Latest / Giant Ideas / Synthesia Co-founder and COO, Steffen Tjerrild: the Future of Video, Trust, and Deepfakes

Transcript

Synthesia Co-founder and COO, Steffen Tjerrild: the Future of Video, Trust, and Deepfakes · Giant Ideas

• 0:00
  SPEAKER_00: So I think we'll start to have to trust BBC even more than we do today.
• 0:04
  SPEAKER_00: Now they will be able to kind of create hundreds of BBCs down the line, right?
• 0:08
  SPEAKER_00: Because the production, the journalistic things, it's gonna be easier and easier and easier, and the barriers to entry is gonna be lower and lower and lower.
• 0:15
  SPEAKER_01: Hello, and welcome to Giant Ideas with me, Cameron McLean, and me, Tommy Stadlin.
• 0:20
  SPEAKER_01: We're co-founders of Giant Ventures, which builds and backs purpose-driven companies.
• 0:24
  SPEAKER_02: At Giant, we're lucky to meet extraordinary people with giant ideas that are changing the world.
• 0:29
  SPEAKER_02: This podcast brings you behind the scenes access to those ideas and the inspiring stories of the people behind them.
• 0:35
  SPEAKER_01: We explore how one giant idea can kickstart a billion-dollar company, shape culture, and transform life as we know it.
• 0:42
  SPEAKER_02: Today on the podcast, we have Synthesia co-founder and COO Stefan Gerald.
• 0:46
  SPEAKER_02: Synthesia is an AI company that allows you to create professional videos with just a few lines of text.
• 0:52
  SPEAKER_02: It's now used by over 1 million customers, including 80% of Fortune 100 companies.
• 0:57
  SPEAKER_01: Synthesia was recently valued at over$2 billion.
• 1:00
  SPEAKER_01: The heart of this exponential growth is Stefan.
• 1:02
  SPEAKER_01: As the company's COO, Stefan has seen the company through its inception to the latest funding round.
• 1:07
  SPEAKER_01: Today we're going to talk about the magic of creating video from a few lines of text and what that means for human creativity.
• 1:16
  SPEAKER_02: Welcome, Stefan.
• 1:18
  SPEAKER_02: Thanks for having me.
• 1:18
  SPEAKER_02: Thanks so much for being here.
• 1:20
  SPEAKER_02: Stefan, you and I met back in 2018 when you were pitching for your seed round for Synthesia.
• 1:24
  SPEAKER_02: I thought you were awesome, thought you your co-founder, Victor, was awesome, but probably like many other people made the bad mistake of passing.
• 1:32
  SPEAKER_02: Uh at my previous firm, Hummingbird Ventures.
• 1:35
  SPEAKER_02: I thought you two were awesome.
• 1:37
  SPEAKER_02: I could see that you were going to do something special, but candidly, I didn't see the giant idea at that moment in time.
• 1:41
  SPEAKER_02: I think you were pitching more around dubbing for media and Hollywood at that time.
• 1:45
  SPEAKER_02: But fast forward to now, you have built a company that's valued at$2 billion.
• 1:50
  SPEAKER_02: It's worth, it's doing over 100 million of revenue, I believe.
• 1:53
  SPEAKER_02: And you've built a remarkable magical product that essentially turns text into video like magic.
• 1:58
  SPEAKER_02: And you're working with over 80% of the Fortune 100 company.
• 2:01
  SPEAKER_02: So it's been quite a journey.
• 2:03
  SPEAKER_02: So tell us, you know, how you started and I guess give the audience an essence of what this giant idea is.
• 2:08
  SPEAKER_00: Yeah, thanks a lot for having me.
• 2:10
  SPEAKER_00: Good to reconnect, Cameron.
• 2:11
  SPEAKER_00: Yeah, of course.
• 2:12
  SPEAKER_00: It's been a few years, but uh, but yeah, back then it was obviously less clear kind of how big and kind of you know how this kind of category would unfold.
• 2:20
  SPEAKER_00: So I think there's definitely to not to only your fault, but I think the pitch was not as clear, a clear as kind of what the outcome has kind of gone back then.
• 2:28
  SPEAKER_00: But uh, but back then it was kind of originally stem from like a research project.
• 2:32
  SPEAKER_00: So we kind of partnered up with the two professors who kind of for the first time showcased how you can change and update video after the fact.
• 2:39
  SPEAKER_00: So when it comes off camera, and the first kind of you know use cases or the things that kind of came to mind there was around kind of movie dopping, enriching and kind of you know taking content that has been shot with a camera and actually kind of you know translating that with the lip movements.
• 2:53
  SPEAKER_00: But back then it was very much like a very unscalable solution.
• 2:57
  SPEAKER_00: Uh we had to kind of film somebody to then kind of create a new piece of content, and it was kind of you know inherently unscalable and not really kind of you know something that was almost kind of VC backable back then, yeah.
• 3:08
  SPEAKER_00: Um, but actually full circle.
• 3:10
  SPEAKER_00: So we actually launched that product yesterday where you can just one-click drop an existing piece of content with lip sync with all these things that historically kind of required a full kind of five PhDs, kind of changing every frame to kind of create just a few seconds of new videos.
• 3:25
  SPEAKER_00: So it's kind of becoming full circle since then, and it's pretty awesome to kind of know revisit that product and kind of getting that actual to life uh with with this new kind of offering.
• 3:35
  SPEAKER_01: Um so let's talk about the idea itself, the current idea, the big idea, the giant idea.
• 3:40
  SPEAKER_01: Um, text to video.
• 3:42
  SPEAKER_01: Is it as simple as that?
• 3:43
  SPEAKER_01: That is the huge idea.
• 3:46
  SPEAKER_00: Yeah, I think if you if you if you kind of boil it down to to to kind of know this this notion of these kind of the technicalities of kind of typing in text and getting video out, right?
• 3:55
  SPEAKER_00: I think that's kind of you know pretty magical.
• 3:58
  SPEAKER_00: But I think what is also kind of allowing companies and individual creators, individual storytellers is also kind of really democratizing kind of movie creation and storytelling per se, right?
• 4:08
  SPEAKER_00: I think if we parallel it out to other creative disciplines such as music creation, we're sitting in a in a music studio now, right?
• 4:15
  SPEAKER_00: I think 10-15 years ago you needed to go to a Hollywood recording studio to create a top chatting hit.
• 4:21
  SPEAKER_00: Today you can synthesize any tune in the universe on your laptop.
• 4:24
  SPEAKER_00: And we think that the same kind of you know creative disruption or creative kind of you know democratization is gonna happen to video, especially with these magical, amazing kind of video models that's coming out.
• 4:35
  SPEAKER_00: I think we're sitting right next to DeepMind who kind of came out with the amazing view free.
• 4:39
  SPEAKER_00: And it's getting closer to this vision.
• 4:40
  SPEAKER_00: When we started the company, right, we had this vision that you can create a Hollywood-style movie on your laptop, and that is actually starting to really kind of come come to fruition now.
• 4:49
  SPEAKER_00: We said back then 10 years, so it's actually kind of you know pretty close now, uh, which is uh which is pretty kind of you know amazing to see that everybody in the in the world with a good story to actually kind of know express that with the techniques that has kind of been been a kind of monopoly in Hollywood with hundreds of millions of dollars for actors and movie studios and so forth.
• 5:08
  SPEAKER_00: I think it's gonna, you know, change what humans are not that good at in historic kind of terms, right?
• 5:14
  SPEAKER_00: With repeatable processes, find this very rudimentary and very kind of repeatable to something that is kind of more of creative nature, where humans are just uh kind of, you know, much, much better at doing some of those uh tasks, right?
• 5:26
  SPEAKER_00: So I think there is no doubt that I think by allowing more people to kind of get into the field and allowing them to have the techniques to kind of do these type of things, I think will be a positive force for sure.
• 5:36
  SPEAKER_00: But I think it will have some kind of replacement of of jobs in in the long run for sure.
• 5:41
  SPEAKER_00: But I think that's very general for a lot of AI today.
• 5:43
  SPEAKER_01: Okay.
• 5:44
  SPEAKER_01: Tell us a little bit about where creativity will thrive then in the world of synthesia, where it is about storytelling.
• 5:50
  SPEAKER_01: So, kind of, you know, why would Cameron's storytelling prompt for synthesia be better than mine?
• 5:55
  SPEAKER_01: What is it about how he's gonna write that that's gonna be, you know, more impressive than mine?
• 5:59
  SPEAKER_01: Is it great writers will kind of thrive and and you know create create amazing videos?
• 6:04
  SPEAKER_01: Will it be more the concept?
• 6:05
  SPEAKER_01: Will it be quite technical things about how exactly you prompt the level of detail?
• 6:09
  SPEAKER_01: Who wins in that world?
• 6:11
  SPEAKER_00: Uh I think it's people who who are great storytellers in general have like a lot of like creativity, have a lot of like good things around how they kind of tell stories, right?
• 6:20
  SPEAKER_00: Uh, but I think it's it's gonna be those people who's working with these systems, there's also gonna be bouncing off ideas with these systems a lot, right?
• 6:28
  SPEAKER_00: So I think it's important that you kind of continue to to kind of work with these systems and really kind of you know are trying to feedback it off it because that's kind of where you can kind of really put these technologies to the edge of where it's kind of you know good enough or where is it not good enough for today.
• 6:43
  SPEAKER_00: Uh, but I think it will be very different from where you are and what is the audience.
• 6:47
  SPEAKER_00: And I think also like storytelling will fragment a lot.
• 6:50
  SPEAKER_00: I think we'll kind of go from kind of very broadcast where we all kind of consume the same media in the in the cinemas to much more individualized feeds as we do on our phones with TikTok, where we have all unique feeds, right?
• 7:02
  SPEAKER_00: We are all consuming very unique information, which is based on our interest, our kind of way of wanting to kind of you know consume information, right?
• 7:09
  SPEAKER_00: And I think this will also just allow kind of storytellers to kind of you know have much more individualized and much more kind of fragmented kind of stories to distribute.
• 7:18
  SPEAKER_02: How close are we to creating video that is indistinguishable from video that was created by humans or or you're already there?
• 7:26
  SPEAKER_00: Uh I think we're getting there.
• 7:28
  SPEAKER_00: I think both with kind of the avatar technology we have, I think we're really kind of starting to get there.
• 7:33
  SPEAKER_00: We're launching Express 2, which is our latest model, which is like full hand gestures that is in sync with what you're saying.
• 7:40
  SPEAKER_00: So some of the earlier versions of the technology was like waist up where it doesn't have hands, and now they're kind of full body.
• 7:46
  SPEAKER_00: They can kind of you know gestulate in the language and kind of know the they kind of know the body language that is kind of tied to what you're saying, right?
• 7:53
  SPEAKER_00: And then also with these magical kind of foundational models, which is you've seen all over the internet, right?
• 7:59
  SPEAKER_00: I think it's also just getting increasingly good enough for a lot of these things.
• 8:03
  SPEAKER_00: They're still kind of hard to manage and hard to kind of you know uh kind of tame.
• 8:07
  SPEAKER_00: Yeah, and they're also kind of pretty short form.
• 8:10
  SPEAKER_00: But I think give this a few years, right?
• 8:12
  SPEAKER_00: I think we'll get to to a level where it's a good thing.
• 8:14
  SPEAKER_00: So you think it's a few years?
• 8:15
  SPEAKER_02: You think a few years till we hit the uncanny valley when it's indistinguishable?
• 8:19
  SPEAKER_00: It depends on kind of what type of content.
• 8:20
  SPEAKER_00: If it's like corporate content where we do, I think we are there.
• 8:23
  SPEAKER_00: Okay.
• 8:23
  SPEAKER_00: But if it's like uh things you watch in the cinema, there will still be a few years.
• 8:27
  SPEAKER_02: And when these when this happens and these lines blur, how do you think the world will change?
• 8:31
  SPEAKER_02: And and you guys are at the real forefront.
• 8:33
  SPEAKER_02: How will we deal with deep fakes?
• 8:36
  SPEAKER_00: I think for we have like had the system since we started the company, since we met you back in back in the early days, right?
• 8:42
  SPEAKER_00: This framework around the free C's, so consent, collaboration, and control, uh, which means that you can't just upload anybody to our system and make like a video of them saying something they have ever said.
• 8:54
  SPEAKER_00: So we have like pretty strong consent systems.
• 8:56
  SPEAKER_00: So every avatar, every person on the platform has given explicit consent to kind of you know giving their likeness and digitizing their likeness.
• 9:03
  SPEAKER_00: Uh, we also have really strong kind of content moderation, which means that you can't just type anything out in the in the platform.
• 9:11
  SPEAKER_00: So we are pretty strict on what you can actually produce.
• 9:13
  SPEAKER_00: And if you think about content moderation on the internet, that's typically only moderated on the point of distribution.
• 9:20
  SPEAKER_00: So that's kind of when you actually upload it to Facebook or TikTok or whatever.
• 9:24
  SPEAKER_00: That's kind of the point of moderation where we actually moderate it on the creation, right?
• 9:29
  SPEAKER_00: That's kind of in the Word document, it's kind of being moderated.
• 9:31
  SPEAKER_00: You can't even type out hate speech and these type of things.
• 9:33
  SPEAKER_00: That's kind of how how strict we are with trying to kind of enforce this on the level of creation.
• 9:39
  SPEAKER_00: Um, and then we also have like, you know, collaboration, right?
• 9:42
  SPEAKER_00: I think one thing is our system not being abused and kind of you know misused for some of these things.
• 9:47
  SPEAKER_00: But it's obviously as these technologies democratize, it become easier and easier to kind of produce these things with open source and so forth.
• 9:54
  SPEAKER_00: It's obviously important that we think about this more on an internet-wide scale.
• 9:58
  SPEAKER_00: So, how do we kind of make sure that we believe what we see and kind of have the trust in the internet and the information that we consume?
• 10:04
  SPEAKER_02: Do you have a second giant idea for solving that problem?
• 10:06
  SPEAKER_00: I think it'll be more about it'll be less about has this been manipulated or changed by AI?
• 10:12
  SPEAKER_00: It will be more is this the original meaning with the piece of content?
• 10:15
  SPEAKER_00: So it's the media provenance.
• 10:17
  SPEAKER_00: Okay.
• 10:17
  SPEAKER_00: Where if you think about all photos on the internet, right?
• 10:20
  SPEAKER_00: I think 99% of them have been through photo filter or Photoshop.
• 10:23
  SPEAKER_00: And even when you take a picture on your iPhone today, it actually does changing the light systems and the kind of manipulation.
• 10:30
  SPEAKER_00: Is that AI too?
• 10:31
  SPEAKER_02: Okay.
• 10:31
  SPEAKER_00: It's kind of, you know, it's less about has it been manipulated or not.
• 10:34
  SPEAKER_00: It's more is this the original meaning?
• 10:36
  SPEAKER_00: And if that provenance changes, right?
• 10:38
  SPEAKER_00: So if I take a news clip from BBC and kind of changing the anchors, kind of motions on that, then the provenance breaks.
• 10:47
  SPEAKER_00: And when I upload that to YouTube, it'll kind of use a shasham style understanding.
• 10:52
  SPEAKER_00: Has this clip been uploaded before?
• 10:53
  SPEAKER_00: And has the original meaning been altered?
• 10:56
  SPEAKER_00: I think that's kind of you know how we think about how this will be solved at scale.
• 10:59
  SPEAKER_01: There's this great BBC drama about deepfakes and and the sort of real problems in that start when you get live deepfakes, right?
• 11:06
  SPEAKER_01: So this person goes on BBC News and and does a live interview, and the the politician, the real politician is talking, giving one message and then is going out broadcast live, a slightly different, well, very different message, right?
• 11:18
  SPEAKER_01: Um, well, how close are we to live deepfakes or it doesn't have to be deepfakes, but it's like I guess live synthesia.
• 11:24
  SPEAKER_00: Yeah, I think we we're working on these capabilities now.
• 11:26
  SPEAKER_00: I think that's also kind of a big picture where we think this is moving from this kind of one-directional broadcast into much more kind of conversational pieces, right?
• 11:35
  SPEAKER_00: You're also kind of talking to AI today with your Grok or with your uh Open AI or whatever you're using, right?
• 11:41
  SPEAKER_00: You're kind of you know speaking and using the voice mode.
• 11:43
  SPEAKER_00: And the natural next step is also kind of having a face to some of those capabilities, right?
• 11:48
  SPEAKER_00: So I think we are we're getting closer to kind of that real-time world.
• 11:52
  SPEAKER_00: Um, and I think kind of the security and the provenance obviously will kind of have to kind of know move with it.
• 11:58
  SPEAKER_00: And uh, I'm pretty sure I feel pretty good about how we're kind of thinking and tackling these type of things today.
• 12:03
  SPEAKER_01: Do you feel good about it?
• 12:04
  SPEAKER_01: I mean, I'm so obviously this is massively positive, right?
• 12:06
  SPEAKER_01: You you talked about the uh creativity you're gonna unleash for the world with your your technology, and so clearly there is so much good there, but it it is worrying, right?
• 12:15
  SPEAKER_01: The the the world of deep fakes and what what effect that can have on democracy.
• 12:19
  SPEAKER_01: What part of you sort of lies awake at night and thinking, I you know, am I gonna be responsible for something kind of horrific or or in the world?
• 12:27
  SPEAKER_01: Do you worry about that at all?
• 12:28
  SPEAKER_01: You've got we talked about it, you've got a young young child.
• 12:30
  SPEAKER_00: Yeah, I I I actually don't worry about it.
• 12:33
  SPEAKER_00: I think we've kind of gone a long way, and I think some of the things we've seen of how I think we all talk about technology will disrupt a lot of things and AI will have a huge kind of impact on our society, and we all agree.
• 12:44
  SPEAKER_00: I think it'll actually strengthen our institutions.
• 12:47
  SPEAKER_00: I think we'll start to have to trust BBC even more than we do today because now they will be able to kind of create hundreds of BBCs down the line, right?
• 12:55
  SPEAKER_00: Because the production, the journalistic things, it's gonna be easier and easier and easier, and the barriers to entry is gonna be lower and lower and lower.
• 13:02
  SPEAKER_00: And that's kind of why we'll kind of gravitate to some of these institutions that we're really kind of trusting today.
• 13:07
  SPEAKER_00: And I think being able to kind of you know thinking and changing our worldview from default trusting to default being skeptical, unless it's kind of coming from critical sources.
• 13:18
  SPEAKER_00: I think that's kind of where in the world we are we're moving to, right?
• 13:21
  SPEAKER_00: And I think this kind of provenance thing is gonna be something that has to be rolled out on an internet-wide scale to to kind of bring that kind of security and that kind of trust back.
• 13:30
  SPEAKER_02: What's the craziest possible future we could imagine with with digital avatars?
• 13:35
  SPEAKER_00: Uh, I think uh I think there's there's obviously kind of the things that we're working on now is kind of this more conversational piece, right?
• 13:43
  SPEAKER_00: Where you can think about a lot of your business operations are actually being automated with kind of conversations.
• 13:48
  SPEAKER_00: There's a lot of like structured calls and businesses that is very repeatable, that has very low value, can kind of be onboarding calls, it can kind of be cybersecurity training, it can keep be customer education, some of these kind of you know, mechanics that is requiring a lot of like very rudimentary kind of human power, but it's valuable enough to kind of put a human in front of other humans to make sure that they can understand certain concepts.
• 14:12
  SPEAKER_00: I think that's kind of the world where we're moving to, where it's less about creating MP4 files, fees cheaper and faster, but it's actually delivering full understanding, right?
• 14:22
  SPEAKER_00: Where we're actually certifying that you've understood the information by being quized, by being kind of, you know, making sure that you understand the materials, right?
• 14:30
  SPEAKER_00: And the thing that's kind of going from these passive communication flows that are happening every day, where people are zoning out, they're watching things on 2X speed, they're clicking through security awareness training uh without like understanding anything, right?
• 14:44
  SPEAKER_00: And I think that's kind of where the world where you're actually certifying people who have understood it before you can kind of complete.
• 14:50
  SPEAKER_02: So it's sort of like the LLMs, the intelligence of open AI matched with, you know, digital avatars that feel very real and kind of that interaction back and forth.
• 14:58
  SPEAKER_01: I want to go back to that idea of this personalized feed.
• 15:01
  SPEAKER_01: So it's one thing for a producer to serve someone else or you know, a consumer, a video which they made with AI.
• 15:07
  SPEAKER_01: But this idea potentially of us all as consumers being able to create our own entertainment, essentially, right?
• 15:13
  SPEAKER_01: Withstead of scrolling mindlessly through Netflix and not finding something for 20 minutes that you want to watch.
• 15:17
  SPEAKER_01: You can just go, I want to watch a drama about, you know, two cops in London who are running around King's Cross trying to solve a murder.
• 15:24
  SPEAKER_01: Like, how how close are we to consumers self-serving and self-creating the content that they want to watch, which is Netflix level quality?
• 15:31
  SPEAKER_00: Yeah, I think it'll still be a few years before we kind of get to that level.
• 15:34
  SPEAKER_00: But will we get there?
• 15:35
  SPEAKER_00: Yeah, we'll 100% get there.
• 15:37
  SPEAKER_01: Yeah.
• 15:37
  SPEAKER_00: I think to get that with a cost and kind of inference cost, I think it'll be maybe five years where you can kind of really produce Netflix style content let us just continuous, where you can make 10 more seasons of uh whatever kind of, you know, pick your favorite show, right?
• 15:52
  SPEAKER_00: Um, but I think it's interesting, right?
• 15:54
  SPEAKER_00: Where we kind of, you know, going from this one size fits all, which is kind of the movie theater, right, to a much more personalized, contextualized, localized version, right?
• 16:03
  SPEAKER_00: Where every kind of piece of content can kind of be very unique and it can also be unique to a situation, right?
• 16:09
  SPEAKER_00: So if you are driving your car, you probably want to hear the information as a podcast.
• 16:13
  SPEAKER_00: If you're sitting home at nine, you probably want to like Netflix style movie.
• 16:17
  SPEAKER_00: If you're at work, it could be like, you know, a condensed kind of document with all the kind of you know the things, right?
• 16:23
  SPEAKER_00: So I think you'll have this ability to kind of you know transform the information to be consumed in different different formats, and you can even get your slide deck as a Netflix video, or you can have your whatever brand pitch book or pitch idea as kind of different formats, right?
• 16:39
  SPEAKER_00: Um, and I think as you also get closer to how the audience wants to kind of perceive and understand that, you can also kind of you know tailor that on the on the recipient.
• 16:48
  SPEAKER_00: And even though that the creator might have some idea of how this should be created, there can be that recipient can also kind of you know perfolize it and contextualize that information in the way they want.
• 16:57
  SPEAKER_02: I don't think great new shows will be built that way, but personally.
• 17:00
  SPEAKER_02: I kind of love the idea of I love the idea of more seasons of the show you always love.
• 17:03
  SPEAKER_02: I think your gutted is over.
• 17:05
  SPEAKER_02: Exactly.
• 17:05
  SPEAKER_02: I think I think if it's already been made, then they could extend the season.
• 17:08
  SPEAKER_00: Or like change the ending from uh don't kill her, kill him, or whatever.
• 17:12
  SPEAKER_02: I mean, it's this great this is sort of you know parody like uh slight tragedy of what's going on with with AI right now is that it needs the internet, or the LMMs need the internet to feed their models, but the kind of uh the incentives for creating that content are going away, right?
• 17:27
  SPEAKER_02: So it's sort of like a victim of the same thing, says it's interesting how how that that element will play out because I think you need original content to feed the AI beast, right?
• 17:37
  SPEAKER_00: Yeah, I think I think with synthetic data and some of these things that is really starting to kind of become good enough.
• 17:43
  SPEAKER_00: I think is also I think we have so much data already that I think we've seen that with the data, right, we can create incredible things.
• 17:51
  SPEAKER_00: But uh of course, original ideas is gonna be the ultimate test for AGI, right?
• 17:56
  SPEAKER_00: How can you make new physics breakthroughs?
• 17:58
  SPEAKER_00: How can you actually kind of really kind of take the the the world understanding to a next level with uh with kind of new new capabilities just for listening to synthetic data?
• 18:08
  SPEAKER_01: Tell us about that.
• 18:09
  SPEAKER_00: Um, so I think as you're kind of you know running out of kind of humanly written content or movies produced on YouTube or whatever, there is kind of you know gonna be a need to kind of you know train your AI models, AI systems on synthetic data that is not created by humans, so synthetically generated data, right?
• 18:26
  SPEAKER_00: That will either kind of know change or kind of remask some of the biases your model have to kind of you know fill fill the gaps, but it can also kind of you know take your models and take your AI systems to the next level by feeding its systems data that is not really have enough in the original world and producing that with the the with synthetic data that then trains AI systems.
• 18:47
  SPEAKER_02: And that's happening now.
• 18:48
  SPEAKER_02: You're already there, already doing that.
• 18:50
  SPEAKER_00: Uh it's it's starting there, it's starting.
• 18:51
  SPEAKER_00: Yeah, no, this is already kind of really big field.
• 18:53
  SPEAKER_00: Yeah, cool.
• 18:54
  SPEAKER_02: Well, shifting gears a little bit to the business itself, many of our listeners are entrepreneurs.
• 18:57
  SPEAKER_02: I think they'd love to hear a bit more about your plans for the for the business.
• 19:01
  SPEAKER_02: One thing we were discussing before this uh this chat is that you know when the marginal cost of creating video is dropping to almost zero, even in this format, how do you think you will create and capture value moving forward?
• 19:13
  SPEAKER_00: Yeah, I think I think especially for um for us, right?
• 19:17
  SPEAKER_00: We're very focused on the enterprise.
• 19:18
  SPEAKER_00: So working with the biggest companies in the world, right?
• 19:21
  SPEAKER_00: And I think what they're buying from Cynthia is less about a video clip generator.
• 19:25
  SPEAKER_00: It's actually a full workflow, right?
• 19:27
  SPEAKER_02: Yeah, was it originally a video clip generator they were buying?
• 19:29
  SPEAKER_02: It was, yeah.
• 19:30
  SPEAKER_00: We were kind of delivering clips for for a long time and the early versions of the platform, you created an avatar, you built kind of you know the um the full video edits, right?
• 19:39
  SPEAKER_00: Where there's more to a video than just somebody talking to you.
• 19:42
  SPEAKER_00: There's animations, transitions, uh music, all the stuff that actually makes a video a compelling video, right?
• 19:48
  SPEAKER_00: So we kind of build uh we build kind of the full kind of product experience around this paradigm that generated media is different from camera recorded content, and that's kind of you know the product experience that we've kind of been tailoring for today, and also leveraging all the facts that every pixel is software rendered as opposed to be sensor captured, right?
• 20:07
  SPEAKER_00: And that has kind of been been been kind of the product experience and the product paradigm that we're building for now.
• 20:14
  SPEAKER_00: Um, but I think what we've seen is our kind of mode as we kind of you know going and maturing as a business from a research lab, can we actually make the technology work?
• 20:23
  SPEAKER_00: To oh wow, can we actually build in a unique product experience that is native to the fact that this is kind of AI?
• 20:31
  SPEAKER_00: And then next thing is oh okay, wow, we have such a great distribution, we have a license to talk with all these great companies, they have a lot of problems they want to solve with visual information.
• 20:41
  SPEAKER_00: How can we take and build custom workflows that is actually delivering really high-end business outcomes from our customers, right?
• 20:48
  SPEAKER_00: So I think there will be, as kind of video creation drops in cost.
• 20:52
  SPEAKER_00: I still think it actually is on the rise because you are constantly also evolving the capabilities.
• 20:57
  SPEAKER_00: We're still not at Netflix style, like customer onboarding in any kind of company, right?
• 21:01
  SPEAKER_00: But I think we we probably want to, right?
• 21:03
  SPEAKER_00: If we can get like a Libron Games style kind of onboarding video, if I'm a new employee at Nike, right?
• 21:09
  SPEAKER_00: Um with a really high production kind of quality of how do you actually become a Nike employee, right?
• 21:15
  SPEAKER_00: We're still kind of a little bit away from that.
• 21:17
  SPEAKER_00: So I still think in the short to medium term the cost will rise, but then it'll kind of eventually kind of go to uh to kind of you know to zero towards zero, right?
• 21:25
  SPEAKER_00: At least.
• 21:26
  SPEAKER_00: But I think at that point, right, we'll kind of be much more embedded in workflows and there's more value than just kind of the frames, but actually kind of inserting ourselves in more business impact, where today, right, we're delivering MP4 files that is an input to an outcome you want to produce.
• 21:44
  SPEAKER_00: If I go to any of the CFOs in the Fortune 500 and say, hey, you're creating faster and cheaper videos with synthesia, they're gonna say, so what?
• 21:51
  SPEAKER_00: I wanna kind of you know have uh lower customer support um cost, I want to have uh higher kind of you know productivity in my sales force with.
• 22:00
  SPEAKER_00: Better enabled sales people.
• 22:02
  SPEAKER_00: All these things, right?
• 22:03
  SPEAKER_00: And I think that's kind of where we're moving to now, where instead of just delivering MP4 files, they're actually delivering business outcomes.
• 22:10
  SPEAKER_02: So you're 100 million in revenue today or thereabouts,$2 billion valuation, how do you become a$100 billion company?
• 22:17
  SPEAKER_00: I think it's kind of moving into we we think there is actually a pretty big opportunity to um to build a really big standalone because the paradigm for generated media is so different than camera recorded content.
• 22:30
  SPEAKER_00: And if you think about the internet today, right, everything is kind of built for camera recorded content.
• 22:35
  SPEAKER_00: You shoot something with a camera, you put it into the Adobe products, and you put it on YouTube.
• 22:39
  SPEAKER_00: Those three kind of you know processes are completely air gapped and completely different, right?
• 22:44
  SPEAKER_00: And you can't really kind of go to a YouTube video and say, hey, make me 200 versions of this and A-B test my ad.
• 22:51
  SPEAKER_00: That's not possible today because you can't talk to the video editing and you can't talk to the camera, right?
• 22:56
  SPEAKER_00: But what synthesis built is kind of this fully connected thing where you can only you can also create the content, right?
• 23:02
  SPEAKER_00: But you also have the editing and the collaboration kind of capabilities, and then you have the distribution of it.
• 23:08
  SPEAKER_00: And that means that when you unify those three things, right, you can also start to build reinforcement or kind of human in the kind of um um self-improving systems, right?
• 23:19
  SPEAKER_00: So if I can see the people are dropping off after five seconds my video, I can render a new version that is becoming better at solving those problems, right?
• 23:27
  SPEAKER_00: Wow.
• 23:27
  SPEAKER_00: And I think that's like a little bit of the future of kind of building these kind of engagement algorithms on behalf of our customers to kind of drive higher performance through their videos.
• 23:36
  SPEAKER_00: And uh, and I think that's also kind of you know where the big paradigm of this is, right?
• 23:40
  SPEAKER_00: Right now we're using new methods to produce old formats, so faster and cheaper MP4 files.
• 23:46
  SPEAKER_00: But in the future, right, we're kind of enabling completely net new formats, new media objects that you can never upload to YouTube, right?
• 23:53
  SPEAKER_00: That you have to kind of consume real-time avatars, interactive, fully localized, fully contextualized, integrate with your CRM.
• 24:01
  SPEAKER_00: So we say hey Cameron, welcome back to uh Spirit Land.
• 24:05
  SPEAKER_00: We can't wait to serve you another Gentano.
• 24:09
  SPEAKER_01: So the classic kind of annoying VC question to founders over the last 15, 20 years has been, well, why can't Google do this?
• 24:15
  SPEAKER_01: The question now is why can't OpenAI do it?
• 24:18
  SPEAKER_01: You've talked a little bit there about the kind of stickiness, the how embedded you are with customers, you're solving a very specific set of problems for them, where even if someone like OpenAI can match the technology, they just don't have the focus, I guess, to go as deep on a very specific problem that you're solving for customers.
• 24:34
  SPEAKER_01: Is that it?
• 24:34
  SPEAKER_01: Or do you still feel like you've got to win on technology?
• 24:37
  SPEAKER_00: Yeah, I I I think we'll we'll gradually become more and more applied where we started as a very kind of can we actually make this work as a scientific breakthrough?
• 24:46
  SPEAKER_00: But I think it's very naive to think that you can kind of keep up with the model layer.
• 24:50
  SPEAKER_00: I think I'll be very, very hard and very few companies that have the capital and the resources to do that.
• 24:55
  SPEAKER_00: So I think over time we'll become less and less and less kind of, you know, scientific blue sky.
• 25:00
  SPEAKER_00: I still think there is some short to medium-term uh time to market that we want to win on and kind of continue to stay ahead.
• 25:07
  SPEAKER_00: But I think what we do is that we work backwards from customer outcomes.
• 25:10
  SPEAKER_00: We want to deliver a very specific customer outcome, and we're gonna take the best kind of off the shelf model, we're gonna take off the best kind of things and kind of really enable that thing.
• 25:19
  SPEAKER_00: And there are some things where we also work with a lot of like, you know, um voice providers, but we needed to build our own voice technology to have the best accent preservation because for our avatars, our executives, our kind of customers creating personal representations of themselves, they want to hear their own accent, and we couldn't get that in the market, so we had to kind of build that and kind of you know invest in that.
• 25:40
  SPEAKER_00: And that will obviously also be a solved problem over time, but they will give us a time advantage and a good market advantage, right?
• 25:46
  SPEAKER_00: But I think um, but I think what we what we see is that we are less of like an interface to an AI model.
• 25:53
  SPEAKER_00: If you're just like a singular modality kind of company, I think it would be very hard to kind of be competitive long term, as that will be very kind of hard to kind of keep up.
• 26:01
  SPEAKER_00: But if you're actually selling a full workload, a full outcome to a customer, they couldn't care less if this is uh UberDI, Claude, or Gemini.
• 26:10
  SPEAKER_00: They care about is this faster, cheaper, more productive, is this keeping my company safe?
• 26:16
  SPEAKER_00: That is the four reasons why you buy software and buy anything as companies, right?
• 26:20
  SPEAKER_00: And I think that's kind of where we kind of moving and kind of being much more tied to business outcomes and kind of you know, allowing us to do that faster and cheaper for our customers.
• 26:29
  SPEAKER_01: On customers, just briefly, how savvy, AI savvy are the customers of the big corporates?
• 26:34
  SPEAKER_00: Uh I think they're getting there.
• 26:35
  SPEAKER_00: I think a few years ago before ChatGPT, there was nobody who had any kind of AI procurement and uh any kind of guidelines for how to use service and kind of procure this type of stuff.
• 26:45
  SPEAKER_00: I think it's kind of increasing, but it's kind of you know overcorrecting a little bit in the short term, right?
• 26:50
  SPEAKER_00: Where that's typical kind of this J curve of everything, right?
• 26:53
  SPEAKER_00: Yeah.
• 26:54
  SPEAKER_00: Where they're kind of overcorrecting, oh, this is a completely new paradigm.
• 26:57
  SPEAKER_00: We have to stop everything, we have to kind of you know have a pause and everything, and then like, oh, okay, now we understand it.
• 27:03
  SPEAKER_00: Let's kind of come back to Earth, right?
• 27:05
  SPEAKER_02: Because you you have a unique seat leading one of these cutting edge AI startups, but working with really large companies.
• 27:10
  SPEAKER_02: Where do you feel we are in the AI hype cycle?
• 27:12
  SPEAKER_00: Uh, I think it's it's hard to say because I think it's also kind of real on the bottom, right?
• 27:18
  SPEAKER_00: But it's obviously extremely hot in at least the talent wall right now.
• 27:22
  SPEAKER_00: It seems seems completely crazy.
• 27:24
  SPEAKER_00: Crazy.
• 27:25
  SPEAKER_00: But uh, but it's um but it's hard to kind of see because you also have a lot of the signs of strong productivity, and we already see that really live in companies, especially on like coding the efficiencies and the productivity gains on like your developers, it's just incredible, right?
• 27:41
  SPEAKER_00: And I think that's kind of where it underpins a lot of the the investments, right?
• 27:45
  SPEAKER_00: But if you think about that global productivity can kind of go up with 10%, that is like tens of trillions in value, right?
• 27:53
  SPEAKER_00: If you could look just look at like the global scale, and that is and we only investing like 500 billion on that scale, right?
• 28:00
  SPEAKER_00: It looks kind of low, right?
• 28:02
  SPEAKER_00: Um, but uh, but yeah, I think I think you'll kind of you know it's definitely hot, but it's also kind of worn.
• 28:08
  SPEAKER_00: So I'm not sure if it's gonna slow down anytime soon.
• 28:12
  SPEAKER_02: Let's talk about 996.
• 28:14
  SPEAKER_02: For our listeners, 996 means working 9 a.m.
• 28:16
  SPEAKER_02: to 9 p.m.
• 28:17
  SPEAKER_02: every day, six days a week.
• 28:19
  SPEAKER_02: What's your take on it, Stefan?
• 28:21
  SPEAKER_00: I think I think it's naive to think that you can just start a startup, especially as founders.
• 28:27
  SPEAKER_00: I think it's less about being 996 in the office every day, kind of you know, all that time, right?
• 28:32
  SPEAKER_00: But if you as a founder, I I think I thought way more than 996 about my company in the early days and probably still is, right?
• 28:39
  SPEAKER_00: It's just ever consuming.
• 28:41
  SPEAKER_00: And I think if you if you think that it's gonna be a hobby thing, it's it's there's obviously many businesses and there's many different levels of ambition.
• 28:49
  SPEAKER_00: And if you are building a crafty kind of e-commerce type of offering, there's obviously more of a lifestyle thing to that, and that's totally fair.
• 28:59
  SPEAKER_00: But if you really want to build a global scale in such a fierce competitive kind of company, uh sorry, industry as AI, and try to be competitive with uh with China and the US and so forth, it's uh it's pretty naive to not think that you have to at least work uh extremely hard to get a chance for that.
• 29:17
  SPEAKER_01: Do you think you got to product market fit through incredible hard work and bulldozing and just grinding, or is it more very clever decisions and and being intellectually agile?
• 29:28
  SPEAKER_00: I think a combination, I think it's also synthesia is a pretty hard story to get, and it's still non-obvious.
• 29:35
  SPEAKER_00: Oh, it's just Amazon.
• 29:37
  SPEAKER_00: What is that, right?
• 29:38
  SPEAKER_00: It's still even today, it's still actually a non-obvious idea, and I think most of those investors that ended up getting it who actually knows the pains of producing videos themselves, and oh, this is fucking painful.
• 29:49
  SPEAKER_00: Yeah, uh, and they actually kind of have the empathy for the problem, right?
• 29:54
  SPEAKER_00: Um, but I think there's we've kind of been also starting this company the wrong way around, in the sense that we start with solutions and applying that to problems, right?
• 30:04
  SPEAKER_00: Where the typical startup is, oh, I have this problem, I want to build a solution for it, right?
• 30:08
  SPEAKER_00: But we kind of started the other way around, which meant that we went with the solution many different paths down the idea maze of kind of is this for gaming, is this for high-end visual effects, is this for Hollywood movie production, is this for high-end advertising?
• 30:23
  SPEAKER_00: And that just gave us those first three years, was extremely painful, and my hairline receded a bit over those years, but it was also just giving us so much kind of confidence once we kind of hit product market fit in uh in 2020.
• 30:38
  SPEAKER_02: How did you know you'd hit it?
• 30:39
  SPEAKER_00: Instantly, instantly, yeah, like it was just we launched kind of the self-service offering of the platform.
• 30:44
  SPEAKER_00: We've kind of been a service, a high kind of visual effects AI service for for many years, and then we kind of you know had this breakthrough where you can type in text, you can then go from text to speech and from speech to kind of expressions, which was kind of the magical kind of breakthrough we had in uh late 19, and then we kind of built the first version of the product of synthesis that you know today in uh in the summer of 2020.
• 31:07
  SPEAKER_00: And when we launched it, it was just like a magical thing that you can type in text, even though the the Avatos was very robotic, very static, very modern tone in their kind of you know uh body language, right?
• 31:18
  SPEAKER_00: But it was just instant that we kind of hit a nerve that this is interesting for people.
• 31:22
  SPEAKER_00: And I don't think we had like the ideas and the ticks and turns and the vision for where it could become back then, but it was clear that we kind of hit something that was interesting to people, right?
• 31:31
  SPEAKER_00: And I think it's also coming a little bit back to this when you I like the saying that it's easy to build a hard company and it's hard to build an easy company in the sense that when you have a crazy idea, back then the idea was to create a Hollywood-style movie on your laptop.
• 31:45
  SPEAKER_00: Um, that is kind of you know inspiring to a lot of people and allows a lot of people that want to kind of come on this journey and like try to kind of you know attack this crazy idea, the giant idea, right?
• 31:56
  SPEAKER_00: Um, versus if you're kind of building accounting software number 704, you're kind of probably gonna get people who's just there for the salary, who's there for like the nine to five, that is not like there for the for the full vision, right?
• 32:08
  SPEAKER_01: I'm surrounded on each side here by two great Danes.
• 32:11
  SPEAKER_01: You're you're uh, if I may say so, a classic Dane.
• 32:13
  SPEAKER_01: You're you're running this incredible AI company, one of the most exciting AI companies in the world.
• 32:17
  SPEAKER_01: You're very humble.
• 32:18
  SPEAKER_01: We're gonna ask you to be unhumble for a minute, which you ask all of our guests as a final question.
• 32:22
  SPEAKER_01: What is it about you do you think and your co-founder have that have allowed you to go and create such an exciting company?
• 32:28
  SPEAKER_00: I think it's kind of having kind of the vision and like foresight.
• 32:33
  SPEAKER_00: I think uh that is what what Victor's magic is, uh, is to really kind of see where the world is moving and being able to kind of see that before other people.
• 32:42
  SPEAKER_00: So we were just at the perfect spot, right, when Chat GPT kind of became AI on everybody's lips.
• 32:48
  SPEAKER_00: We already had a very fully functional product, we already kind of were AI first.
• 32:52
  SPEAKER_00: We it was not like a bold-on thing that we had to kind of know insert into our product as many SaaS companies are trying to do now.
• 33:00
  SPEAKER_00: Um, but it was kind of really building on the backbone of kind of you know this breakthrough.
• 33:05
  SPEAKER_00: And uh, I think in the back in 20 early 23 when AI and Chat GPT was kind of coming onto the scene, there was like five things before beyond ChatGPT you could do, right?
• 33:17
  SPEAKER_00: Go into ChatGPT, write your script, put it into synthesia.
• 33:20
  SPEAKER_00: So we were just like blowing up like crazy back then.
• 33:22
  SPEAKER_02: You were very Danish, you didn't you it's not about yourself.
• 33:24
  SPEAKER_02: What about what about you?
• 33:26
  SPEAKER_02: What about you as a person?
• 33:28
  SPEAKER_00: I think I think my superpower is probably kind of you know to actually make it happen.
• 33:34
  SPEAKER_00: I'm extremely executional oriented and very impatient.
• 33:38
  SPEAKER_00: So actually kind of transforming these uh lofty ideas into actually what do we do?
• 33:44
  SPEAKER_00: Not tomorrow, but right now.
• 33:45
  SPEAKER_00: Sounds like about these things, could match in that way then.
• 33:48
  SPEAKER_00: Uh yeah, so I think that's kind of you know the left and the right brain between the two of us, where I'm kind of you know overseeing kind of sales, operations, finance, kind of been doing everything else, and he's kind of you know having a lot of division, thinking about the future, and I'm thinking more about the present and uh reporting a bit on the past to the investors.
• 34:05
  SPEAKER_01: Brilliant.
• 34:05
  SPEAKER_01: Stefan, thank you so much for that giant idea.
• 34:08
  SPEAKER_01: Really appreciate it.
• 34:08
  SPEAKER_01: Thanks for having me.
• 34:10
  SPEAKER_01: If you like the episode, don't forget to like, subscribe, and follow the show, and you can find many more giant ideas wherever you get your podcasts.