Diriger l’avenir : savoir-faire et créativité à l’ère de l’IA

[MUSIC PLAYING] [CHEERS, APPLAUSE] MIRA LANE: Welcome. My name is Mira Lane. And this is, I would say, one of the most exciting times to be a filmmaker. The tools are changing, the possibilities are expanding, and these three amazing people are at the center of it. As you just heard, Doug Liman has redefined action genre through movies like "The Bourne Identity," "Mr. and Mrs. Smith," "The Edge of Tomorrow." And the three of them are co-founders, Julina and Jed are also co-founders of 30 Ninjas, a new AI filmmaking studio and a new way to work on interactive content. So I thought we could start by just talking with you, Doug. First of all, maybe take us into your mind a little bit and talk about the way you think about emerging tech and what got you excited about AI to begin with. DOUG LIMAN: Well, in terms of getting excited about AI, I think art has always reflected the technology of the time. And this is definitely the technology of the time. So as a filmmaker and artist-- I don't necessarily think of myself as an artist because I make commercial Hollywood movies-- but still, in terms of artists embracing AI, it's the technology of the time. And really, my career, I started with "Swingers." And you wouldn't think of technology if any of you have seen "Swingers." But the reality is, back then, films were shot on film. And Kodak had come out with a new high-speed film stock. And I was able to use that to make "Swingers" on a shoestring budget that would not have been possible two years before. And so from the beginning of my career, technology has empowered me to tell stories. And any time there's been an opportunity to look at a new piece of technology, I've jumped in because I got rewarded in different ways. We're in the Guinness Book of World Records, not for anything athletic, unfortunately, but for the first VR series. And then a couple of years ago, I went to one of these Google dinners where they were like, would you think about doing something with AI? And I jumped right in. I was like, actually, I've got movie script-- maybe 2 and 1/2 years ago. And Google was like, wait, wait, wait a second. We were thinking maybe a little short film. But for better or worse, I've been willing to jump into journeys. I don't know how they'll end. I used to guide whitewater rafting trips-- Julina's been on them-- where they don't always end so well. But the one thing that's for sure is I don't necessarily know what's coming up around the bend, but one way-- I'll try to navigate us through it as best we can, but sometimes we get to the other end and not everyone's in the boat. So right now, 30 Ninjas has powered me to make the first fully AI studio film with Casey Affleck, Gal Gadot, Pete Davidson. MIRA LANE: Which is amazing. DOUG LIMAN: And then a cast of about 100 other actors. And my kind of brand of movie is big, big movies, big ideas. Like "Edge of Tomorrow" is a very big idea. It's time travel and aliens. But the biggest thing in that movie is Tom Cruise and his character. And for "Bitcoin," which is the film I just did with AI, it's a giant world that's set in the world of cryptocurrency. But the biggest thing in the movie is Casey Affleck's character, Pete Davidson's character, Gal Gadot's character. And I think when I was making "Bourne Identity," I came from making "Swingers" and another indie movie, and then suddenly I'm making this big Hollywood action movie. And it was very rough going because it didn't look what the studio was expecting a movie to look like. And because I was bringing some of this new technology I used in "Swingers," in which the studio was saying, "Bourne Identity" looks like shit. Literally, it was said to me at the time. It didn't look like what they expected a spy movie to look like. And at some point, Stacey Snider was running Universal was starting to yell at me that "Bourne Identity" was not my $50 million film school. And the fact that I'm sitting here on stage admitting to that now means that I've completed-- and not with therapy, by the way-- a journey where I'm willing to look back and then go, you know what? It was my film school. Of course I didn't know how to make "Bourne Identity" when I started that. I'd made "Swingers." I figured it out as I went along. And I think more than a therapist helping me get to that place was an "Edge of Tomorrow." We were having some script issues. Because if you try to do anything with time travel, you'll come to a realization that humans are never going to travel through time because it's almost impossible to even write a script with time travel in it. There's so many paradoxes. And we were having a meeting where we're really struggling. And Emily Blunt made a suggestion. And I snapped a little bit. And I was like, that's never going to work. Because we had a looming start date and I was feeling a lot of pressure. And Emily was like, whoa, easy. I've never made a movie like this before. She'd come off "Devil Wears Prada." She'd never been in an action movie. And I was like, well, I've never made a movie like this before either. And my producer who is in my office, was like, oh, my god, you can never say that out loud again. Warner Bros. is giving us $150 million to make this movie. You cannot say out loud, you don't know what you're doing. And Tom Cruise said, I love that Doug doesn't know what he's doing. I want to be on a journey with him as he figures out how to make this movie. And I think not therapy, but Tom Cruise saying that to me made me look back on "Bourne Identity" and go, yeah, I'm willing to stand up in front of you guys or on the set and be like, I don't know how to do this, but God knows I'm going to try because I'm going to have to face an audience when it's all done. MIRA LANE: And tell us how you did it. DOUG LIMAN: So that's probably why I'm probably a pretty good person to venture out and make a film with AI, entirely with AI, when not that long ago we were anticipating post-production was going to take 40 years, 4-0. And I was like, but I know we're going to figure it out, and we did. And we've cut it down to six months. But when we embarked on the journey, it was like, this may be, at my age, it may be like that church in Barcelona, where the person who designed it doesn't get to be there when it's done. The Gaudi church. MIRA LANE: The Gaudi church. DOUG LIMAN: But I was counting on Google and the other people in this world to accelerate the curve. And so I showed 10 minutes of the movie last week at Cannes, and it's extraordinary. MIRA LANE: And we'll get a little bit of a behind the scenes look at this in a moment, actually. Well, maybe we could talk a little bit about 30 Ninjas, this company you guys have started. And maybe Jed and Julina, tell us about the company and the types of projects you guys are working on. JED WEINTROB: Sure, yeah. First of all, thanks for having us at Google I/O. I know this is a developer conference, and we have very strong roots in a large part of our team who are software engineers. Who in the audience comes from software development, creative technology engineering? Awesome. And I'll assume the rest of the people are a little more film and entertainment. 30 Ninjas was founded, really, to use new technology to empower and supercharge storytelling. And that's meant many, many, many things over the years, going back to interactive cinema. We did the first-ever long-form virtual reality series in 2015 called "Invisible" with Google Cardboard and Samsung at the time. And we're actually in the Guinness Book of World Records for the first-ever long-form VR series, which is cool. We've done a lot of interactive media. We tend to partner with large brands, like Google, when they're coming out with something new and exciting, to be their storytelling partner. And we started going down this trip with the team at Google about 2 and 1/2 years ago, first around the launch of Android XR. 30 Ninjas, right now, our studio has two different arms that will eventually become one. One is an interactive story extension and character extension company. We work with XR. We have an upcoming Google Glasses project, which we haven't really talked about yet, which is going to be really exciting. We work with XR headsets. We work on mobile platforms really to take and create-- and Julina will talk to this a lot, I'm sure-- AI-driven characters who are tied to linear properties. So we're very much of the belief that as audiences, we are always going to sit around together and watch or listen to stories told by visionary creators that are linear. And in addition to that, technology and current audiences are really expecting to be able to go beyond that with the characters they love, and the worlds they love, and the stories they love. So we use generative AI to supercharge an audience connection with a character, a story, in a world where they can talk to them or have an adventure with them. And that's great. And then the other part of the studio is our immersive film studio and our AI film studio. And that's where, as we were working with the XR team, they sort took us behind the doors and were like, hey, take a look at what Veo is before it comes out. And we've been working with both our own engineers and the DeepMind team to refine what we do really well, which is work with top-tier, sometimes Hollywood, sometimes sports, sometimes music creatives, and creators, and endemic digital creators, to tell their stories using AI. And lately that has meant a very artist-first, human-first approach to AI storytelling with video. Do you want to talk a little bit about-- because I think we'll look at a little "ASTEROID," a little bit about characters, and maybe "ASTEROID." MIRA LANE: Julina, you want to tee up "ASTEROID," and then we can watch the clip and talk about-- JULINA TATLOCK: Can you say it one more time? Sorry. DOUG LIMAN: "ASTEROID." MIRA LANE: You want to tee up "ASTEROID"? JULINA TATLOCK: Oh, sure. Tee it up. Yeah, yeah. So "ASTEROID" is one of the projects that Jed mentioned. We created it for the Galaxy XR headset. It's a film. It's a story, a space story, that we did, that Doug directed, in 180 stereoscopic. But after the film, you discover that someone-- well, I don't want to spoil maybe the trailer if the trailer was there. But after the film, you discover that one of the characters is still alive and you assume a role, and you then try to rescue the character. You engage with the character who is written with generative AI, Gemini Live, actually, API. DOUG LIMAN: So DK Metcalf, a football player, is one of the characters in the rocket ship of the film. And he's playing himself, a version of himself, a football player who's manager said, this would be a good thing, do you want to go up in this rocket? And it turns out to be anything but a good thing. And by the end of it, it doesn't end well for most of the characters on the asteroid. And he ends up stranded and left behind. And originally, he was left for dead. And then Google approached 30 Ninjas about, could we extend the story using agentic characters? JULINA TATLOCK: Yeah. So the idea was, how do we incorporate that into the plot so that basically that the viewer can become a player and participate in the story and assume a role in order to rescue DK? DOUG LIMAN: But normally, I have to do a sequel. But they were like, could we do something where the story could continue if you liked it? And you could imagine this being applied to movies and other things. You like the character, you like the world, and now you can interact with it, and it's been trained. So you interact with DK, but it's not just DK Metcalf, the football player. He's DK Metcalf, the character in our show. So he can talk about football because he is playing himself, but he also is DK stuck on an asteroid trying to get rescued. "Bourne Identity" is immersive filmmaking, right? It's like, you're in the car with Jason Bourne. That's why I jumped at the VR series that we first did. We had to wear this cardboard thing on your head to watch. And then when the Android XR headset was going to be coming up and there was the opportunity to make that as a film for a really proper piece of equipment, I jumped at that to really put you in the rocket ship. And so whatever they're going to show you here, I dread because it's not meant to be seen on a flat screen. It's meant to be experienced like you are there. You are in the rocket ship. And Hailee Steinfeld plays the lead. And Ron Perlman's in it. And you're really on the journey with them. And that's when Jed talks about 30 Ninjas having these two sides. Julina has been doing interactive storytelling and branching narrative since it was first possible, and trying to figure out lag. If you're going to make a choice, how can it go seamlessly to where we are today? And 30 Ninjas is very much a story-driven company. And how can we use technology to further storytelling? Obviously, if you've seen Bourne, you know I want to get you really invested in it. You're there with Tom Cruise and on the beach in "Edge of Tomorrow." I want you to really be there. And the ability to control some of the story-- some of it, because there is a role for the filmmaker. We need to build the world and build the characters. It's not just explore the world and anything happens. But the ability to maybe incorporate the best of what storytellers, like myself, can do with giving you some control within that world is something that's really exciting to me, and something that Julina has really been pioneering at 30 Ninjas. And so that's, when Jed talks about the two sides to the company, there's the interactive side, and then there's the other side, which is what empowered me to make "Bitcoin." MIRA LANE: But before we jump into "Bitcoin," why don't we show you all a little bit of "ASTEROID"? So we've got a short clip that we'll play and we'll dig into "ASTEROID" a little bit before we jump into "Bitcoin." [VIDEO PLAYBACK] - DK, sit down. - Get out of my face. I have training in two weeks. [TENSE MUSIC] - [HEAVY BREATHING] I can't feel my face. - Keep moving! - Jesus. Jesus, don't pull those. [EXCITING MUSIC] - An asteroid-laden with gold and precious metals passing 200,000 miles from Earth. It's a space gold rush. How was the landing? - Textbook. - You came back on this spacecraft without them. - Oh. - God damn. - We had six hours on the asteroid. The real problem was the cold. - What? [GRUNTING] - Stop it. - [GRUNTS] This isn't over. [END PLAYBACK] [CHEERS, APPLAUSE] MIRA LANE: It's a really cool experience. If you guys haven't had a chance to do it, you can see it on one of the headsets. Really immersive. And let's talk about the character at the end and how you use AI to interact with this character. So as Doug mentioned, one of the characters is left at the asteroid at the end, and the movie extends. So you get to interact with them in real time and talk to that character. DOUG LIMAN: And this isn't the first time, by the way, that I've killed off a character to have producers come back and be like, bring him back to life. Because when I did "Roadhouse," a key part of that film was that Jake Gyllenhaal had to kill Conor McGregor at the end. It's like an essential part of his arc. And at some point, Joel Silver came to me and said, Conor is too iconic in this movie. We have to keep him alive for sequels. And I was like, no, it's the whole arc of the movie that he dies. That's Jake's entire arc. He's worried that there's a killer inside of him. If he doesn't kill somebody at the end of the movie and then have to leave and go back to the shadows-- that's essential to the movie, to Jake Gyllenhaal's arc. And Joel Silver's like, no, I'm not a dummy. I know that that's essential for Jake's arc. Yes, of course, he has to kill Conor. I'm just telling you, he also has to be alive at the end of the movie. [LAUGHTER] And we had to figure out a way to do that. And here, it's a similar thing where I've killed DK, and then they're like, OK-- MIRA LANE: Bring him back. JED WEINTROB: It's not the end. DOUG LIMAN: It turns out he's not dead. And that's how we got to suddenly an additional chapter to the "ASTEROID." JULINA TATLOCK: The real groundbreaking thing that happened with "ASTEROID" is that we conceived of it as the 180 film and the story extension, the story extension then using generative AI in order to create the conversation and drive the interactivity after. And right, I would say, about four months before we were going to premiere, Gemini Live native audio came about, which was actually the ability for natural language conversation to drive the engagement with the character, was really a game changer. So then you move past the NPC and you can still stay within the story world, stay within the context of the story with the stakes, and carry that tension from Doug's film and the immersive experience that you're in in headset, and then talk to the character and try to save him. And it's a Unity app, too. So you're within a representation of the set as well, using and leaning on the construct that you, the person, the player, you've assumed someone from NASA. So you're on Earth, and so you're not seeing a full-- it is photorealistic, but it's abstracted in terms of what the visuals are on the asteroid. But DK, the way that Doug wrote DK and that character, he's extremely likable. Not everybody in that film is a likable character. They end up, some of them, killing each other. But DK is funny and likable. And so you really do also want to save him. And that was the key point and really necessary for the character to work. MIRA LANE: And I'm sure you've gotten questions about, how did you use AI in this? And did this take the job of screenwriters? And so maybe, Julina, talk a little bit about that. And what that's actually like in reality. JULINA TATLOCK: I mean, it hasn't been our experience, I would say. DOUG LIMAN: It sort of predates AI in a way. JED WEINTROB: But it's a great question, the AI character question. JULINA TATLOCK: Right. So for an AI character, for the interactivity, number one, there was an enormous amount of writing that goes into it. I mean, it's a novel worth of writing. It's 1,000 pages for DK for a short experience. DOUG LIMAN: They had to write every-- not everything that DK might say, but they had to train the model. They had to put DK through a ton of experiences and someone had to write that. MIRA LANE: Yep. JULINA TATLOCK: And then create the game in a way that there's a dramatic situation. And so then you have the overlay of game design, linear storytelling, interactive storytelling, and then needing to care and want to rescue-- DOUG LIMAN: I mean, our experience at 30 Ninjas is, obviously, because we're a startup, we're generating a lot of jobs that didn't exist before that because we're an expanding company. JED WEINTROB: We're creating the jobs. We're not generating jobs in AI, just to be clear. DOUG LIMAN: Fair enough. JED WEINTROB: And the script is a script written by a writer, right? So there's already a writer, and a movie, and a script with this character. So the amount of writing it then takes actually to craft an AI character that not only has a complete backstory and is trained and tuned, but also has enough guardrails around them so that when people try to mess with the character and get DK to trash talk other players in the NFL, he won't turn off, but he will say, well, I think everyone who gets to the level of the NFL has to be a super talented athlete, and they're all amazing. That takes a lot. And it takes a lot of humans not only to write, but also to test. And then DK tests it. And he's like, yeah, that kind of sounds like me, and that's pretty cool. DOUG LIMAN: But what I was going to say is that, there is the interactive part of 30 Ninjas, and then there is the AI side, which did "Bitcoin." Because I come from the film business, and so many of my friends aren't working right now. The business is broken. It is, for sure. The number of my friends who are not working as writers, as editors, as directors-- and what I find as sort of extraordinary is, how many of the people like that 30 Ninjas has hired, whether it's top film editors who just are not currently working because there's a quarter of the number of movies being made that were being made, big movies like that. And they're fantastic. Screenwriters who are working, generating the shots for "Bitcoin." And they're bringing all of that expertise as storytellers in the industry and from different parts of the industry. So my experience on the job side is that, because we are a startup, we're creating jobs for people where maybe the traditional industry is not currently. MIRA LANE: Well, so speaking of "Bitcoin," why don't we jump into that? And, Doug, maybe you could give the audience a summary of the movie, and then we'll show a clip after that. DOUG LIMAN: So "Bitcoin" is the story of Craig Wright, played by Casey Affleck, who may or may not have created Bitcoin, and the forces that come out to destroy him. And Pete Davidson plays this fantastic billionaire magnate who's making the biggest gamble of all history on Casey. And Gal Gadot plays a reporter trying to figure out what's true and not true. And the reason why I thought this story would be fantastic to tell through technology is it's set in the world of cryptocurrency, which already is sort of pitting humans against machine. The whole thing is, do you trust-- the whole thing about Bitcoin is don't trust, verify. Like, don't trust the humans, verify. Let the computers do it. And are humans corrupting the original vision of what Bitcoin was supposed to be? So within the world of "Bitcoin" is this battle between humans and machine. And I try to never repeat myself as a filmmaker, and part of how I do that is by working with different people. There's other filmmakers who their films all look sort of similar because they work with even sometimes the same actors. Because I'm always going to be the same person. All I can do is change the people around me. So this time, I was like, well, one of the people I'm going to change around me isn't going to be a person. It's going to be a computer. And so I'll make something different than what I've made before. And this seemed like the perfect story to do that with. MIRA LANE: This is a full feature movie that's using AI. Describe to the audience what's happening with AI and what you're trying to do with it. DOUG LIMAN: So we started down this journey 2 and 1/2 years ago, because of Google, of being like, hey, come let us show you what the future's going to look like. And where I settled at in terms of how AI would work with a director-driven movie, not some random AI slop movie, but I want something that represents the best that I can bring to the table, the best that my production designer can bring to the table, my DP can bring to the table, my actors can bring to the table. That was the most important part was-- Casey, and Gal, and Pete are fantastic in the movie, and I know that because not only can I tell that, but when I show scenes, which unfortunately I'm not allowed to show here, people are like, they're fantastic. And that's figuring out how to get their performance into the computer, along with the department heads and the best of all of us, and then, yes, the prompters, who then can add a special sauce so it doesn't look like what I've done before. And that was a really challenging process. And I have to give a real shout out to-- or he's going to kill me for doing this-- to my producer, Ryan Kavanaugh, who raised the money for this movie and embarked on this journey with me. It's one thing for a director to be crazy and be like, OK, I actually have no idea how this is going to work out, but I'll go down that-- and sort of crazy that you would go down those rivers with me, when I literally have no idea how we're going to get through it. But for a producer to have the courage to back a filmmaker and go like-- because when we started this, my estimation, with no exaggeration, was that post-production was going to be 40 years, 4-0 years. And it was like, we're just counting on that there's going to be technological breakthroughs while we're prepping and figuring it out, and that will continue. And now we've cut it down. Two months ago, it was three years of post, and now we're down to six months, which fits within a traditional film schedule. But the entire film was captured-- all the performances were captured on a stage that we designed over the course of six months of full prep R&D. MIRA LANE: And it's one set, right? DOUG LIMAN: It's one stage. Everything was shot on it, no matter whether they were inside a swimming pool, whether they were on a boat, on the beach, in Vegas. JED WEINTROB: It's like this tent. We could do it here. DOUG LIMAN: And the thing that was unexpected is once-- I knew I was like, I'm going to embark on this journey, and I don't exactly know how it's going to affect me, but I'm just going to go with it. And one of the things that I discovered was happening was that normally, a filmmaker and the producers have this kind of dialogue where I want to shoot a scene in Vegas, and you can't afford to go to Vegas. And then it's like, but how about this conference room in the basement of a hotel in London? And I'm like, no, but I kind of want the scale of Vegas. And then, well, if you go Vegas, you got to give up this. And suddenly I realize, well, actually, I don't have to choose. I can do Vegas and I can go to Antigua, neither of which were in the original script. And I ended up going to 150 cities around the world over the course of the film because I wanted to do something that captured the excess of cryptocurrency. I wanted to make "Wolf of Wall Street" seem sleepy. And so it started changing the script. Because normally you take a story and you have to condense it down to 40 scenes and X number of locations. I was like, oh, I don't have to do that. And so it changed my entire approach to how you even write a script, because those limitations are no longer there. You can imagine a playwright has to write with this idea of, how do I have everything take place in this living room, because all I have is that set? And filmmakers, you wouldn't think they have limitations, but we really do. And how do we fit it into that box? And I'm like, well, actually, there is no box anymore. The box now is just making sure that everything I do furthers the story. And there's a lot talk about job loss or creation. The reality on "Bitcoin" is that when you suddenly go to 150 different locations, as opposed to 40 in a traditional movie, those locations have to be populated with people and characters. So suddenly you have to cast actors, because that was the thing I wasn't going to replace. I wasn't going to do any computer-generated performances, even though there are some actors out there in the world who computer might actually do a better job than they do. But I wasn't going to do that. I was just going to cast fantastic actors. And so it was causing an issue during prep because my line producer, who was letting me shoot wherever I wanted in the world, because we're creating it, was like, Jesus, you're hiring so many actors in this. There's 125, 140 actors. Can we cut it down to 70? For instance, like "Bourne Identity" has 50. And we didn't. So certainly "Bitcoin" is different than my other movies in terms of the scale and the number of little of small characters. Obviously, Gal, Pete, and Casey are the heart and center of it. MIRA LANE: And then under the hood, maybe, Julina and Jed, talk us through, what is it like to work with leading-edge tech? How did you stitch this together? JULINA TATLOCK: So I was on set with Doug as the AI supervisor on set with the capture. We were also running a team of AI artists within the editing so that we were generating, working with the art department and the production designer, in order to create the environment so that we could then pass stills and frames and to the edit, and they would start to use that generative footage inside the animatic, so that Doug could then see and evaluate how well it was working. And that was our first phase, essentially. And as Jed said, that's a cross-discipline team. They're storytellers from different walks of life, essentially. But we also worked, then, very closely with, obviously, the VFX. There was a VFX team that was doing-- we did 360 capture on set as well, so that we would have people's likenesses. DOUG LIMAN: And the capture on the set was-- my films sort of benefit from whatever I experienced before. Like, "Mr. and Mrs. Smith" is the movie it is because it's a rejection of "Bourne Identity." Because "Bourne Identity" was like, look how cool it is to be a secret agent. And then I read the first draft of the script to "Mr. and Mrs. Smith," and I was like, oh, wait a second. Why are we celebrating James Bond? That guy can't maintain a relationship across two movies. The bigger thing to celebrate would be somebody who's holding a marriage together after seven years and isn't gallivanting around like a secret agent. So "Mr. and Mrs. Smith" is very different because of my experience having made "Bourne Identity." And because I had made "ASTEROID," which was all performance capture, and true performance and motion capture, all we ended up in the editing room was a bunch of data. And I was like, holy fuck, how do we edit this together? It's like, we have to generate shots from that data first, even being able to figure out how to start to edit. And I was like, we can't go through a whole movie like that. So we decided that the process needed to involve us composing shots on set of my actors. There are witness cameras, and in post, I can decide to change it, if I want, change the angle. But I have to commit to something on the set so I have something to edit with. JULINA TATLOCK: And also, in the process of the animatic, some scenes were working or not working in terms of how we were generating the AI and the scene was conceived. And we rewrote and reshot some of the scenes as well, which that's not unusual for a film, if you want to rewrite during that. But that was just really the tip of the iceberg. The performance capture and those days, there were six months of prep. There was 4 and 1/2 weeks of production, that we were in production. So all of that prep, that sprint, that small sprint, and now the hours and hours of post-production. JED WEINTROB: Yeah. No, it's interesting. I mean, look, we're an artist first, technology-driven company. And I like to think we take the best of both worlds. The film production world, I've worked in all my life, and it's a very bi-road-- VFX studios have pipelines, and you have a process of getting from pre-production to production to post, that it's kind of written, to a degree, in stone, as opposed to the technology side of our business, which is agile We are always adapting everything we do to whatever the latest technology that comes out there is. So we have our own proprietary pipeline and workflows that we can make these films with. But that changes. We're not doing it right if we're not changing that every couple of weeks. And what that means is not that we're cutting costs or saving money using AI, it means that in the current realities of making movies and television, creators like Doug are like, I don't have enough to do what I want to do. And we supercharge doing what our creators want to do by bringing them new tools, new technologies, and building a workflow around it that enables us actually to deliver something. And the way that Doug interacts with this technology is going to be very different from the way a lot of other names we can't talk about who you've heard of and worked with and seen a lot are going to engage with this kind of technology. And that's where it's really exciting. And where us also partnering with Google not only getting a little bit of an advanced look at some of these tools and some of the things that Doug's saying now we can do this in six months-- thank you, Gemini Omni and similar tools. That's great. It's also about being able to work with the research teams and work with DeepMind and really learn from each other and say, where we really need to get is here. And sometimes we get there and then we change things, and sometimes we don't get there and we go back to some more traditional filmmaking techniques. So it becomes a process. You can change the film in post as you're going through it. You can iterate and the line between pre-production, production post sort of almost doesn't matter until you get to a deadline, just like with agile software development, where you need to deliver, and you're like, OK, here it is, here's this version of it. MIRA LANE: So we have a few minutes left. And I thought, why don't we wrap up with some advice that you think filmmakers ought to know or have based on your experience and what you've been playing with? So maybe Jed, we'll start you, and we'll end with Doug. JED WEINTROB: I mean, I feel like now the advice is always, just go do it. As a young or old filmmaker-- and I say old filmmaker, because a lot of our team are in addition to the technologists and the AI-native artists we find, are also-- one of the top feature film editors in Brazil is one of our lead creatives because they don't make big budget feature films in Brazil anymore. So he's working with us. And it's not a different career. It's a different kind of career. So we found him because he started making really amazing AI shorts. And Julina judged a festival and found this short and was like, that's amazing. So my advice is always use all of the tools that are out there. Google and other companies like Google have handed you billions of dollars of investment to go tell your own stories now. So go tell them. MIRA LANE: Julina, your advice. JULINA TATLOCK: I mean, not much different than that, except for that I would say, also, I think it's a really interesting idea to figure out what the tools can do really well, but also try to break it and try to make it look untraditional. And how can you use and harness that within the writing and within the story? I mean, there's a sense within the story that some of the art direction, some of it's told essentially back in memory. It's told from the perspective of the present to the past. So where is that? Where can you lean into what AI can do in the special effects and the visual effects of that? But I think, figuring out what the boundaries are so that you can develop your own style, because a model is also difficult to sometimes not make everybody else's and have person in center frame. And there's a couple of quirks that all of the models have, essentially. So break it and find it for yourself. MIRA LANE: I love that. What about you, Doug? Advice for filmmakers here, or even the film industry? DOUG LIMAN: Well, it's almost similar to the advice I used to give just speaking at film schools. And I love sharing-- if someone can learn from the mistakes I've made, that makes me feel better about them. And it would always be like, don't try to copy me or copy some other filmmaker. Chart your own path. If they want someone to make a film that looks just like "Bourne Identity," they'll come to me or whatever. Do your thing. "Swingers" doesn't look like anyone else's movie. So in a way, I would be thinking about giving that same advice now that-- I don't know if there's any screenwriters in the room. But if there are and I said, OK, who here has cheated and tried to ask ChatGPT or Gemini to write a scene for them? And we'd all raise our hand. Yep, hand came up there. And it's terrible. Figure out a way how to use these tools to bring out your humanity, but not to cheat and just let it do it for you. But figure out how to empower and let your humanity come through. And that would be my advice. MIRA LANE: I love that. All right, audience, please join me in thanking Doug, Jed, and Julina. [APPLAUSE] [MUSIC PLAYING]