 2: First Amendment right is a big concern for everyone. And
20:07
Speaker 2: us included.
20:08
Speaker 1: I mean, it's a hard one. It's a hard one because.
20:12
Speaker 1: As a matter of sort of public interest, you want
20:15
Speaker 1: people to understand what is real and what is not real.
20:17
Speaker 2: Yes. Right?
20:18
Speaker 1: Like, I like satire. I'm a reporter. I like the
20:20
Speaker 1: First Amendment. But, like, people being confused about whether the
20:24
Speaker 1: president is really saying something or it's a deepfake of
20:27
Speaker 1: the president, that seems quite bad.
20:29
Speaker 2: I do think that it's going to be. So when
20:31
Speaker 2: you did a political cartoon before, it was in the
20:33
Speaker 2: political cartoon section. of the newspaper or The Economist or whatever.
20:37
Speaker 1: And it's a drawing. Like, no one is going to
20:39
Speaker 1: be confused. Yeah.
20:40
Speaker 2: Well, I think that there'll just be kind of a
20:42
Speaker 2: time and place in the internet for that, which is
20:44
Speaker 2: there'll either be a certain part of the site or
20:47
Speaker 2: it will be labeled clearly on the video. Right? There'll
20:49
Speaker 2: be a restriction.
20:50
Speaker 1: Yeah, like a label. A label is good.
20:52
Speaker 2: If you labeled it, it would be just fine.
20:54
Speaker 1: Is that functional? Like, will that work?
20:57
Speaker 2: Well, this is This is where, you know, and not
21:00
Speaker 2: to harp on us or to focus on what we do,
21:03
Speaker 2: but anyone really that wants to build a business like ours,
21:06
Speaker 2: what you would say is what the time it goes
21:08
Speaker 2: to generate it, you'd say, oh, I see that you're
21:10
Speaker 2: actually making a satirical thing. That doesn't cost you anything
21:12
Speaker 2: because that's your first amendment, right? So we're not going
21:14
Speaker 2: to charge you. You get a bypass on it. But
21:16
Speaker 2: then when we see it out in the wild, if
21:17
Speaker 2: you don't have labeling on it, or if you're not
21:20
Speaker 2: using it in a satirical way, you're using it to like, say,
21:22
Speaker 2: look at what the president said, right? You're trying to
21:25
Speaker 2: like make a stance on it and trying to, you know,
21:27
Speaker 2: push it off as being real. then we'll do a
21:29
Speaker 2: takedown on it. So there's a way that you kind
21:31
Speaker 2: of need both systems to work well.
21:33
Speaker 1: That makes sense. You mentioned the federal legislation. I've heard
21:38
Speaker 1: you talk elsewhere about state laws as well. Is the
21:41
Speaker 1: federal legislation going to sort of preempt or render irrelevant
21:45
Speaker 1: the state laws? Are there state laws that are important
21:47
Speaker 1: that are doing other things that are worth talking about?
21:50
Speaker 1: I mean, I seem to recall California being kind of
21:52
Speaker 1: at the vanguard of name, image, and likeness stuff in
21:55
Speaker 1: the 20th century because of Hollywood, basically, right?
21:58
Speaker 2: Yeah. And then, oh, Tennessee is right now with the
22:00
Speaker 2: Elvis Act.
22:01
Speaker 1: Huh.
22:02
Speaker 2: They've made the No Fakes Act essentially for the state.
22:05
Speaker 1: I see.
22:05
Speaker 2: So if I have a client in Nashville, I have
22:08
Speaker 2: a lot of power for them to go and get
22:11
Speaker 2: things taken down. If someone made just a benign someone
22:14
Speaker 2: walking down the street in AI, if they live in Tennessee,
22:17
Speaker 2: it gets taken down. If they don't, I need to
22:20
Speaker 2: make a pretty good case about why this harms that
22:22
Speaker 2: person and not use AI rules to do it.
22:25
Speaker 1: Oh, interesting. So just to be clear- Outside of Tennessee,
22:29
Speaker 1: which is to say in most of the country at
22:31
Speaker 1: this point, if I make a deep fake of whatever,
22:34
Speaker 1: George Clooney, and he's just like, whatever, reciting Shakespeare or something. And, well,
22:42
Speaker 1: is that okay?
22:45
Speaker 2: Yeah, I mean, it would be allowed. Like, we wouldn't
22:47
Speaker 2: be able to necessarily take that down.
22:49
Speaker 1: If I sell ads against it, if they're just like
22:52
Speaker 1: programmatic ads that show up in the middle of it,
22:54
Speaker 1: now can you take it down?
22:55
Speaker 2: That's right. It's a violation of right of publicity.
22:57
Speaker 1: Because I'm making money off it.
22:59
Speaker 2: Right, off of their likeness.
23:00
Speaker 1: But if this bill becomes law, then it won't be
23:03
Speaker 1: okay anymore.
23:04
Speaker 2: Right. Those people will be able to pick what gets
23:05
Speaker 2: made of them. With a 1A carve-out, right?
23:09
Speaker 1: The First Amendment carve-out, potentially.
23:10
Speaker 2: Yeah.
23:11
Speaker 1: It's interesting that it's okay to do it now. I
23:13
Speaker 1: think that is not intuitive. Yeah.
23:16
Speaker 2: Well, I don't want to encourage people to do it. Um,
23:18
Speaker 2: you literally, like these people are upset that you were
23:20
Speaker 2: stealing part of their private nature, right? Because the people
23:24
Speaker 2: are also doing it for people that aren't celebrities. Um,
23:26
Speaker 2: you know, doing it, there's deep fakes made of, uh,
23:29
Speaker 2: teachers saying crazy things and then them losing their jobs.
23:32
Speaker 2: Like there's still harm, right? That comes from it. So
23:35
Speaker 2: I don't want to really tell people go on the
23:37
Speaker 2: boom rush now and go make it while it's legal. Um,
23:40
Speaker 2: you still are creating harm, even if it may not
23:42
Speaker 2: be a law.
23:43
Speaker 1: Is there a universe where the big platforms just do
23:46
Speaker 1: this themselves? Like, why when I sign up for whatever, Instagram,
23:50
Speaker 1: can I not just say, no deep fakes of me?
23:53
Speaker 1: Why do we need an extra company?
23:55
Speaker 2: So every platform, most platforms, and we're good partners with them,
23:58
Speaker 2: so I don't want to be disparaging about this, but
24:00
Speaker 2: they've tried over the last few years to do this,
24:02
Speaker 2: and there's issues with doing it. One is, it's hard
24:06
Speaker 2: for them to set up similar rules across all systems. Two,
24:10
Speaker 2: it's hard for anyone to believe that they will properly
24:12
Speaker 2: gatekeep themselves.
24:14
Speaker 1: Because the platforms want more content. They're not incentivized to
24:17
Speaker 1: take things down. Right.
24:18
Speaker 2: And so you're like, hey, you as the gatekeeper, can
24:20
Speaker 2: you generate things? And then if you're a celebrity, it's
24:22
Speaker 2: even more complicated because you're saying a biometric release is
24:25
Speaker 2: needed for that. And most humans do not want to
24:28
Speaker 2: have Google give a full biometric release because they know
24:30
Speaker 2: everything about you, right? You also don't need to know
24:32
Speaker 2: what you sound like and look like. And these large
24:35
Speaker 2: tech companies have been sued for face recognition usage in
24:37
Speaker 2: the past because of how they have this weird interstitial
24:41
Speaker 2: data set of like everything you've ever gone to online
24:43
Speaker 2: and what you look like in real life. Places like
24:45
Speaker 2: Illinois and Texas have sued them. So platforms aren't in
24:48
Speaker 2: a great place to say, I'll be the person that
24:51
Speaker 2: steps up and does this because the technology is largely
24:54
Speaker 2: handicapped for them because of this past litigation. And then
24:57
Speaker 2: the other part is people don't want to trust them
24:59
Speaker 2: with the biometrics. Those releases are pretty insane to think about.
25:03
Speaker 2: For us, like this is our only job, right? It's
25:05
Speaker 2: easy to give Lodi those biometric releases and Because all
25:08
Speaker 2: we're doing is protecting your biometrics.
25:10
Speaker 1: I've heard you talk about user agreements as weirdly important
25:16
Speaker 1: in this broader conversation. Why? What's going on there?
25:20
Speaker 2: There's a land grab for AI rights right now, for
25:22
Speaker 2: likeness rights. So if you're a tech company, it doesn't matter.
25:26
Speaker 2: It's not even the big ones, even small startups. They're
25:28
Speaker 2: trying to figure out how to get as many people
25:30
Speaker 2: as possible to do releases on their data, their likeness,
25:33
Speaker 2: their biometrics, because that data is really valuable.
25:36
Speaker 1: And like if I'm just a normal guy with an iPhone,
25:39
Speaker 1: like when I do unlock my phone with my face,
25:42
Speaker 1: am I participating in this without knowing it?
25:45
Speaker 2: No, but actually Apple, if you use an iPhone, they
25:48
Speaker 2: are aware of the same kind of problem sets that
25:50
Speaker 2: we are, which is they make sure that that's all
25:51
Speaker 2: local to the phone.
25:52
Speaker 1: Oh, interesting. So like Apple doesn't know what my face
25:55
Speaker 1: looks like. It's just my phone that knows what my
25:57
Speaker 1: face looks like. And that's good.
25:59
Speaker 2: There's a world that's coming in which all that's going
26:01
Speaker 2: to get legislated.
26:02
Speaker 1: I see.
26:03
Speaker 2: I think that's the next thing after no fakes. Now
26:05
Speaker 2: we're all going to decide what constitutes the data that
26:08
Speaker 2: goes into making a fake, right? I think it's coming
26:11
Speaker 2: and it's probably going to hurt a lot of smaller companies,
26:14
Speaker 2: but the bigger ones now are starting to see that
26:16
Speaker 2: is what's on the horizon and they might as well
26:18
Speaker 2: just design their program accordingly.
26:20
Speaker 1: Yeah, I mean, often regulation... ends up being a moat
26:23
Speaker 1: for big companies, right? It is essentially a barrier to entry.
26:26
Speaker 1: It makes it more expensive to do things. And so ultimately,
26:30
Speaker 1: it's good for the big incumbents.
26:32
Speaker 2: That's absolutely true. You have that plus regulatory capture. And
26:36
Speaker 2: it's definitely something that we'll see in the AI industry soon.
26:40
Speaker 1: Interesting. Let's talk more about the future. We've talked about legislation.
26:47
Speaker 1: We've talked about technology. We've talked about rights. But like,
26:52
Speaker 1: When you think, uh, you know, about whatever amount of
26:58
Speaker 1: time it makes sense to think about five years out,
27:00
Speaker 1: I don't know, 10 years, who knows what's going to
27:01
Speaker 1: go on in 10 years, right? It's so crazy. It's
27:04
Speaker 1: so fast. Um, first of all, what do you worry about?
27:10
Speaker 2: Man. Um, I worry about us being stuck in 2024
27:17
Speaker 2: art and culture for a long time because of how
27:20
Speaker 2: machine learning is trained that, uh, You know, you basically
27:23
Speaker 2: have the corpus of human creativity that built models a
27:26
Speaker 2: couple of years ago that we all started using. And
27:28
Speaker 2: then everything that comes out of those models will be
27:31
Speaker 2: like what the past thousands of years of creative look like.
27:35
Speaker 2: And what we won't have is something that kind of
27:36
Speaker 2: jumps out. A lot of machine learning is based on
27:39
Speaker 2: training on the top of the bell curve of items,
27:42
Speaker 2: like what looks like the most, like 50th to 75th percentile.
27:45
Speaker 1: Sort of the average, the average. Yeah.
27:48
Speaker 2: So you have the average of 2024 that's coming out
27:50
Speaker 2: of models now. And you have to have something that
27:52
Speaker 2: looks truly unique. And AI doesn't want to make that.
27:55
Speaker 2: It doesn't know how to. And that's where the crap
27:57
Speaker 2: lives on one side. And that's where like the really
27:59
Speaker 2: amazing stuff lives on the other.
28:02
Speaker 1: Of the bell curve. You're talking about sort of the
28:03
Speaker 1: tails of the distribution, things that are really bad and
28:06
Speaker 1: things that are really good.
28:07
Speaker 2: Correct. Right. So you have like this mathematical preventative thing
28:11
Speaker 2: that says, hey, I don't know how to do math
28:13
Speaker 2: that allows the really cool stuff in without letting the
28:16
Speaker 2: bad stuff in. So I'm just going to go ahead
28:18
Speaker 2: and say all this big stuff in the middle is
28:20
Speaker 2: what I can produce. So the fear I have is
28:22
Speaker 2: that we've kind of walked down this path not really
28:25
Speaker 2: understanding that this is a feature of the mathematics. It's
28:28
Speaker 2: not like someone at OpenAI is making the decision. It's
28:31
Speaker 2: that they are using math that is forcing this decision.
28:34
Speaker 2: And then it's in 10 or 20 years and we're
28:36
Speaker 2: basically generating the same kind of content because nothing can
28:40
Speaker 2: really like pop out without getting hammered down by either
28:43
Speaker 2: algorithms or or by the generative models themselves.
28:45
Speaker 1: Well, if that's true, there's a weird, happy, a silver
28:50
Speaker 1: lining to that story, which is that's an opportunity for
28:53
Speaker 1: talented humans, right? If it's true that AI can only
28:57
Speaker 1: generate kind of mid-creative work... That's good for people who
29:03
Speaker 1: are weirdos and make things that are maybe really bad
29:07
Speaker 1: or maybe really great or a few people love even
29:09
Speaker 1: though most people don't. If that's the case, that's actually
29:12
Speaker 1: a happier outcome than I'm worried about.
29:14
Speaker 2: And I have a 15-year-old, a 13-year-old, and a 9-year-old,
29:17
Speaker 2: and that's what I tell them, is that the world
29:19
Speaker 2: is going to get really average quickly, which is going
29:22
Speaker 2: to be good for you because you're all very weird.
29:25
Speaker 2: You kind of have to figure out what your voice is.
29:28
Speaker 2: And in a world that wants to collapse you down,
29:30
Speaker 2: that wants to hammer the nail that sticks out, just
29:32
Speaker 2: know that your competitive advantage is that you won't.
29:35
Speaker 1: I mean, I feel like there's a universe where you
29:39
Speaker 1: can play with the AI and get it to make
29:42
Speaker 1: stuff out in the tails, right? You can use AI
29:45
Speaker 1: to make weirder stuff by kind of iterating, right? And
29:51
Speaker 1: to what extent you need a person doing that, you know,
29:54
Speaker 1: like to what extent it's just one shot where it's
29:56
Speaker 1: like make something really weird and different in this way
29:59
Speaker 1: versus to what extent it's like, well, I write something
30:02
Speaker 1: and then I ask the AI about it and it
30:03
Speaker 1: gives me some ideas and I write some more. is
30:06
Speaker 1: to me kind of the open question and a meaningful
30:08
Speaker 1: open question because the less you need of me, maybe
30:12
Speaker 1: the worse at some level.
30:14
Speaker 2: So this is where I get optimistic about AI is
30:16
Speaker 2: because I do think that humans can push AI to
30:19
Speaker 2: make something extraordinary, but it's the difference between being the
30:22
Speaker 2: painter or the paintbrush, right? Like if AI is the paintbrush,
30:25
Speaker 2: you're good, right? But as long as you, as soon
30:28
Speaker 2: as you say, well, actually AI write a script for
30:30
Speaker 2: me or paint the painting for me, Um, and it
30:33
Speaker 2: becomes more of the painter. You kind of are screwed.
30:36
Speaker 2: You're in that spot of like, it's going to make
30:37
Speaker 2: something that's average, but you're right. If you're a human
30:39
Speaker 2: that's involved in the process and you kind of push
30:42
Speaker 2: it beyond its boundaries, you are then taking it out
30:45
Speaker 2: of like the architect role and more into the, like
30:47
Speaker 2: the engineer, like just do this thing. I know it
30:49
Speaker 2: looks weird or sounds weird, but do it anyway. Um,
30:52
Speaker 2: that's like taste, right? That that's human creativity getting imbued
30:55
Speaker 2: into the AI product are very bullish on that. I
30:58
Speaker 2: think that that's where AI will actually help our species.
31:02
Speaker 2: I just don't like the idea that some people will say, well,
31:04
Speaker 2: they'll be writing movies soon. Like, hopefully not. Or they'll
31:07
Speaker 2: be very mid. Or they'll be writing music soon. Hopefully
31:10
Speaker 2: a human that wants to make a movie or wants
31:12
Speaker 2: to make a song says, like, there's something that is
31:15
Speaker 2: so out of this world I have to use AI
31:17
Speaker 2: for that clip of it. That's cool. I'm all for that.
31:20
Speaker 1: Yeah, I mean, I hope so. You used the word
31:22
Speaker 1: taste in there, which is a word people have been
31:24
Speaker 1: using a lot, right? People have taste. I don't know.
31:29
Speaker 1: That's like an interesting big... conversation about the future of
31:34
Speaker 1: sort of human creativity and AI. If we think more
31:37
Speaker 1: narrowly about the business that you are in, you know,
31:41
Speaker 1: people's own likeness in this world of exploding AI, what
31:48
Speaker 1: do you worry about in the, you know, next few years?
31:50
Speaker 2: I would say the problems I want to solve for
31:53
Speaker 2: are making sure that there's very low friction so people
31:57
Speaker 2: can control their likeness, whether you are a celebrity or
31:59
Speaker 2: you're like me and you're just some average guy walking
32:02
Speaker 2: around I think that that's a difficult problem to solve. Like,
32:05
Speaker 2: how do you do that at scale? Which friend of
32:06
Speaker 2: mine is allowed to generate me as well as a
32:08
Speaker 2: weird problem?
32:09
Speaker 1: I'm going to just say none.
32:11
Speaker 2: Yes. You might say none or, or what you might
32:14
Speaker 2: actually find out is, um, you have a parent that
32:17
Speaker 2: is ill and they're like, actually, um, I do want
32:21
Speaker 2: to make sure that even after I'm gone, I'm reading
32:23
Speaker 2: books to your kids. Right. And then you're like, okay,
32:27
Speaker 2: that's an exception.
32:27
Speaker 1: Right.
32:28
Speaker 2: And then you'll end up finding these kinds of outliers
32:29
Speaker 2: where you're like, can my partner make something? They had
32:31
Speaker 2: a dream of me and they want to like show
32:33
Speaker 2: me the exact dream they had of me. And it's
32:35
Speaker 2: going to be cathartic for them. Yeah, sure. I don't
32:37
Speaker 2: mind if my partner, I trust them with everything. Might
32:39
Speaker 2: as well trust them with that. I think there'll be
32:41
Speaker 2: slivers of things. You're going to have like high school
32:43
Speaker 2: kids that'll be able to give their likeness because their
32:45
Speaker 2: friends always love to create crazy stuff.
32:47
Speaker 1: Right.
32:48
Speaker 2: Um, And you need good guardrails around that too. And
32:51
Speaker 2: so I think there'll be pockets of things that people
32:54
Speaker 2: that even traditionally would be staunch, be like, I'm not
32:56
Speaker 2: going to do anything with AI, will say, well, actually,
33:00
Speaker 2: I think that that's cool. And then it'll just grow
33:02
Speaker 2: from there.
33:05
Speaker 1: You do seem optimistic. So give me your happy vision
33:08
Speaker 1: of the world five or 10 years from now.
33:10
Speaker 2: The happy vision is one in which we don't ever
33:15
Speaker 2: use AI as the painter. It's humans that have figured
33:18
Speaker 2: out how to push the boundary of creativity with this
33:21
Speaker 2: new kind of paintbrush. That's the first part. And then
33:23
Speaker 2: the second part is people don't see the internet as
33:27
Speaker 2: this big wild west of anything can get uploaded of
33:29
Speaker 2: you whenever, wherever, and can destroy your life overnight. And
33:33
Speaker 2: so I think both of these things, like the privacy
33:35
Speaker 2: component and the control and autonomy and agency, hopefully will
33:38
Speaker 2: come to the internet. And then the second part, of course,
33:40
Speaker 2: being creativity, getting solved for, but solved in my mind
33:45
Speaker 2: is more paintbrush and less painter.
33:50
Speaker 1: We'll be back in a minute with the lightning round. Okay,
34:03
Speaker 1: let's do a lightning round. What's your favorite deepfake?
34:10
Speaker 2: The bunny rabbits on the trampoline. Totally harmless. It got
34:14
Speaker 2: a million people laughing and enjoying and seeing that AI
34:16
Speaker 2: could make some funny, cool stuff and tricked a lot
34:19
Speaker 2: of people into thinking they were real as well. So
34:21
Speaker 2: I actually did like that.
34:23
Speaker 1: I'll have to check it out. I haven't seen it.
34:26
Speaker 1: What celebrity likeness would you license to advertise for Lodi,
34:31
Speaker 1: to be a spokesman for Lodi?
34:33
Speaker 2: Cannot say. Yeah. It's coming out actually in the next
34:36
Speaker 2: few weeks, so cannot say.
34:38
Speaker 1: Okay.
34:38
Speaker 2: Yeah.
34:41
Speaker 1: I'm very interested to see. Is there one or will
34:45
Speaker 1: there be a bunch?
34:46
Speaker 2: There's one that's on the forefront of all of this. Okay.
34:50
Speaker 2: Someone that we deeply respect in this space.
34:54
Speaker 1: I'm very interested. How do you want your kids to
34:57
Speaker 1: use AI?
35:00
Speaker 2: I told my 15-year-old that he could probably cheat on
35:02
Speaker 2: everything with AI. And that there's probably no way that
35:05
Speaker 2: me or a teacher could figure that out. Because I
35:08
Speaker 2: know that kids are really crafty. So that he has
35:10
Speaker 2: to make the active choice that if he wants to
35:12
Speaker 2: be successful, he has to do hard work. And that's
35:14
Speaker 2: a completely different skill set. So it really is coming
35:19
Speaker 2: down to just, especially if you have a smart kid,
35:21
Speaker 2: just tell them, look, I know you can do it.
35:23
Speaker 2: But here, let me make the case for why you
35:25
Speaker 2: should work harder instead. Is if it's in your head,
35:28
Speaker 2: you could then leverage it to do things that other
35:30
Speaker 2: people cannot. So it's a little competitive, but It's essentially
35:33
Speaker 2: to tell my kids, don't shy away from it. Like
35:36
Speaker 2: my 15-year-old's on the $ 200 a month cloud plan. Like
35:39
Speaker 2: he uses AI a ton, right? He's like building apps
35:42
Speaker 2: and stuff. It's crazy. I'm really proud of him. But
35:45
Speaker 2: at the same time, he knows that you can't use
35:47
Speaker 2: it to fill out all your homework without you robbing
35:49
Speaker 2: yourself of something.
35:51
Speaker 1: Yes, right. The point of homework is actually not the whatever,
35:57
Speaker 1: the math you're learning. It's learning to just sit there
35:59
Speaker 1: and do it, right? That's actually the work.
36:02
Speaker 2: Yeah, those are the... So a lot of our discussions
36:03
Speaker 2: around AI are actually the reasons why you wouldn't want
36:06
Speaker 2: to use it in a home that deeply uses it. Yeah.
36:09
Speaker 2: So it's, but I think that that's like the, the
36:12
Speaker 2: brake and gas pedal kind of thing. Like you need
36:14
Speaker 2: both to run the car.
36:15
Speaker 1: Yeah.
36:15
Speaker 2: And so a lot of the conversation is around the
36:17
Speaker 2: more complicated use of a brake system. What?
36:21
Speaker 1: I was trying to think, frankly, of a good question
36:23
Speaker 1: about like AI and the future of movie stars. Like,
36:27
Speaker 1: I don't know the first movie with an AI genre,
36:28
Speaker 1: but I couldn't quite get the right question. Like, make
36:33
Speaker 1: a prediction about AI and the future of movie stars.
36:37
Speaker 2: I think that there's going to be, there's going to
36:40
Speaker 2: be a demand for like authentic performances. That's what they'll
36:44
Speaker 2: be called. Something to that effect, some euphemism for it's
36:47
Speaker 2: really them in front of film. And I think they're
36:49
Speaker 2: going to make movies that portray that a little bit better.
36:52
Speaker 2: Like it'll be like human expressions that AI is not
36:54
Speaker 2: able to get or more complicated scripts that land in
36:57
Speaker 2: a way that's not like uncanny valley. This is a
37:00
Speaker 2: good example. There's the, The Silver Springs, Fleetwood Mac, where
37:05
Speaker 2: the lead singer and the guitarist basically are in this
37:07
Speaker 2: long-term situationship, and they're essentially singing about themselves on stage,
37:12
Speaker 2: and everyone can kind of feel that tension. It's a
37:14
Speaker 2: human thing. And if you were to swap them out
37:16
Speaker 2: for robots, you probably would not feel the same way
37:19
Speaker 2: as a human. I think there's something about knowing that
37:22
Speaker 2: that person suffered in the way that you've suffered that
37:25
Speaker 2: makes you identify with their art that is going to
37:27
Speaker 2: be much bigger, and we talk about Taste, I know,
37:30
Speaker 2: is a cheesy word, but still is something real here.
37:32
Speaker 2: I think taste for that will go up because it'll
37:35
Speaker 2: be more rare. And then I think there's going to
37:37
Speaker 2: be a bunch of superhero action movies that are made
37:40
Speaker 2: in AI and no one's going to care.
37:42
Speaker 1: I mean, superhero action movies are already largely made in AI.
37:46
Speaker 2: Correct. That's it. Yes. Yes.
37:49
Speaker 1: Like, it is interestingly not as binary as as we say, right?
37:55
Speaker 2: If your art is making really generic romance songs for radio,
38:00
Speaker 2: you probably are in for some competition with AI. But
38:03
Speaker 2: if you have a unique sound, unique bass, whatever it
38:07
Speaker 2: is that you bring to the table that is truly
38:09
Speaker 2: not like something else we hear everywhere else, the demand
38:12
Speaker 2: for that's going to go through the roof.
38:14
Speaker 1: That's encouraging. I appreciate your time.
38:18
Speaker 2: Of course.
38:25
Speaker 1: Luke Garagoni is co-founder and CEO of Lodi. Our show
38:30
Speaker 1: is produced by Gabriel Hunter Chang and Trina Menino. Our
38:33
Speaker 1: editor is Lydia Jean Cott, and our engineer is Sarah Bruguier.
38:37
Speaker 1: I'm Jacob Goldstein, and we'll be back next week with
38:39
Speaker 1: another episode of What's Your Problem?