Martin Pagh Ludvigsen, Goodby Silverstein’s AI Director on the Future of Creative Work

Meet Martin Pagh Ludvigsen, Director of Creative Technology and AI at Goodby Silverstein & Partners, as he breaks down how AI is actually changing creative work inside one of the world’s most respected advertising agencies.
In this episode, we talk about how AI is being used in advertising, why Martin’s role sits inside the creative department instead of tech, and how agencies are using AI to ideate, prototype, and sell creative ideas faster.
We also dive into AI slop, human taste, creative judgment, and why Martin believes the future of creativity is not about replacing people, but helping great ideas move faster.
Martin shares how Goodby Silverstein & Partners used generative AI on projects like Giraffes on Horseback Salad, what it was like working with Google DeepMind and Google Cloud, and why AI video still needs human craft, editing, and post-production to become truly production-ready.
Whether you’re a creator, agency owner, filmmaker, or creative professional trying to understand where AI fits into your workflow, this episode is for you.

Learn more about Martin:
Portfolio: https://martinpagh.com/
Goodby Silverstein & Partners: https://goodbysilverstein.com/

CONNECT WITH US
Website: https://vidpros.com
Instagram: / vidprosedits
Facebook: / vidprosedits
LinkedIn: / vidprosedits

MORE ON VIDPROS
Need pro video editing? Vidpros helps creators and brands produce retention-focused content that performs.

📩 sales@vidpros.com

Transcript

SPEAKER_01 0:00

Martin works at one of the most respected advertising agencies in the world. But his role sits right at the intersection of creativity and AI. In this episode, he breaks down how AI is actually being used inside top creative teams, why it's more powerful for ideation than execution, and how the real advantage today comes from taste, judgment, and the ability to go beyond what everyone else can generate with the same tools.

SPEAKER_00 0:26

My name is Martin. Martin Peck Ludvigsen. That's the English pronunciation of my Danish name. So let's do the Danish one as well. Martin Peig Ludwigsen. I am a Dane living in the Bay Area in California, and I work at the legendary advertising agency Goodbye Silverstein and Partners as the director of creative technology and AI.

SPEAKER_01 0:51

Awesome. And your role sits inside the creative department, not tech or strategy. Why would you say that is important?

SPEAKER_00 0:58

Yeah, that is that's a really good question, and something that probably at this point in time sets us apart from other directors of AI in the advertising industry. I have worked as a creative my entire career, but with a background in technology, where you in advertising you typically have creatives who come from art, design, or writing is sort of the classic background for it. But but lately we've seen creatives come from different backgrounds, and my background is technology. Um I actually, somewhere in my baggage, have a degree in computer science from the IT University of Copenhagen. It's now been more than 20 years since I got that degree, but that really was my background and what brought me into the industry. But you know, there are obviously a lot of engineers, a lot of coders in the world, and our interests go in a lot of different directions. And I've always been very interested in using my craft to bring things on screen that can excite people. And that is that is honestly what brought me to advertising in the first place, that I can put my work out into the world in a way that excites people. And there are lots of other components of technology in advertising, the whole industry is built on technology. We have what we call ad tech, which is essentially the technology that makes it so that the ads are served in front of you, whether that's on your phone or on a digital billboard or on TV or online on a banner or whatever it is. That's probably the much bigger side of technology inside advertising, but we also play a significant role in the production of the core of advertising, which is the creative output, the stuff that hopefully excites people about the brands that we work with and make them want to learn more about the brand or buy the product or subscribe to the thing or whatever it is we're trying to do for our brands.

SPEAKER_01 2:53

How would you say your role is actually different from how most companies are thinking about AI right now?

SPEAKER_00 3:00

Yeah, that is that is a big and broad question. We we did have a bit of a claim to fame, that we were one of the first agencies, maybe the first major agency to actually hire a director of AI. I was hired uh three years ago. Nowadays, you have lots of agencies, most of them will probably have someone with the title head of AI or director of AI or something like that. Um, and I think that most of them are focused on AI in the process of everything that we do because AI is obviously a pretty big deal. Let's just say that it's not just another hype, it's not just another technology trend that we're trying to turn into uh the next shiny object. It is a paradigm shift. It is completely changing the way we do knowledge work, it is changing the way we do creative work, and and that means that it it plays a really big role throughout different organizations. But I was brought in first and foremost to consider how we could use AI as a way to come up with new genres of creativity inside the agency. Um and when I joined Goodbye Silvestine and Partners in in 2023, they had already been doing generative AI for years. And actually, one of the reasons that I wanted to come work for this agency out of all the agencies was that way of leveraging creative technology and AI in a way that I didn't see anyone else uh doing. There was work out in the world for the Dali Museum in St. Petersburg, Florida, that our agency, I can now say, have worked with for a long time, have done some really exciting stuff for. Uh, there was a project that had launched recently for BMW, um, the um an AI a generative AI dream car, like projecting generative art onto a car, which was just like this is this is just completely different from from from what you see out there. And they wanted to bring in someone who could lead those kinds of initiatives and figure out how to how AI could could evolve their creativity and take it to exciting new places. And then what changed since then, and you know, it's been it's been about three years since I've worked there, is that AI became part of everything. You know, we obviously the the the big moment was November 2022 when the first version of Chat TBT launched and uh the the world collectively had a holy shit moment. This this can actually generate new text, it can come up with text that didn't exist before. Those of us who were a little closer to to the industry had already seen this would would probably come, but but um yeah, but it's it's it's always interesting to think about that chat GPT was not a product that OpenAI intended to launch, it was just a demonstration of GPT, of what the platform could do. But it really became the moment that unlocks generative AI for for you know the whole world. Politicians, normal people, you know, school children started cheating on their tests with it and and and all the stuff that that came from that. So it became it became something that everyone everyone understood. But but we had already been able to generate new images for for a while when that when that moment happened. And that capability has obviously only gotten better and better since then, uh to to the point where you know I'm I'm sure we've reached the point already where there's where there are more AI-generated images on the internet than there are images made with in the in the traditional way. So so things are moving very quickly, the tr the technology is changing all the time, and my remit is to try to be a little bit ahead of that all the time so that we can keep breaking through the noise and get attention to our brands for coming up with new and innovative ways to tell stories, and a lot of that is is is now through AI, let's be honest.

SPEAKER_01 6:48

Yeah, one thing that I'm very curious about is like how what was the process like for you adapting into this new technology? Like, what what was like your day-to-day life at work before AI became this mainstream thing, and what is it like now as a creative technologist? Like things changed. What what was the process for you?

SPEAKER_00 7:14

I personally have a few benefits in in this in this paradigm shift. One of them is obviously that I do have a computer science background, so you know I know at a deeper level than most people how computers work. So that's obviously that's obviously important to keep in mind. But uh in addition to that, I've always been focused on and been asked to focus on creative technology, which means what are the cutting-edge technologies that exist out there in the world that we might someday be able to use to tell stories for our brands. So the idea of adapting to new tools, to new platforms, to new ways of thinking is something that has been part of my work for my entire career. Whereas for other people, that might be more challenging. I think from what I understand, because I'm not that old, but there was a transition when you sort of went to working on computers as as as a as a as a as a paradigm shift that happened in our industry. So so we're not just talking about work that is displayed on a computer, you know, web banners and stuff like that, but it's actually doing your work on a computer, sending and receiving emails, working with cloud documents as an example, was something that a lot of people had to figure out how to do and adapt to because this this curiosity about new tools and platforms that sort of that I have and is also uh asked to have, that not everyone necessarily will have that desire to like go out and try new platforms and tools. And that may that means that the transition to using computers for work was difficult, but you know, we managed as an industry, all knowledge industries, creative industries managed, and now we're seeing the next transition into into this this brand new world where we might be learning that you know 80% of what we used to do, of what used to be the process of creative work, of of knowledge work, can now be handled by automated systems. And we now will have to focus on the challenging, the heart, but also the wonderful part, the last 10 or 20% that makes our work better than the slob created by by everyone else. Because, you know, that's the that's the world that we live in. Everyone has access to the exact same tools, so everyone can use the AI tools to generate something that is 80% of the finished thing. So, in a world where that is our competition, we now have to figure out how to go above and beyond what everyone else is able to use the AI tools for because we all have access to the same tools, the tools are more amazing than they have ever been before in history. So, what do we do with the access to those tools to make sure that our work is better than that that comes from our competitors or even from you know new types of creatives or agencies or people out there competing for the business that we have, the business of advertising and creativity?

SPEAKER_01 10:18

One thing that is interesting is that usually this type of curiosity comes with fear, right? So I I I know I I don't really want to get into that AI versus human thing, because that's not really a thing, I believe, but more of like as an integration. What does that actually look like in practice? And what do you think people misunderstand most about AI in a creative workflow?

SPEAKER_00 10:43

Yeah, I mean that that is a that that's a really touchy subject, right, right now, right? Um what is how is it going to change how we work? And I think it's important to understand. I mean, I'm just expressing my opinion here. But very recently, OpenAI decided to shut down Sora, their platform for generative video. And that came as a shock to a lot of people that they would, you know, abandon the platform that allowed them to generate video. And I think that is because OpenAI's end goal was never to generate the best possible video and you know destroy Hollywood with with with amazing new generative video tools. Their goal is to understand the world so that they can achieve this you know AGI goal that they have that they have set set out in the world to accomplish. And as a step on the way to achieving that AGI goal, they needed to create video models that could understand the world and try to recreate the world. So so all this all these outputs that come out of their models, even you know, ChatGPT, as I said, was kind of an accidental product success. It's it's it's people get to experiment with and use for productivity tools that the tech companies put out in there in the world without them being sort of the end goal. This was these were never meant to be the final products. But we we take them and we find that okay, there's we can actually we can do something interesting with this. We can do something with Sora or with ChatGPT or with Vio or with Runway or with Kling or whatever your your preferred tool is to come up with something that the creators of those tools likely never intended. They they just created these foundational models that can, you know, in theory generate anything that you can imagine, even some things that you can't imagine, but they don't know what what what can come out of that once they put them out into the world. So so really what we as human beings suddenly bring to the table is what makes us uniquely human. We are the ones who have the original ideas, we're the ones who come up with something that has never been done before, and now we just have more ways of expressing that. I think that's I think that's that's wonderful. And it's so, you know, when you browse through your your feeds on your whatever your preferred social platform is, you see yet another video, Hollywood is cooked, and then they show some recreation of something that Hollywood has done before. You know, it's Brad Pitt fighting Tom Cruise, it is a new Transformers movie, or whatever it is. It's like, well, I you know what, I don't think Hollywood is cooked just because someone is able to recreate with like 80-90% level of quality the same stuff that has already been done before. Because at the end of the day, they're just repeating what has been done before, they're just regurgitating art that already exists out there in the world and and producing it at a lower quality uh and producing content that no one actually has any emotional connection with or really even care about. So, you know, it's good for the headlines to say that that all the creative industries that we know cooked, but at the end of the day, what makes us creative, what makes us unique and human is our talent, it's our ideas, and it's our hunger to go out and create new things. So I think if you are if you're someone who has become a creative because you're fascinated by the tools, then no matter what, the tools will always evolve. There will always come a new version of the tool, a new way of working that will be better than than than what happened before, you know, final cut will be replaced with Premiere or the other way around, or or someone will come in and create a new workflow for VFX that that that is better and more efficient than what you've done before. But so the tools always change. The way that we express our ideas will always change, and now there's a new way. But but at the end of the day, the ideas come from us, and the ideas is what takes us from the 80% work, which is all the slop that's out there, and and and truly elevate it to something that that people will care about, that people will want to engage with.

SPEAKER_01 15:06

If you're a content creator and editing is hitting up your week, pay attention. Vid Pro is a professional video editing service built specifically for creators, not random freelancers and not AI tools. You get a dedicated human editor who learns your style, pacing, and brand from day one. You film your content and upload the raw footage, and your editor handles everything: cuts, captions, color grading, sound design, and short form clips. We deliver everything through Google Drive and frame.io so you can leave timestamped feedback and requests and limited revisions. We start with a $100 trial week that includes 10 hours of editing. Go to vidgrolls.com and start your $100 trial today. Stop editing and start growing. Where in the creative process for you, does AI actually like create the most leverage today? Would you say it's ideation or do you think it's execution? What what what is it like for you?

SPEAKER_00 16:02

It is far more on the ideation side than it is on execution. I I love that you in a way, if we're being a little bit cynical here, you no longer have to like articulate a completely coherent, perfect thought and idea. Because one of the things that I actually really enjoy about working with AI tools is just to have sort of a stream of consciousness conversation with your tool of choice. There are, you know, all the different platforms have some sort of live mode where you can actually have a conversation with it. Um and what I really enjoy about that is that that is a mode where I can sort of just like talk out into space and come up with my ideas as I am talking, sort of like work on them and keep saying things, and I will say nonsense, I will not complete my sentences, I will interrupt myself. But that doesn't matter because I know that there's this live feature of the of the AI that's constantly listening to me, and that is responsible for for organizing all the weird stuff that comes out of me and hopefully help me sort through it later and and and figure out what's good in there so that so that so that I can decide it. It's it's it's something that came to me from I've always been a terrible note taker. Just like even when I was in school, it was like everyone else was like, oh, they had their like highlighters and their different pens, like why I should I be taking notes because I see that everyone else is doing it, but I was just never good at it. And it was just it's just been so amazing with AI now that you can just like record the meeting and then say, can you summarize this? And you will get much, much better notes than any of my classmates with their highlighters were ever able to do. And I'm like, hey, wait a minute, the way it can structure unstructured conversations and thinking is probably something that I can use myself. So so even when we have sort of the closest we have to a writer's room situation where at our agency, where we have a bunch of people in a room workshopping ideas together, it's just like record the conversation, and then we can just talk and not worry about actually writing down what we come up with. We don't even need a whiteboard anymore because everything that we say is being recorded, and at the end of it, we just hit stop and then we say, hey, let's organize everything that we just spoke about. And it's like, well, there it is, these are all our ideas. So, so so it's a tool to assist us in coming up with ideas. It's it's absolutely amazing. And you can even you can take it uh one step further than that, uh, since it is a live conversation with the AI, and you can instruct the AI to start asking you questions to better strengthen your idea. Because I I firmly believe that the idea should come from you. AI is not creative in itself, it cannot come up with new things, but you can. And AI can be really good at asking you pointed questions to help you go deeper into your idea, to break through the surface, to really strengthen it so that you can articulate and come up with all the answers that that that turn your idea from half-baked into something amazing, or sometimes that's also a desirable outcome, help you realize that maybe this was not such a good idea in the first place, so let's get rid of that and move on to something else. So, so as a as a way to help you organize all the wild things that are flying around inside your head, it is it is you know it's it's truly remarkable for that. On the on the execution side, since that was the other half of your question, then that is obviously something that that evolves over time. But even when we use the most modern, you know, video models at our disposal, we find that the quality of the output is typically not good enough for production usage. Uh maybe in on like smaller social things, we can get to something that at least as a can be a starting point for something that can be final. But we do find that that there's a ton of post-production, old-fashioned post-production required to turn something from you know a series of unstructured generative clips into something that is actually good enough to put out into the world on behalf of one of the brands that we work with. Um but but sort of in between, you know, the the the idea and the final product is also what we call the prototyping of our ideas. And and and this this really is another huge unlock for us in in the way that we work. Um advertising is is an old industry, and for a very long time we have been used to the process of selling a video idea to our clients, which is that we will present them with a script and with some supporting art for that script, some boards, a few keyframes to sell through the idea. And that has worked okay for us, but but it's always been a challenge that our level of our ability to abstract from from scripts to final product uh are different, especially between you know creatives and and and the decision makers on the client side. In other words, they may not see in their heads what we're seeing in our heads when we read out a script and show some supporting art. So there's a sort of a gap in between what what what what we as creatives see the idea is and what what what what they may see. And that can lead to some bias remorse, let's just say say it that way, right? Like you you you may think, okay, I I said yes to this idea, but it didn't turn out the way I had planned. And and and thanks to generative AI, we can now go much, much further than that script plus supporting art than we've than we've been able to before. Obviously, very quickly we discovered that that gener using AI to generate comps was was much faster and a better use of our time than building those comps in Photoshop, which is you know, there are lots of creatives in our industry who probably. Have built their career at being really good at Photoshop. And it's always a good idea to be good at Photoshop because it's a great place to finish things. But as conceptual creatives, it is not the best use of our time to spend, you know, six, eight hours building the perfect comp uh to illustrate an idea when we can use generative AI tools to do it in you know 30 minutes and maybe even come up with a bunch of different iterations and and and land at something that's that's better than what the comp was capable of. And then the next step is obviously adding motion to those comps, which is which is what what AI video now enables us to do, that we can start to visualize how these stills will come to life as as the final video. And obviously, we we work in a lot of different genres of of video, whether it's live action, CGI, claymation, or you know, what have you, all sorts of different things. But this idea that you can actually start to prototype what what the finished film would look like with AI, that makes that gap much, much smaller than it's ever been before, because our clients can now start to see a much clearer visualization of what it is that we have in mind with uh with with with that idea. And that is that is helpful, but it's also it's also a challenge for us that that that we've gotten to that because we are that has now become the expectation that we take these visualizations, these prototypes of our idea further than we've done before. And it is time consuming and something that we have to you know use strategically when we when we develop ideas. When we when we give a brief to a creative team at our agency for how to solve a business problem with creativity for one of our clients, they may come up with 15 different ideas or 20 different ideas, and it obviously is not a good use of their time to prototype every single one of those 15 to 20 ideas. So that means that everything that we've been so good at for a long time, which is to like find the good stuff in those 15 to 20 ideas, that's the role, that's the job of our creative directors, that still applies. And then at some point later, when we can decide, okay, from these 20 ideas, these are the three that we really believe in. And out of these three, this is the one that I think we're going to recommend to our clients. So that is the one where we're gonna put in the extra effort to to prototype it, to show a deeper visualization of what this idea can finally be before we bring it to our client.

SPEAKER_01 24:36

Do you think like the whole idea that we had in the past that we used to learn in film school of vision board is completely shifted now?

SPEAKER_00 24:45

Do you think it's dead? No, I don't, I don't, I don't think it is. I don't think it is because I still think it makes sense to be able to see the entire film displayed in one on one page, right? Which is which is essentially what what what a board can be. At least when we you know, we we typically our films are typically no longer than 30 seconds, maybe 60 seconds when we make the the big Super Bowl spots. But that means that you can actually squeeze the entire thing into one board, and that still makes sense, and that is still a part of the process for us. So so I don't think that I don't think that's that. I actually don't think that the the steps in developing an idea they're they're they're they're they're unchanged. It's just that some of them we can go deeper than we used to before. We can some of the some of the steps in the production of a film that we used to wait until pre-production with or even production with, we can now start to experiment with those steps earlier on in the process than we've been able to before, before we go out and hire anyone to start uh to start making this. I I mean I think it's sort of we all agree that the traditional storyboard artist, at least for our industry, that's we're not hiring as many storyboard artists as we used to, because that that step might be something that that that that has been replaced by AI. But the people who were storyboard artists before can now, they too can now take their work further and actually start to to produce visualizations, animated visualizations of of what it is that they created as stills before. So I I think everyone, everyone who has been involved at some step in the you know conceptualized conceptual development, pre-production, production, and post-production of ideas can now take it one step further. Or in you know, in in some cases, if you sit in post-production, you can actually like reach back into the conceptual phases and and even be part of that. So the I think the waterfall model, which is something that comes from software, right? The that the idea falls down a waterfall and you have to finish the previous step before you move on to the next one. Um that might be changing. That we actually, thanks to these generative tools, can suddenly go way back to you know to the script phase and make changes, even though we're technically in the schedule in the post-production phase. We even we didn't actually do it, but um we made a film that came out last year where very late in the process we discussed recasting one of the one of the actors in the film because not because well it was a generated, it was a fully generated film, right? But we we were like, I think I think we we need a different type than what we are. It would be interesting if we had a different type. So just the idea that something where we are in full post-production, we're editing the finite film, can suddenly start to discuss what if we recast that person as someone completely different, and we didn't actually do it, but we could have done it. And that to me is wild that you can go that far back in production um and and make changes. It's it's also scary that suddenly, you know, you used to be able to tell your client or whatever it is, oh, sorry, nope, that is locked. We have script lock or we have casting lock or whatever it is. Most of the time we still do, but but at least in theory, you can you can now make really radical changes much, much further back in the process than you've been able to before. And I don't think that is that is a paradigm shift that that that it is at least theoretically possible to do that.

SPEAKER_01 28:14

But one thing that a lot of people fear is get into that zone of AI slop. So for you, how much of uh human taste and judgment is put into this? And where where in the process do you actually put like the human creativity, the taste, the judgment, and so we don't replicate just some sort of slop?

SPEAKER_00 28:38

Well, that that's why we're still here. That's why we'll never be out of a job, right? Going back to when Sora was shut down, which I really think is sort of a milestone in the history of AI and creativity. But there were people out there saying, hey, look at all the people who were creators thanks to Sora. They're just they just went away now because the the tool, the platform that they happen to use to create their art is not there anymore, so now they don't know what to do and they just disappeared. I don't know if that's necessarily true. Some of them may still resurface because maybe they did have good ideas that can be expressed through another platform, but but it is what separates good work from from slop. There's there's something fascinating about you can have work that goes out into the world that is, you know, you you look at the sort of the quality of the final output, and it can be it can look, for lack of a better word, perfect. But because it is generated by AI, uh, there will be a lot of people who reject it. And I don't necessarily think it's because they think it it looks bad, it's because of the way we perceive art that that exists in the world. We can we can tell when there's no authorial intent inside it. We can tell when the machines have taken the steering wheel when the machine is in control of the output and and the human is only like making tiny little adjustments here and there. Um that's something that we we we know from reception theory, which is which is not something that most people think about, but something that I actually also went to school for. It's like, okay, what is it, what is it that that the author encodes into their work, into the literary text, as we call it, of themselves, of their experience, and and encode and send to the person who will watch the work at some later point. And that person will then also bring themselves into the work and it becomes something new, something that they own. And and and when there is the missing connection between the author and and and the viewer, you can you can tell something feels off. All the pixels might be perfect, but there is no human connection there. And I think you know, if we might get to a point where where AI will create art that is seen by other AI, and that's fine, they can make a connection with each other. But as long as the intended audience is human, they want to feel a human connection with the human author that that that that is there in the first place. And and that's why we can tell when when when when when something is off. I've I've been revisiting Rick Rick Rick Rubin, the uh the producer lately, because he he has such an interesting approach to what he does, which is that he, you know, as he says, he has uh taste and he has a great confidence in his artistic choices in that taste and being able to express what is good and what isn't. And he also famously said that he's not he doesn't have any technical skills to speak of because that's that is not what he brings to the table. So so we all need to channel our inner Rick Rubens as you know there was definitely a moment where that when that video came out a couple of years ago where where he said, I have no technical skills, but I have great confidence my taste, or whatever it is that he said, we're like this fucking guy, but at the end of the day, he's right. And and if you have that good taste, which is something that that you can practice, you can practice having good taste, then you have a really good chance at being successful in a world where 80% of the boring work is done by an army of AI interns, and you then come in and and apply your intent, your artistic sensibility, your taste, your creativity to the work to make it something that people will actually want to connect with. So the other piece of advice of his that I want to relay to everyone else is to go out there and immerse yourself in the good stuff, listen to great music, watch great movies, go to the theater, watch, watch great plays, to like learn what great art looks like because by being exposed to great art, your taste will develop and you will be able to create the same great art that that it is you're seeing. So so it is it is the most important thing. Taste is the most important thing that we have left, and what will set us apart in the in the age of AI slob.

SPEAKER_01 33:28

I think just talking about what you're mentioning right now just came to my mind um the project Giraffes on Horseback Salad. Because it's a great example of looking back at like an artistic piece at art and it's like recreating it and bringing it to life. Can you talk a little bit more about this and why this is a good example of this shift?

SPEAKER_00 33:50

Yeah, absolutely. Yeah, it's so funny. Giraffes on Horseback Salad. It's been, you know, we're talking today in early April, it's been about a year since it came out, and and that is that is a long time in in AI years. AI years are like dog years, or maybe even longer than that, right? So but I think it has aged remarkably well. Um, what we've done. We other people seem to think that too. We've been nominated for two Webby Awards, so I'm very proud of that. By the time this this podcast episode comes out, we will know whether we won or lost. But but just being recognized and getting two nominations means that this is obviously something that that that resonated with people. Um, and the the way it came about was, you know, as as these video models were being developed, Google Cloud and Google DeepMind came to our agency with this new model. It was actually Google VO1 back then, and reminder we're on version 3.1 now. But but they wanted to find out what can a sort of a flagship creative agency like Goodbye Silverstein and Partners do if we put this model in their hands. Um, and then we went off and worked with a with with an absolute absolutely wonderful team, Lucas and Katsu. Props to you. Great stuff you came up with, uh, two creative directors who who I worked with on this. And one of the things that that we decided to do was, hey, why don't we actually embrace the hallucinatory nature of AI? Uh, this was a point in time where everyone was very focused on, you know, AI images having six fingers on each hand. That has obviously improved since then. But but but to us, that could be a feature instead of a bug. And going back to our long-standing relationship with the Dali Museum, we uncovered this script screenplay collection of notes called Giraffes on Horseback Salad that Salvador Dali had tried to sell in around 1937 when he was in exile here in the United States because of the civil war in Spain at the time. So he was over here. And he had this wild idea that he wanted to make, and he was told by the Hollywood studios at the time, this is too wild. There, we can't we can't do this, the technology does not exist to bring this to life. So it was forgotten about for a very, very long time. It's been since recreated as a comic book, and it is something that's been discussed that maybe it's worth figuring out how we can bring this to life. But but thanks to generative AI, we actually could bring this to life in in a way that was it's you know, let me don't get me wrong, this was still a lot of work and a lot of late nights, but it was a lot less than it would have been if we had tried to do this as live action and a massive post-production CGI effort that that it would have been to truly do this at the Hollywood level. So, but but but but really this this thing about the the domain of of Salvador Dalí is obviously the the surreal domain, and in the surreal domain, you're already asking yourself, am I hallucinating or is this real? So so this idea of of hallucinations might not be a bad idea. In fact, it might actually help bring the idea forward. Um so the I mean I am also guilty of AI slop because there are definitely things in that film in the in the uh it's it's about a 17-minute short film, and even in the trailer that we released, there are things that people can point out and say, okay, that's sloppy. Like we have a a clock with Roman numerals on it where we where the number eight is written wrong, and and we're like, okay, you know, we could fix that, but but we we don't have to fix that because this is not a world where everything has to be, where the expectation is that everything has to be pixel perfect. This is fortunately not a car commercial where where there is an expectation that everything everything is picture perfect. It is the domain of the surreal, so so so we can get away with it. Um so so that really felt like the the the right thing to work with. And it was it was so interesting because you know, Google, who is behind this model, they are also a big tech company, and they are also the end game for them is also not to create a video model that can create perfect uh videos, and then they're gonna go home and write off into the sunset. They're also on their way to to doing something much bigger with with AI in the world, but but they also have demonstrated an interest in creativity in what we do, and an interest that is expressed through the questions that they ask us about okay, what what what would you like to get out of this? What are you what are you what are you actually interested in? And I remember we had some conversations with the team there where I complained that, well, you know, it's a shame that when when when they output clips, they're H264, because they're already encoded in in a in a super lossy format with with tons of artifact. It would be much better for us if if it came out in some you know intermediate codec or maybe even fully lossless, so that we have much more information uh to work with in the post-production uh process. Uh because right now it's it's the equivalent of like editing with videos you rip from YouTube. And and that that came as a surprise to them because but we thought you you wanted the system to spit out finished films. They they were like concerned about, yeah, we're so sorry, it can only output eight-second long clips. And we're like, that that's fine. I mean, we're gonna we're gonna edit it. No clip will be will probably be longer than five or six seconds. It's very much a music video style. So they were focused on on putting different features in than we actually needed as professionals. But they listened to us and and were actually able to understand how we worked as filmmakers and and and how what what they as you know AI researchers and engineers put out in the world and could help us sort of bridge the gap there. And it's been a real pleasure to work with with the DeepMind and with the cloud team in their you know understanding of what it takes to actually bring film into the world, more than just you know what's what Sora was, which was essentially like a like a quick content platform where people could have fun and make a video and put themselves into whatever and then move on with their day. So you could tell that that Google was interested in the professional, in the artist community in a way that that no one else was. And they were they were super excited to see where we where we took giraffes on horseback salad. It was it was created with VO2, so which means it's the it's the previous generation of the model. But when you look at it today, it's like wow, it is it is super impressive. I I would say it still has a little bit of that animated slideshow vibe that that AI film, a lot of AI film still has, and and around that time that was that was that was what most of it was because it still didn't understand action that well and couldn't really like make things happen in the scene. Um but but but but other than that, I think it it has aged well, and I'm I'm I'm I'm very proud to have it in my portfolio.

SPEAKER_01 40:56

You know, one thing that I think is really interesting is how AI makes filmmaking accessible. Because I remember back then when I got to film school, in order for us to create a project, it was such a hard process. It was a lot of money involved, it took a lot of work, and now it seems like anyone can be a content creator, a filmmaker if they actually understand the rules and they use AI and they apply those rules. And I think it helps us to uh open doors to younger talent that in other ways would not be found. And one thing that I'm very curious about, like for a project like this, uh, do you think it could have been sold to a client without showing those high fidelity visuals? What would be the process like of selling a project like that?

SPEAKER_00 41:47

Oh, yeah, that is that is a good question. I don't think I have a good answer to it. I was actually I was talking to one of our founders, Rich Silverstein, about it the other day. He has sort of famously come up with a concept for a musical about the Cuban Missile Crisis, where we've uh created a concept video and he's he's he's trying to sell it to Broadway and not having a ton of success with it lately. But what he relayed to me is actually that when you work with with with with Broadway or with A-list Hollywood, then maybe maybe you don't want to bring something that looks too finished to them because that makes them feel like they can no longer influence this. It's like, well, you've already made this, why would you why would I want to make it there? I I can't, as a director, as a filmmaker, have any influence on it. So I still think there's there's room for the you know the more old-fashioned way, which is to write a script and and and try to sell that, which has always been hard. Uh it will continue to be hard. So so I don't know that the this idea of like prototyping every idea that you have is is is right for them because they might just look at it and say, oh well, you've already made this, I'm not not interested anymore. But but elsewhere in the ecosystem of getting your content out into the world, I I do actually think it it makes sense. Um because you can you can get people interested in in what your idea is, but but I I I would and and by that I mean that you can you can prototype your your film idea and and actually get it to a point where where people can start to see what it is that you have in mind. But I would not try to finish it. I would it's it's something that you know I find myself doing when you're working with generative AI tools. Suddenly you're like, oh, you're working on this tiny little detail getting that right, and like, what am I doing? I'm not making a finished film here that's meant to go out into the world. I've already gotten my point across, so who cares if the logo is wrong down there in the left corner? It's not important for me to pitch the idea, and yet you find yourself trying to fix it anyway because if you're like, oh, I just just one more prop, bro, just one more prop, then it'll be right, and it never will be, right? You always suddenly it's 2 a.m. and it's it's worse than the first generation. That you made a 2 p.m. So just do enough to get your idea across. Just do enough to get people excited. Definitely do enough so that you yourself are excited about your idea. But then recognize that in order to stand out from the sea of AI slob that's out there, you need to involve more people in filmmaking. I've never really been too involved in sort of the independent creator community. I've always worked at you know advertising agencies that that sort of work at a certain tier of advertising where it's bigger productions. So for the TikTok content creators of the world or the real creators of the world, I don't I don't know what what's what's in it for them. They're interested in the person making the art, the person telling the story. That's probably what brings a lot of them to the table. So brings in a lot of their fans, right? So I don't think that they have anything to fear from these like pure AI celebrities out there, the ones that exist only in AI. Because once again, we're human beings, we want to connect with other human beings, and we can tell even when the pixel perfect character on screen everything looks right, but it's like it's not a real human being, it's not a real actor, it's not a real person. There's no real intent behind what they're doing there. So so so keep keep focusing on on your what makes you unique. Keep focusing on being on being human and and and then you can take control of what AI gives you and and and keep keep doing what you do and use it to your benefit, use it to make your ideas better, but don't let AI take control away from you because it can never replace your talent, your humanity in what it is that you're doing.

SPEAKER_01 46:02

That's a discussion that I've been getting with a lot of creators lately, which is like, is it easier now to become a creator or more difficult? Because there's so much content created with AI that maybe the competition of like human content is maybe you know much smaller now. We never know. That's a question that a lot of people are going through and they're asking themselves. Because we go online on YouTube and there's a lot of AI slop content every single day. There's people making money with content that they didn't do anything, they just generated fake images and they asked Claude to generate a script and they just put it together and like for some reason it created something. So I think it's interesting that from your perspective, you feel like this won't affect a lot of people in the content creation industry and the creator economy, because that's a fear that a lot of people have. But I think your point makes a lot of sense.

SPEAKER_00 46:58

And I think there are still people who will have bad ideas who maybe are not as talented as as they hope they could be. That's always been the case in the artist community. But but they're they're not gonna lose this battle because AI is is better than them. They're gonna they're gonna lose it because they unfortunately don't have good ideas or don't know how to express their their good ideas. I think I think someone who has it who has a good idea now has has a better chance than ever to actually articulate that idea in a way that other people will find engaging and interesting.

SPEAKER_01 47:33

When it comes to having good ideas, do you think because of AI the expectation has shifted one or two years uh from one of two or two years from now? Like our clients becoming harder to impress or expecting more clarity? Uh, because I I know you said that you don't want to show the full project, the full process yet. You don't want to give them the full vision. But where do you stop? How much of it do you show, and how much do people expect you to show once you start working with AI?

SPEAKER_00 48:03

That is we're still figuring that out. I I think that's the that's the honest answer to that. We actually just Friday we shipped a video to one of our clients, which was an internal video. And this was an interesting project for us because they asked us if we could make a video for an internal meeting, a big internal meeting, but still something internal, right? And and back in the day, we would have said, well, yes, we can do that for you, but the cost is going, it's not going to be much different from if it was, you know, a commercial that we made for you, because we still need to spin up the whole production apparatus to make this happen. But but this time we said to them, well, you know what? If you're if you're up for it, we can we can we can do this with generative AI, and we can create something that we believe will be good enough for a short internal meeting for getting a reaction when when they present something on stage, which is was what what this was about. Um so they're sort of like, okay, let's let's let's let's go ahead and do it. So so that meant that this production was handled entirely by you know my team at at the agency, uh labs as as we're called, generative AI video production. And it's the interesting thing is that we we went in there with them and and we said, well, this is gonna be AI, so we all understand that it's not gonna be perfect. Uh like there were some things about their logo and things like that that we told them ahead of time isn't gonna be good. But we made it, and it it actually got really it it got close to being to being you know good enough to send out into the world. Not not quite there, but I would say we reached that like 90% level. And one of the one of the interesting things that happened was that our client now started providing feedback as if this was meant to be a perfectly crafted production, like came back with you know brand guidelines and and and guidelines for the characters that we used in there, which which is the kind of feedback you would expect in a real production that we're gonna send out into the world. But we had to like, oh yeah, well, it's great that you're giving us this feedback, but reminder, this is you know, this is generative AI, and and in order to accommodate this level of feedback that you're giving us right now, we probably have to go to like a full CGI production to actually make that that that happen. Um and they were like, oh, right, right, yes. So like they like remembered what we agreed to. It's just that we got it to such a level where they got back into the habit of providing real production feedback, and we had to remind them that that is not what this is. So you you want to make sure that that you you focus on what is what what are what are we actually what are we actually doing here? I I also have a background in user experience design, and it's it's sort of a classic thing that when you create your wireframes for your website or your app or whatever it is that you're making, that you make sure that everything is in grayscale, in black and white, because as soon as you start to put color on it, clients will start to make comments about you know the color palette and your design choices, and they're like, no, no, no, that's not what we're supposed to be talking about right now. We're supposed to be talking about the user experience. So you want to make sure that you you make it good enough so that you can have an informed discussion and make decisions about the the aspect of the concept, the project that you're that you're doing right now. But when you start adding that extra level of polish, you are gonna get feedback on that extra level of polish. And that's probably not you're not at that phase in the project right now. You're trying to sell an idea, you're not trying to sell a finished thing. So so we're we're trying to figure out like, okay, how how much do we need to bring them? Because if we overdo it, then everyone also suddenly gets an idea of like, this is, yeah, we saw the finished thing, but it's like, no, no, no, no, this is this is not the finished thing. This is a concept. And and when we you know, when we shake hands and you sign the check, we will actually go out and make it. And and we don't want to lock ourselves into specific directorial author author choices in that in that that project that we shouldn't be making until later. For instance, when we sign a director to make this film, because you know, reminder, when we work in advertising, by the time we sell through ideas, we have not yet gotten a director's treatment for it yet. So that's a whole nother step where that director's artistic vision also is just as important as it's always been to make sure that the finished commercial is a success.

SPEAKER_01 52:34

And once you go through that director's uh treatment, what is the post-production process like? Would you say that editing style for AI content is different than traditional editing?

SPEAKER_00 52:48

Not too different. I I I think one of the things that impresses me the most about the post-production process is working with you know with professional editors who can you s you you articulate an idea out into the to the edit suite and they make it happen in real time. And that is that is that is real craft, that is real skill that I just don't see that that that AI will be able to do that um at any point in time, anytime soon. Like this understanding of what it is that you're doing. There are other steps in in the in in the in the process where you know it's maybe it starts to make sense to to to to use generative AI um along the way. It's you know, when we put our finished product together, sometimes we buy stock video to fill a gap somewhere, sometimes we we use b-roll, sometimes we can use generative video to maybe make a plate that we that we didn't already have, that's you know, maybe we it used to be we would like go out and get a stock version of that to to fix it, but now hey, it's actually really nice that that we have some options here for for doing things differently. And then the other part of it, when we do actually have clips, components, scenes, whatever it is, in our in our editorial process that are made by generative AI, the quality is usually a lot lower than than than than what we uh than what we produce in the good old-fashioned way. So, you know, the dynamic range isn't there. It's it's as I said already encoded at H264. It is maybe the resolution is is is low. I mean, we you know we we try not to do it, but sometimes okay, a 720p video is all we've got, so we we gotta try to put it in put it in there and and and make it look as good as possible. And and I still it's just especially for someone like us, the the we still need a better quality um than than than than what we can get out of it. So so so that process really really hasn't changed much. But but the there we are also seeing that the post-production companies that we work with, they are really good at making sure that something's when we like sorry guys, we only have the 720p generative video. That's the that's the highest quality asset that we've got. And they said, okay, we'll make it work, and they have to put a lot of work into making it work, and and that's where where we're super impressed with some of the production partners um that we have. And we we famously, was it last year, did a Super Bowl spot for Mountain Dew, where uh Seal, the singer, was a SEAL and sang a kiss from a lime. And that was such an absurd video of a seal oddy with seal the artist's face, that a lot of people assumed that this was surely this must have been something that they did with generative AI. And and we went out, well, there was no generative AI whatsoever involved in the production or the post-production of this film. This was the old-fashioned way with a face rig and tons of like postwork required to to make it look as good as that. But but it's just, I guess we had already no, it's more than a year ago now, arrived at a point where you see something that is just so wild and absurd, you're just like, okay, surely this must be this must be AI. It's like, oh no, it's this is good old-fashioned craft and really good craft, I might add.

SPEAKER_01 56:12

One thing that I want to talk about a little bit is uh the process, the creative process for you guys inside the company that you're working for right now, because that's something that I'm very curious about. Like when you're creating a project, how many people are involved in this project? Are roles emerging inside agencies and production studios right now? What have you seen happening and how many people usually are needed to create a project?

SPEAKER_00 56:43

That has not changed that much for us. We we still operate with the old-fashioned um advertising agency model, which is that a project will typically have a creative director who is you know sort of responsible for all the creative on a specific account. And a brief will then get assigned to one or more teams of typically an art director and a writer, a copywriter, and they will be the ones who sort of come up with the ideas that that come to us from our from our strategy department. You know, our strategy again, our strategy department will get a brief from the client. This is the problem that we need you to solve with creativity. We're launching a new product, we need brand building. Those are like classic problems that that we're solving, and we will turn that into a brief based on the insights that we have about the product, the brand, the audience, the channels that it's going out into. And then the the creative team will come up with ideas. But what what what is interesting and and maybe a little different about our advertising agency is that we have an in-house innovation unit that I lead, I'm the director of, it's called Labs. And we're the ones who are responsible for coming up with ways to articulate ideas that maybe haven't been done before. So what happens then is that these creative teams working across our various clients, whether it's you know Mountain Dew or Doritos or Lace or BMW, they will come to us and say, Hey, labs, we have an idea that somehow uses technology to solve this creative problem, and this is sort of what we have in mind. Can you help us figure that out? And and that's where where our sort of experimental nature comes in, because we will see it, and and and sometimes it's just a conversation, sometimes just well, this is really interesting, you're onto something here, and you can use this social platform or this technology or uh or whatever it is. But sometimes we look at it and say, okay, this is this is interesting. I don't know what the answer is to this question, but we will find out for you. And then we will set out and and and do some experiments inside our little innovation department to try to help them articulate that I that idea through through technology. And and and you know, this is a case where AI has really changed a lot because we can suddenly be much more productive at trying things out quickly with AI. Um, and now we're not just talking film ideas, we're talking, you know, app ideas, technology ideas, use of AI. You know, we famously created a lobster telephone last year that you could pick up and have a conversation with Salvador Dali in the Dali Museum in Florida, which is obviously that is not traditional advertising. I think we can all agree about that, but it is still a very, very creative idea that came from a brief from the museum about how can we connect a modern audience with the art of Salvador Dali in ways that haven't been done before. So, so so our responsibility, our remit is to do those experiments, try things out using technology that haven't been done before. Um, and we have a privilege in our little innovation unit that most people don't have in the advertising industry, and that is that we are allowed to fail. And what I mean by that is that we're called labs, and labs do experiments, and experiments sometimes they go wrong, and sometimes they go right. And if they go wrong, we will not have bet everything on this particular horse. The creative team that came to us will probably well, they will always have other ideas that solve the business problem in you know more traditional way. Maybe that's that is a video idea, maybe that is an experiential installation. But but we're the ones who are responsible for figuring out what if we did it this specific way? Um, and and and that can actually lead to some some really interesting cases. We had a a truly wonderful example of that in a campaign we did for the Ritos and Stranger Things for for Netflix Netflix last year, where the the idea was that they came up with a video idea where people you had an old school 80s telethon where you could hear people calling into the telethon because they wanted to help the people of Hawkins, Indiana, uh, the fictional city in Stranger Things. And and that was basically the idea because we launched a product with Doritos, which was called Collisions, which was a sort of a merger of the 80s design style of what Doritos looked like back then, and it was also the meeting between the upside down and the real world. So that was that was the product that we were launching, and we were going to launch that with a with an old school telethon. But the creative team came to us and said, Hey, what if you could actually call into this telethon and have a conversation with with the host of this telethon, and the host of this telethon would be 80s style celebrities. And and we said, you know what? I I think that might actually be possible. So we already had the idea of recreating a telethon, but what we brought to the client was in the meeting, we actually called up an 80s celebrity and had a real conversation with that 80s celebrity through AI to demonstrate to them, hey, we think people can actually call in to a phone number and really take this idea to the next level. And you know, our Doritos client, the Netflix client, they absolutely loved that idea, and we sold it as an add-on to the bigger campaign idea, and that was absolutely wonderful. But if, say if if we had gone back in time and we had found okay, this technology is not quite ready for this, we can't quite do it, we would still have been able to do an 80-style telethon video, and that could have been the campaign in itself. But because we were allowed to fail, we could add this thing to it that really took it over the top and and made this into an absolutely wonderful campaign. It's been a huge success for us and for Doritos as well, uh our client.

SPEAKER_01 1:02:48

What do you think people that are doing what you're doing right now need to do to stay relevant? Because um it's scary for a lot of people.

SPEAKER_00 1:02:57

Yeah, I so I didn't go to South by Southwest this year. I I usually go, but you know, just it's it didn't fit in the calendar. But there's there's a session that I always want to check out, which is when when Amy Webb and the Future Today uh Institute release their annual trend report. This year, uh to my surprise, they decided to retire the trend report because it doesn't what they're saying is it doesn't make sense to talk about trends anymore. By the moment the report comes out, the trends will already be outdated. So they've released something they call the convergence outlook instead, which is this idea that a lot of ideas, a lot of technology is converging at the moment and creating disruption both forward and sideways, and it it really is quite chaotic. And their advice to organizations, which I also think is a really good advice to creatives, and a scary advice, is to bet on something, find something that you think is going to be a really, really good idea, even if it's not yet, even if it's not quite there yet, and focus on that. Go go all in on that. Because right now everyone can be an 80% expert on everything, thanks to AI. So if you're if you try to do if you try to be a jagger all trades, you will be a master of none. So I think the advice right now is to actually try to be the master of one thing, and then really cross your fingers that it turns out to be the good thing, that you made the right choice. And that's why it's scary, because you might you might end up you know uh betting on Betamax and it turns out to be VHS that that that wins the competition. That was a dated reference, but I I still think that most most people understand it. So and if I I wanna I wanna talk about passion here, the one thing that you bet on should be something that you're passionate about, because you will always be better at something that you're passionate about than something that you don't care too much about.

SPEAKER_01 1:05:01

Now I have one final question, which is another advice, but to creators or agencies, what's one mindset shift they need to make about AI?

SPEAKER_00 1:05:14

I don't think it's intelligent. I think you need to start thinking of AI as what it is. It is a technology to automate things, and it can automate a lot of things for you. But but you are commanding an army of interns that honestly they don't know what they're doing. But sometimes they will come back with something that's absolutely amazing, but they don't know that it's amazing. Um AI has no taste. It comes back with something and it has literally zero idea if what it came back with is good or bad. It just knows that according to probability and statistics and the mathematical model inside of it, it addresses what you asked it to do. But if it's good or bad, it has no idea. So you just, you know, again, going back to the advice I just gave, if there's one thing you want to bet on, it is your test. Taste. And it is to be confident in your taste and and believe in that. And then you you want to make sure that the one thing that you bet on, um be really, really good at that, because all the underlying parameters, everyone will eventually get to the same level on them, because that's the stuff that that that these automated systems, these automation technologies can replace for you. So try to find out what is it that makes you really good and unique as a creator, as an agency, and and then figure out a way to let AI handle all the busy work and you focus on that thing that you're really good at.

SPEAKER_01 1:06:39

Awesome. Thank you so much for joining us and for doing this. And I've learned so much from you, and I was so interested about everything that you've been doing and your workflow. And if there's anything you want to promote or any project that is about to be released right now, or maybe, you know, just a shout out to Webby's. Anything, feel free and do it right now.

SPEAKER_00 1:07:01

I mean, I would I would say go vote for us in the Webby Awards. We're really, really proud as an agency that we have eight projects nominated across various categories. But I know that but by the time this episode comes out, the the voting will be over. But we're still we're still very proud. We're you know, we're not a huge agency, so for us to have eight works nominated at the Webby Awards is is is just such an accomplishment in itself. So I'm very proud to work at Goodbye Silverstein and Partners.

SPEAKER_01 1:07:28

Awesome. Thank you so much for doing this. Thank you. If there's one takeaway from this video, it's this AI gives everyone access to the same tools, but the people who win are the ones with the taste and judgment to turn that into something people actually care about. Subscribe, leave a comment, and I'll see you in the next episode. Bye.