Latest / Newcomer Pod / 'I Fell in Love With an Algorithm' (w/Cristóbal Valenzuela & Alexis Gay)
Transcript
- 0:05Welcome salad. Hey it's Eric newcomer, welcome
- 0:15to dead cat. Another great episode.
- 0:17I'm excited to invite two people onto the show.
- 0:21We have a Lexus gay who is the host of the podcast
- 0:25non-technical. And I went on her podcast and
- 0:28we've stayed in touch and I Thought, maybe she would enjoy
- 0:31coming on a play. The co-host role on a very sort
- 0:35of technical podcast. So hey, Alexis Erik, thank you
- 0:39so much for having me. I am so ready.
- 0:41I'm ready to be technical. I know you.
- 0:44You you've got metaphors Galore, you've been prepping so I'm
- 0:48excited. Absolutely.
- 0:49Alright and then the the one who gets to you know be subject to
- 0:54the actual questioners here, we've got crystal ball,
- 0:57Valenzuela the CEO of Runway who I actually encountered in one of
- 1:01my early crazy generative AI hype cycles and I thought
- 1:05normally, you know, I DM Founders just like am writing
- 1:08about your company and, you know, they ghost me but you were
- 1:10like oh yeah let's get on the phone and talk about it.
- 1:12I thought we had a great call so then I thought you'd be a fun
- 1:16guests to sort of explain. What exactly is going on in
- 1:20generative AI. So, welcome to the podcast.
- 1:22First, thank you for having me. And I'll do my best to try to
- 1:25explain what's going on. It's crazy, right?
- 1:28What? Just to start off.
- 1:29I mean, can you You just give us some back.
- 1:31Story of how you got into generative Ai and the artificial
- 1:36intelligence world. Like how do you come to this
- 1:38space in the first place? Yeah, for sure.
- 1:41So, it all started back in like 2016 2015, I fall in love with
- 1:47us algorithm and technique called Deep dream which is I
- 1:51would consider like one of the earliest.
- 1:52Like a I are kind of like models that was how there was very
- 1:56neat. She was just like a handful of
- 1:58Artie's and researchers were With it.
- 2:01And I felt too much in love with it algorithm to the point, I
- 2:04just left my job in Chile, got a scholarship, I didn't were you
- 2:07and spent two years, researching kind of like, computational
- 2:11creativity or applying like, AI models for Creative kind of like
- 2:15purposes, and were closed and that's rabbit hole.
- 2:18That's hard to get out of. How does one fall in love with
- 2:21an algorithm? I was going to absolutely not an
- 2:24experience I've ever had. No can't say the same, that's a
- 2:29great question. I think for me was the aesthetic
- 2:31quality in this study, kind of like possibilities that I was
- 2:34saying, and it's kind of like earlier going them.
- 2:36This is again before kind of the wave of outputs that you might
- 2:39have seen today. It was more about that deep.
- 2:42I would say two things. One I fall in love with the idea
- 2:45of using neural network techniques to generate images to
- 2:49generate video, right? To generate the kind of like
- 2:51Explorer and type of aesthetic and creative possibilities.
- 2:54Just wasn't like, even imaginable before a tree like
- 2:58that idea of like exploring Uncharted.
- 3:00It is unimaginable territories and I thought and I saw this
- 3:04early kind of like researchers and our he's kind of like as a
- 3:08way as a first kind of like step towards like just experimenting
- 3:12with us and seeing how we could go.
- 3:14I think I followed a model of that.
- 3:15One single algorithm would be a video of experimenting with
- 3:18technology, right of trying and use it in ways that perhaps
- 3:22weren't meant to be used. Writing you go in a that's the
- 3:24rabbit hole. I feel I fall in love with but
- 3:27long story short, that one thing led to the other and Runway was
- 3:30my to whether my co-founders kind of like a research project,
- 3:33was my thesis at art school that I went to.
- 3:36And the whole idea is basically still, it's basically the same
- 3:40that we have today, which is how do you take these algorithms is
- 3:43generative AI systems and use them in the context of
- 3:47art-making in the context of filmmaking in the context of art
- 3:50or design, just in the consists of creative kind of like
- 3:53processor. And I think what we've seen so
- 3:56far right now, like last year 2020 was like a A year where
- 4:01we're models got really good. Right?
- 4:03And for me, there's first way was perhaps 2015 to 2020 to
- 4:07wear, like some techniques and algorithms, some researchers and
- 4:10Artis were playing with 80 was mostly seen as a niche thing,
- 4:13right? So now, when I remember pitching
- 4:15like and telling people like, hey, here's the thing.
- 4:17You can use to generate an image, the first reaction I got
- 4:20was like, Chris is a toy, right? This is you being an art student
- 4:24doing art gold. That this is just, I will never
- 4:27use this, right? And I think that like this
- 4:29missing tonight, You're like that happens often when it's
- 4:32very early and I think the way that we see now which is I would
- 4:35say 20, 22 onwards and we just started on the second wave is
- 4:39models. Have gotten really good, right?
- 4:41So no, my mom like asked me about like and I saw Seaside
- 4:45understand what I'm doing, right?
- 4:46We just like the outputs of them are just so much, crisp and
- 4:50investable that. I think it changed the
- 4:51understandings of what we can do with it.
- 4:53You said, you went to art school?
- 4:55I did. Yeah.
- 4:56So did you spend a portion of your life considering yourself
- 4:59an artist Do you currently I tried, I tried I did, they
- 5:04started like econ and then like design and also like doubling
- 5:08like software engineering and then spend time just in like art
- 5:12and research. And I quickly realized that I
- 5:15guess my I wasn't really a good artist depressed and you're
- 5:18barred. No, I think my art, I always
- 5:22find excuses to make tools to make art.
- 5:25And then I realized that those tools were far more interesting
- 5:28and useful that the outcomes of those.
- 5:30Right. At least for me, I think I've
- 5:32come to, I come to the realization that perhaps my art
- 5:35is just to making any really like making something that then
- 5:39a very talented artist can take and use in ways that I can't
- 5:42even like, imagine, right. And I think that just, that's a
- 5:45very kind of, like, interesting position debate and in terms of
- 5:48our Runway, is specifically, like a video tool is that right?
- 5:52Or am I thinking about it in a, to limited way?
- 5:54I mean, it's helping introduce sort of these generative AI
- 5:58tools into video making. Is that right?
- 6:00Yeah, definitely. I mean Runway you can think
- 6:02about as a company. First of all we are unemployed
- 6:06AI research company. So we we develop and build kind
- 6:09of like low level, the basic fundamental algorithms for
- 6:13Content generation and it mean content because and work with
- 6:16images with video, we just multimedia formats, right?
- 6:19So we have a research team that sits at the core of the company
- 6:22of a product that researching drives innovates and develops
- 6:27novel techniques. For kind of like driving
- 6:29generation of Content. And then we will products that
- 6:32the leverage that sort of like algorithms and techniques.
- 6:35And there's things you can do on the video side.
- 6:37So we have films and post production, companies and
- 6:41broadcasting companies, and Studios, and just teams in
- 6:43journal, working on the video side of things leveraging and
- 6:46using our kind of algorithms and products and we also have a set
- 6:50of like image base Solutions and products as well.
- 6:53Do you have a dolly or like a stable?
- 6:56Do you have sort of an image Creator?
- 6:58This public. Yeah.
- 6:59But think of me, Stable diffusion, we run.
- 7:01I know you're part of that. I'm going to talk about it.
- 7:03But do you have your this Runway, the company have a
- 7:05separate sort of image product? Yeah, we have our own 30
- 7:10different tools, right? And every tool uses a different
- 7:14algorithm where I could be nation of different algorithms
- 7:16to either generate a video edit, a video generate an image or
- 7:20translate from an image to a video, right?
- 7:22And so what's interesting for us is that we're not for our
- 7:25customers, we don't you get access to all these models
- 7:27behind the scenes, but you are not necessarily Kind of like
- 7:30aware of those models, right? So, from a filmmakers
- 7:34perspective, if you're in our director of your designer, if
- 7:36you're editing the video, you don't have any study in care or
- 7:40have to know about, like the complicated like, ins and outs
- 7:43of these models and the techniques and the names, and
- 7:45all this in the same way that if you use Photoshop, you don't
- 7:48really care about the blurring function or the algorithm that
- 7:50allows you to blur an image. We just use the image right, you
- 7:53just use the tool. So for us, it's that parameter,
- 7:56it's like it's about human intention.
- 7:57It's about you coming to the tool.
- 8:00You working with this algorithms in very intuitive ways to the
- 8:04pain. You forget.
- 8:04You're working with this kind of like, quote, a I register you
- 8:07work with that tool that helps. You achieve something bring
- 8:09particular but behind the scenes.
- 8:11Yes, we have like those models that basically just power the
- 8:14whole set of features and products that we have my
- 8:16understanding. Correct me if I'm wrong is that
- 8:19sort of generative, a, I really blew up in the public
- 8:22Consciousness. When stable, diffusion an open
- 8:25source project with ties to stability AI a company.
- 8:30You and your co-founder, I believe that Runway was one of
- 8:33the head researchers on that project and then that project
- 8:37became open to the public right before that, you know, these
- 8:41projects had all been sort of behind the scenes open a.i.,
- 8:44despite the name hadn't released, there's and then
- 8:48stable, diffusion comes out, and sort of the world goes crazy,
- 8:50open, a.i. puts out Dolly, am I getting that right?
- 8:54And where do you sort of fit in to that project?
- 8:57And how much are you involved with?
- 8:58Yeah, I think. Actually you're right.
- 9:00I think so. It might look like there's an
- 9:03overnight success over a specific model that lets say the
- 9:06space has gotten like so much attention.
- 9:07The last couple of months right in might look like, oh something
- 9:10happened in a specific, like they are mouth.
- 9:11But the truth is that the research has been brewing for
- 9:15like years, right? We working on runway for the
- 9:17last four years and he takes a lot of like iterations on
- 9:20learning and training and learning a lot from like what
- 9:23works, what doesn't work, what data sets you can use what
- 9:26models. Can you train?
- 9:27What are the possibilities of that, right?
- 9:29So the The actual code and model and paper that powers and the
- 9:34fundamental architecture behind Civil Division was was released
- 9:37actually like almost two years ago, right?
- 9:39And that's open source. It's been open source.
- 9:40Since we publish it that work is the collaboration of Runway and
- 9:45Ellen do you mean which is a research organization in
- 9:47Germany? And we do this a lot, we
- 9:48collaborate with other researchers, the Patrick Sr,
- 9:51who's our research, principal research scientist lead, those
- 9:54efforts, right? And we publish that code, an
- 9:57open source it for a conference that was like, released on 20
- 9:59lat. 20 21. And then seems then that was
- 10:02actually called late and efficient.
- 10:04I was out late and efficient work, right?
- 10:05It's diffusion is a technique and Layton is like a word to
- 10:08describe like a the space of like possibilities in like am an
- 10:12Nei so we basically call it later.
- 10:14Diffusion, there's a technical term that's actually longer than
- 10:17that. It's high resolution image
- 10:19synthesis with late in diffusion models.
- 10:21Sure slate and very good consumer facing so that yeah you
- 10:26can see and then we treat him on the model we first.
- 10:30In the first model. And then each rated on a more,
- 10:32he took training on more data sets for longer and then a
- 10:35company called stability, kind of like offer compute, basically
- 10:38came in as an infrastructure and just say, hey guys, I saw your
- 10:41open source project, I'll just throw in some computer and you
- 10:44guys can train it on a larger data set.
- 10:46And that's that basically was were simple, diffusion was born.
- 10:52And then there was, you know, I love drama.
- 10:53There was a little drama when you updated you guys updated
- 10:57that software, right? And there was a little back and
- 10:59forth, With about, whether you are allowed or not, or what was
- 11:02that whole thing. Yeah, there was I think there
- 11:07was some confusion to Venice again.
- 11:09We published that we're going most two years ago and it's been
- 11:11open source. We've been improving 8 and
- 11:14somebody donated computer. And I think that had an internal
- 11:16like confusion or conflict about like ownership that they kind of
- 11:21like apologize about both publicly and privately.
- 11:23And I think we just like steer and unnecessary conversation for
- 11:26us, it's like we've been working on this for love a lot of time
- 11:29for a lot of years. And we publish newspapers, we
- 11:32have will continue to upgrade our During diffusion work and we
- 11:35have already like new models that are going to like change
- 11:38the landscape. If table division was like
- 11:40inflection point like q1 if - this year's going to be like
- 11:44probably ask you guys that one, but I think I'm blessed answer
- 11:47but like the drama to be honest, I know people like like the
- 11:49drama, but I think it's just like a instrument confusion.
- 11:51I'm like, I'm glad that something like this.
- 11:54Yeah, I'm glad I'm glad it's figure out like that.
- 11:57I know. I think it's just like, God,
- 12:00this Stakes here are just and this is classic, you know, we
- 12:03experience An open source software projects before, but
- 12:06it's who owns what and the stakes feel bigger with AI
- 12:11because it's being presented. Is this sort of you know,
- 12:14everybody's gonna lose their job or well you know like it's going
- 12:17to change the world. And so I guess my question is
- 12:20just like yeah like all these new you know, when when open
- 12:23a.i. comes out with GPT, for when you update your software,
- 12:26new versions of like Sable diffusion, how much of those
- 12:29projects will be in the public domain where any Company can use
- 12:34them and how much are they going to be owned by particular people
- 12:38because my understanding with Runway was that?
- 12:40Like part of how you were trying to differentiate was sort of
- 12:42like the tooling and making sure you can apply the publicly
- 12:48available. Artificial intelligence
- 12:50technology to specific use cases or are you also trying to own
- 12:55proprietary algorithms? I think every company might have
- 12:59a different strategy and a different like Focus.
- 13:01I think for us something with we've defined barrier.
- 13:03On that has taken us some time to like, really nail down has
- 13:07been disabled of owning your stack, right?
- 13:10So we don't build our products on third-party like models for
- 13:14software data sets, like we tried to build the whole thing
- 13:17and to end. And the truth is that that's
- 13:20that's hard because it requires you to really nail or understand
- 13:24how to build research, how to build infrastructure, how to do
- 13:26products, right? The advantage of that is you've
- 13:28build that muscle and you trade enough is that you can change
- 13:31any part of the stack, right? So if I want to change.
- 13:34You want to improve a model because specific customer wants
- 13:37a better version of whatever we have.
- 13:38I don't have to ask someone else or I don't have to wait for
- 13:41someone else to do it. I know how to do it and I can go
- 13:44there and like do it, right? I can go to the lowest level
- 13:47possible have a model is trained.
- 13:49How mall is optimized other markets?
- 13:51Deploy and then go back to the surface level and change that
- 13:54for them? Right.
- 13:55I think that's to your, my proprietary algorithms that
- 13:58you've built. Yeah.
- 14:00Of course we have with that, we have custom research and Open
- 14:04source, some research. So he built our research
- 14:05website. You'll see that the latent
- 14:07efficient, the Civil Division work was something.
- 14:09We open sirs. There's other work that we have
- 14:11an open source yet will perhaps oppressors at some point.
- 14:14It's like but it's a combination of both.
- 14:16I think our open source strategy so we start to community
- 14:20building like the the wave of creativity that emerge.
- 14:23After releasing stable diffusion was just like really, really
- 14:27incredible to see and that just helps Drive the field forward
- 14:30which is exactly what I think happened.
- 14:32There was like an inflection point It's a great great of
- 14:35hiring as well. You attract talent and your
- 14:37truck researchers who want to work with you and like, people
- 14:39read the papers and people read the author's, and you go back
- 14:41and kind of like, we get a lot of, like, really tiny delegation
- 14:45from that. But we're not our kind of like,
- 14:47approach perhaps different from other companies is that we're
- 14:49not an open search company. Like, our product is not open
- 14:52sourcing products, right? Or products to help creatives.
- 14:56You use these techniques in comprehensive ways in their
- 14:58day-to-day lives crystal ball. I have some questions about
- 15:01that. So one good thing.
- 15:04I'm curious about you just use the term wave of creativity said
- 15:07after these tools came out, there was a wave of creativity.
- 15:10I largely saw on Twitter, people typing like bird flying over a
- 15:17truck, that's on fire and also elon's driving the truck.
- 15:21You know what I mean? So I my question is when you say
- 15:23wave of creativity using these tools, like can you help us
- 15:26understand a little bit about what kinds of things were
- 15:28happening? Using the tools?
- 15:30Yeah, for sure. I think every new technology and
- 15:34Enables us to think about our creative in different ways,
- 15:37right? But like the ability for us to
- 15:38type something and generated image, just like it's novel,
- 15:41it's new, its we've never seen it before, like some people were
- 15:44experimenting with it 10 years ago but now it's just you open a
- 15:47browser and just type something, you get something out, right?
- 15:50And I feel like there's a lot of times where dislike novelty
- 15:54component of it is just trying new things.
- 15:56You just tried to we're distant possible, reduced, Sherman,
- 15:59alexie you weren't exactly what you're saying.
- 16:01You want to see how far you can go, right?
- 16:04Andre, like that's like a Curiosity Company.
- 16:07Like, everyone is trying to break that systems.
- 16:09Like, okay. And when I like really try
- 16:10something like really hard, right?
- 16:12And I think that's like when you realize you're dreaming and
- 16:15you're like, oh my God, I'm Dreaming right now.
- 16:17What should I try? And I think that's the, that's
- 16:21the 100, a good that sticks fermentation finger to curiosity
- 16:24face. You're like Justin, just okay,
- 16:26there's a experimenting with it, right?
- 16:28You're like this is new. This is different.
- 16:30I want to understand that. What do you mean?
- 16:32But I think true creativity. Edie comes after that, right?
- 16:35After you've like, okay, that's has settled down a little bit
- 16:39and you start experimenting with seriously, right?
- 16:41You start to think about it. How you might incorporate it in
- 16:44like a filmmaking process, right?
- 16:46And if you talk to filmmaker for an art, director is not about
- 16:48the craziest like prompt that you can generate.
- 16:51Something is a lot about like control.
- 16:53Like, I want to generate a variation of an image.
- 16:55I want to have I want to change the color or the laugh and a
- 16:58quarter over the right. How do I do it at scale?
- 17:00Can I do it? Like in 1,000, like iterations
- 17:02per second like Those are different kind of like
- 17:05components. And for me, something I hear a
- 17:08lot about is this idea that AI or like the systems are
- 17:11automating creativity. Right?
- 17:13I think for me, so curious about that but I'm a comedian by the
- 17:16way for contacts. I'm a comedian and I have a
- 17:18podcast. But prior to that, I worked in
- 17:20tech for seven years, just like a classic career path, as, you
- 17:24know. But anyway, so I have a lot of
- 17:25fun. I have a lot of thoughts and
- 17:28questions about this idea of a, I supported creativity or a I
- 17:32supported during I say content or a I supported art and all
- 17:37kinds of implications of that and so I would love to hear a
- 17:39little bit more. How much riding Alexis have you
- 17:42tried to write jokes with, okay, so here I tried to what did I
- 17:47travel? I saw the, I saw a i generated
- 17:51versions of my own tweets, that's what I saw using some
- 17:55software that someone sent me, which I'll have to look up the
- 17:57name and it was terrifying because they sounded like me,
- 18:02mine would be like, weird. Our fights and like, I would
- 18:05reside article just what I need to look at it and I generated
- 18:09version of Arabic a tweet, just be a bunch of fake Scoops.
- 18:12I need to look. Yeah, that's so good.
- 18:15Yeah, so I love this idea though.
- 18:16That it's a phase you're saying right now, or maybe perhaps when
- 18:19it first came out, we were in the Curiosity phase and that
- 18:22implies, you said true creativity.
- 18:23Comes after that. So what do you think that means?
- 18:26I think for me, creativity is a state of mind.
- 18:30It's a way of looking at the world, right?
- 18:31It's not at all. It's not our process, right?
- 18:34And I think what you can truly automate it is the processes of
- 18:37executing a creative idea, right?
- 18:39I think you can you can go back to like like previous moment in
- 18:42time or like there was a major technology like change and
- 18:46understand the changes that that transfer that broad to
- 18:49creativity, right? So I always go back to like one
- 18:52of the earliest examples of technology changing art and
- 18:55creativity and how we look at the world, which was before,
- 18:58like, the 1700s painting was like, mostly the realm and the
- 19:02space and the possibilities. Realities of like folks who have
- 19:05the money, the resources and the time and the knowledge to create
- 19:09these very complex pigments, right?
- 19:11And creating a pigment was a very sophisticated thing.
- 19:13It has like this history and tradition, you had to like her.
- 19:16These Masters that were very expensive and you have to like,
- 19:18mix these very obscure colors, right?
- 19:21And then storing the pigments with like, you had to use a pig
- 19:23bladder and Link silly with a string.
- 19:25Grant is very like obscure, right?
- 19:28And then 1800's, someone basically, John Brown's inventor
- 19:32came up with the collapsible pain too.
- 19:34Right. Just that's the piece of
- 19:35technology, right? You can take colors and put them
- 19:38in like a pain tube. And the most important thing is
- 19:42that you can take that paint tube out into the world, right?
- 19:45And so impressionism was born, right?
- 19:47You had people going outside and having like canvases like small
- 19:50canvases, you don't require like a whole studio to just paint,
- 19:54right? And so what that meant is that
- 19:56you start looking at the world in a very thing way, people
- 19:58started painting know just like portraits of like the royalty
- 20:01but like mountains and trees and like so.
- 20:04She'll like Advanced, right? And like go this career TV to
- 20:07like the merchant. That's what I mean by that.
- 20:09Gertie the opportunity that's like your that's a piece of
- 20:11technology that enables you to look at the world and express
- 20:15your view of the world in a very different way that everything
- 20:18you've used before, right? And fast forward, it's like that
- 20:21the naval huge revolution in art and like the transfer ever
- 20:24transform, how we think about painting art, it set up, right?
- 20:26And you can think about very similar moments in time where
- 20:29technology is like like that have changed, right art and
- 20:32creativity and Photography is another one right photography
- 20:36change. We also get the world and a lot
- 20:38of the times early on the further, graffiti kind of like
- 20:41wave. People said that like
- 20:44photography was kind of like the and of hard, right?
- 20:46You, it's gonna be the end of like paintings, right?
- 20:49Like you, I never just like, and the truth is, it's not.
- 20:51It's just a different medium, right?
- 20:53Just need to explore that beat him in a way.
- 20:55There's new artists that will emerge from that.
- 20:57New ideas will come after that, right?
- 20:59And you open the door for something that's the business is
- 21:03very hard to. Predict, right?
- 21:04Because you tried to predict something.
- 21:06You haven't even yet, experimented with red.
- 21:09And it's very hard to like make this set of assumptions over
- 21:11time about everything that needs to happen.
- 21:13I think I come to embrace that like uncertainty, as in like
- 21:17early on something that's going to massively transform, a lot of
- 21:20things, and we should embrace this like just experimentation
- 21:23phase like that Curiosity face, right?
- 21:25And the kind of things that you'll build after that, as yet
- 21:28still to be like uncover and discovered and that for me is
- 21:30really, really exciting. I mean I have such conflicting.
- 21:34In thoughts on this, like, on the one hand so many like
- 21:37amateur visual artists to me, like, what they enjoy is like
- 21:42the physical process over the output, right?
- 21:46And even if they can get better outputs from a sort of an AI
- 21:51system is that cutting into some of the enjoyment, I get, the
- 21:55digital are also in photography and everything is cut into that,
- 21:58and people get their choices on the other hand.
- 22:01I'm like, so excited by the idea of like Pairing the human mind
- 22:06with like another sort of intelligence or I mean there's
- 22:09such there. I feel like you know, Alexis is
- 22:13wearing a chest sweater. It's like if you get good enough
- 22:15at chess you feel like a certain connect.
- 22:18You know, you can use these Bots to like, figure out how you
- 22:21should play better. There they're like these, I feel
- 22:23like they're already experienced this today where you can sort of
- 22:26feel in sync with like a robot system or I play a lot of bridge
- 22:30and my partner, and Bridge online is is a robot, I sort of
- 22:34and what they do. And so having sort of this
- 22:36partnership with some sort of intelligence is, you know, a
- 22:41very Charming experience. But yeah, it's definitely a
- 22:46terrifying future. Are you worried about it at all?
- 22:48Like, do you ever wear? It's just inevitable to you.
- 22:53Yeah, I'm not worried a lot. I think actually, it's a very
- 22:55exciting time to be in a very optimistic about it.
- 22:57I think like radical technologists.
- 22:59Like, this are, like, one of the things I did do is I deleted the
- 23:03cause And enable things to be more convenient, more easy to
- 23:06use, right? So think about like filmmaking,
- 23:09the people who had access to have, like, professional fee,
- 23:10making tools, even nowadays is like a handful of people that
- 23:13know how to like a date on professional software, right?
- 23:16And I will dislike complicated. VFX work flows and need to
- 23:19understand computer graphics and a bunch of things picture.
- 23:22Having everyone with like ilm or like, bfx professional like
- 23:27skills, right? It's a wave of like,
- 23:29storytelling in ways that we haven't even thought of before.
- 23:33It's just like right around. The corner?
- 23:34What kind of tools we need for that?
- 23:36Well, let's just build them, I don't know yet.
- 23:38I'm more excited. I feel like more excited about
- 23:40the outcomes of those and enabling more people to express
- 23:43themselves in creative ways. And I think that's ultimately
- 23:45the role of Technology like, at large.
- 23:47I think I'm less interested in like me, save you again, that
- 23:51you will automate it, and that we come like, some sort of like,
- 23:53one click of solution and like everything will become the same.
- 23:56I think that's just boring and like, it doesn't really like
- 23:59interesting yet, to be honest. And it's also like a partial
- 24:02view of how the world actually works.
- 24:03I think. I think there's always this
- 24:05human component on that that what gives us like meeting at,
- 24:08Alexis you're saying like, we someone just generates tweets
- 24:10that are, like, just actually there's a I've heard there's a
- 24:13generator podcast, right? So you can generate the script
- 24:16and you can generate the voice. And you have those Supply
- 24:18conversation between like Steve Jobs and like Lex Friedman,
- 24:21right? This is M.
- 24:22Everything is generated, right? And if an interesting it's like,
- 24:25yeah, I mean, works very well, you're convinced halfway through
- 24:27this is actually happening, but I feel like there's still a lot
- 24:30of potential of using that in way, more creative ways and just
- 24:33generating the Entire thing, like n to n, right?
- 24:35I hear what you're saying, and I agree with a lot of at night,
- 24:38especially the point about the way, access to pieces of
- 24:42Technology can really draw out. Someone's creativity.
- 24:45I think my personal experience is a good example of that where
- 24:49I was able to teach myself to edit video because I movie came
- 24:53on my Mac, right? And then I know how to use
- 24:55YouTube and then I graduated to other software but it was all
- 24:58accessible, it was all just downloading thing.
- 25:00Look at it. How to do this some
- 25:02twenty-year-old, dunas basement told me on the And there's a
- 25:04great I know how to make videos now.
- 25:06But here's what's different to me about generative Ai.
- 25:10And some of the other tools that have preceded it and assisting
- 25:13creativity, which is that generative AI models.
- 25:16And please correct me if I'm using the wrong terminology,
- 25:18they have to be trained on something, right?
- 25:21And so when we're talking about visual art or God forbid comedy
- 25:26or anything else, do we have to feed it, existing artists work
- 25:31in order to teach it how to do it and does not raise him.
- 25:34Implications around ownership credit and profit participation
- 25:37of the output of the AI. Yeah, totally.
- 25:39I think those are very valid questions and things that are
- 25:42really worth and fasting ourselves as we develop these
- 25:45new technology. And and again, I'll go back to
- 25:46like premiums in time. There's this video and this
- 25:49recording of 35 years ago. Same questions being asked when
- 25:53Photoshop was first released, right?
- 25:55See ya. It's the same conversation older
- 25:57show you guys the same condition, right?
- 25:58Was original. I was like, this is an
- 26:01interesting question. But it's but it's a valid
- 26:05question. It's part of the commits
- 26:06perfect. And there's, yeah, it is right?
- 26:08And and and the interesting part is like they were discussing of
- 26:11the time. It's like, are you allowed to
- 26:13edit a photograph, right? A photograph is like the truth,
- 26:17right? You can edit it, right?
- 26:18You can't go into National Geographic.
- 26:20Going to like TV magazine and edit any much, right?
- 26:22And the half of the panel was like, it's just a little, this
- 26:25technology should be banned, right?
- 26:26And it can imagine now, like not having like a tool like
- 26:29Photoshop in our disposal. So obvious, and like, Photoshop
- 26:32became a bird. Right.
- 26:34And so for sure there's a lot of things that you need to figure
- 26:36out like Photoshop had to figure out a lot of things and kind of
- 26:39like concerns about how to use the technology.
- 26:42People were just need to like for like fake bills from the
- 26:45like, create fake money, right? So if you try to open a photo
- 26:48shot bill in Photoshop, Photoshop will like prevent you
- 26:50from doing that, right? You can't just like it's your
- 26:53battery. And so you build you build over
- 26:56time ways of like, secure in the system, right?
- 26:59But the 99% of the outcomes that are emerging from this They're
- 27:03going to be net, positive are going to make Society new for
- 27:0699%. I'm sorry.
- 27:08I just want to go back. So you I'm just you on the
- 27:11record just said that 99% of the outcomes.
- 27:14That's a, don't you guys go? I think that's just a Boulevard
- 27:20but I think. Yeah.
- 27:21Most of the I would say that old put that on a slide somewhere.
- 27:25Oh, of course. I mean, it's a, it's a very
- 27:28powerful technology. It's going to be as massive as
- 27:30cell phones were like. I feel like mobile was like,
- 27:33Really net positive for everyone, right?
- 27:35Like you're able to hold like a lot of knowledge in your pocket
- 27:39and able to connect with anyone in the world, right?
- 27:41That's just think that's not positive.
- 27:43I think a eyes and gently vir similar Technologies, right?
- 27:46Your there's of course like people they're going to try,
- 27:48they're bad people in the world and then you try to use
- 27:50everything they kind to do bad things for sure and like will be
- 27:53contrary to prevent those but overall the outcomes of the
- 27:56Technologies net positive like well we're allowed to
- 27:59democratization of content. For example, as you were saying
- 28:02of anyone having Access to the sings just like the to their
- 28:05magical change. Yeah but I think to your point
- 28:07there is there's a lot of conversation that needs to be
- 28:09had and people are having around like data set and training the
- 28:13models and one which data is and like how you do it.
- 28:15But again like Google had the same kind of like discussions
- 28:19early on right? I don't want my data to be
- 28:21constrained by by Google, right. Kind of have and kind of have
- 28:25like a waste opting out if you don't have one.
- 28:27I have my website being like searchable and the answer is it
- 28:30you can like just put this file on your website and I Google not
- 28:34scrape. It and I would not index you in
- 28:36the search file in the search kind of like options.
- 28:39And so there's I think there's still a lot of things yet to be
- 28:41uncovered and models are not 100% like ready yet to be like
- 28:46using all sort of like professional environments were
- 28:49still very very early. How important do you think the
- 28:51humans are to generative AI or will be in the next couple
- 28:55years? What we're talking about is
- 28:58human sort of as creators experimenting using this tool,
- 29:02but I another conversation On this podcast, we've talked about
- 29:05Tick-Tock and you know, how sort of the content sorting algorithm
- 29:09is so good. And, you know, there was this
- 29:12question asked of like, how far away are we from Tick-Tock?
- 29:15Just saying, oh, you know, not only do I can guess which videos
- 29:20that exists do people want. But I can just create the video
- 29:23that you are likely to want. Like, how far away do you think
- 29:26that world is where somebody like Tick Tock, can just deliver
- 29:29the video that I want without a human involved at all over.
- 29:33I'm very, very close. I think we're not far away from
- 29:36having laughing. Good, I'm not like, Oh, yay.
- 29:42You saying that makes me scared. That does not sound fun to you.
- 29:46Sir we I don't want to commit to specific data or moment but I
- 29:50think like we're heading towards her world in general.
- 29:53When a lord of the content of your consumer line will be
- 29:55generated picture like a YouTube, like generated like
- 29:58stream, right? You can do Netflix generated
- 30:00content like Tick-Tock generated content, right?
- 30:02I think that's Somehow visible like today, if you combine a
- 30:06bunch of different things but there's that's still a lot of
- 30:08things to be developed to get to a point where like you'll do it
- 30:10real time. I think it's under happened, if
- 30:13it works, if it's possible. Any further comments and allow
- 30:17us to explore new avenues of like creativity.
- 30:19I think that's that's going to be great.
- 30:21I think a lot has to do yet with developing this bottles in more
- 30:24safe, Miners and more kind of like a line ways to our human
- 30:28intention. And there's a lot of work to be
- 30:30done still yet there to prevent possible.
- 30:32Misuse is for sure. Sure, but overall, I think we're
- 30:35going to be in a time in moment, where like you're going to be in
- 30:38every single movie you ever wanted to be right, you are
- 30:40going to be, you're going to eat them.
- 30:42A nice thing about it is not only that it can create
- 30:45something but it can create something specific to me that I
- 30:49will. Yeah, and compelling.
- 30:51Yeah, 100% 100%. And but and we you think this is
- 30:54like three laying out like, yeah.
- 30:56What's your soon? I'm like which how do I need to
- 31:00prepare my job security? Like what does my timeline Need
- 31:03to be to learn to code. I just need a sense for when I'm
- 31:06going to be at replaced by G, PT 3.
- 31:09I think very soon, a couple of years.
- 31:11I would say yeah that's good. Your therefore there's
- 31:15definitely like an exponential like like progress tray that you
- 31:20can see and perceive more clearly.
- 31:21Now that like here's the progress of what Take Years with
- 31:25two gears, for progress is now taking like months, right?
- 31:28Yeah. And we'll what will start to
- 31:30take months with Senators start to take like weeks, right?
- 31:32And you go by. That like all the time like
- 31:35large language models and now going to be writing like papers,
- 31:38I'm going to be writing code, I'm going to be and you started
- 31:40accumulating. The amount of progress that you
- 31:42can make, right? And I think that's yet, it's
- 31:44happening and it will continue to have him, but I think that's
- 31:46going to happen like sooner than we think.
- 31:47I mean I respect you for doing your best.
- 31:50Not to give us a date because you know I covered the
- 31:52self-driving car industry. Any of that was an artificial
- 31:56intelligence powered industry. That always said you know full
- 32:00autonomy is just over the horizon.
- 32:02Like what do we take From that experience.
- 32:04Why is this different than that? I mean, things.
- 32:07People have gotten excited about chatbots before things can
- 32:09appear, you know, if you solve 90% of the problem, but you need
- 32:13to solve the whole hundred percent, it can be easy to
- 32:16convince yourself that you're like almost there.
- 32:18But the last ten percent can be can be extremely difficult to
- 32:21Impossible. I think it depends on what your
- 32:24goal is and what you're trying to accomplish writings I guess
- 32:26in the case of self-driving cars like full autonomy, right, can
- 32:29you drive a car in any Street in the world and like with no?
- 32:33Get right? That's that's the end goal right
- 32:36for us. It runways not.
- 32:37I can you make a piece of art or can you make a video with no
- 32:40human intention and no control at all?
- 32:42Right, that don't think that that's not the goal.
- 32:43Their goal is like, the goal is like, can you take an idea and
- 32:48execute that idea in the fastest way possible, right?
- 32:51And right now if you want to create a video like Alexis you
- 32:54want to make a video 20 30 years ago you would have taken your
- 32:58like months, right? You got to like rental equipment
- 33:01and then like and have someone have been?
- 33:03Cost prohibitive as well as yeah.
- 33:05It's too expensive, right? And so your idea is like, I want
- 33:07to make comedy. Want to make this, you should
- 33:09you hand. I mean, usually it's when I take
- 33:11you months and you need to find a producer.
- 33:12I like all this. Yeah.
- 33:13And the fact it's like technology and like smartphones
- 33:16on the internet. God you like 90% there, there's
- 33:18still, of course work. You have to do like coordinating
- 33:21on recording and editing, Etc, but like it's made, like 90% of
- 33:25the work, right? You were like, so it's so much
- 33:27better so much visible even to even consider doing something
- 33:30like that. So if the goal for us is like
- 33:33to, Help create. It's just like, get over that
- 33:36like 80 90 percent of the work that you don't want to do as a
- 33:38creative. Maybe just like no one wants to
- 33:40like spend time. For example, searching through
- 33:43hundreds of videos and then copying those videos and then
- 33:45placing them to express the idea that you want to have in your
- 33:48head. It would be great if you can
- 33:50just like how to make that, right?
- 33:52And that's like that's 90% of the process, right?
- 33:54And I think that the goal for us is you measure by that?
- 33:57I think we're very close to getting to that point, right?
- 33:59And that's the ultimate goal. I would serve dirty eyes like
- 34:02taking the cause of Ten creation down to zero, not the cause of
- 34:05like, ideas, right? IDs are still like ideas and
- 34:08like the best ideas will still win, right?
- 34:10I think this is the time where everyone will become more of an
- 34:14editor and curator and like AV has the best ideas will be
- 34:17executed, not because they have the funding on the resources
- 34:20just because they're the best ideas.
- 34:22I will bet we have to we're playing sort of the cynics even
- 34:24though I'm extremely excited about this technology but you're
- 34:27doing a good job of making the bull case.
- 34:29Do you worry that we're going to have even more of like a
- 34:32bullshit? Mmm, in the world of social
- 34:34media, there's already been the problem.
- 34:36You know, if you have a big account, you have people writing
- 34:38to you and just the work of like, sorting through among
- 34:41humans. What's like a reasonable
- 34:43critique to spend time an electrically and engage with and
- 34:46what sort of a waste of your time is a taxing exercise.
- 34:49Even if you're like, sort of a with it person and then if you
- 34:53add to the mix, I literally on Twitter.
- 34:55Had someone, I'm 90% sure, use GPT as like an auto response to
- 35:00a bunch of tweets to like, juice engagement.
- 35:03I responded like GPT and they send me a smiley face and it was
- 35:07obvious to me. But it's still like, oh my God.
- 35:10If I'm gonna run into that all the time where I'm going to have
- 35:12to engage with the cognitive task of is this gbt.
- 35:15And of course, the problem with it being GPT is G, PT or
- 35:19whatever text text tool, they're sort of bullshitters.
- 35:22They're very they're 100% confidence and maybe 90%
- 35:26accurate and that's very challenging for human beings.
- 35:29Yes. What do you say to the
- 35:31bullshitter problem of AI? That's a good question.
- 35:35I think. Overall I'm optimistic that that
- 35:37we get solved in the same way that like, spam and like
- 35:40scammers got solving like mail. And in the Internet it's like if
- 35:45you can receive an email from like, a prince of Nigeria,
- 35:47you're like, yeah, this is, this is right?
- 35:50And this is like a very common like thinning internet.
- 35:52I like you build filters and you've been making its into the
- 35:54tech bad and like people lab figure out and governors.
- 35:58Like there's ways of like improving those systems to like
- 36:01avoid that right now. What's happening Bending, is
- 36:03that it's still very early, right?
- 36:05Like, if you consider like chop, GPD tragedy was like a research
- 36:09release, right? It wasn't like a product, right?
- 36:11And it was just like, right the day when taken by surprise how
- 36:13popular was a, all right? Yeah, it in the bottom.
- 36:17It says like this is our research released, right?
- 36:18There's a lot of things that you have in yet consider, but just
- 36:22trying to learn how people use it, the possible, like things
- 36:25that need to be fixed. And I think that a big
- 36:27misconception is to look at this and be like, okay this this is
- 36:29it this is the final thing will use ever and it doesn't work.
- 36:33Right? I'm like, no, it's just like a
- 36:34point in progress and time and I will continue to improve it
- 36:38until our point where like, all that bullshit that's like you
- 36:41will find and they say things will be like, either reduce or
- 36:44remove or like prevent it and still a lot to be to be done for
- 36:48sure. I think that there is cool
- 36:50stuff. I agree with a lot of what you
- 36:52said around taking the burden of some of the more like task task
- 36:58pieces of the creative development process.
- 37:00I think that there's a lot a lot that could be good there.
- 37:03I'm just scared. That's why I'm scared at all.
- 37:08We've talked a little bit about like, oh, eventually people will
- 37:11figure this out, and I'm sure that like things will happen
- 37:14down the line. I'm kind of like this seems to
- 37:16be going awfully fast, and I don't know.
- 37:19Who is ultimately responsible? Like who's the adult in the room
- 37:21here? That supposed to put a hand up
- 37:23if you like, we should probably have some rules socialist
- 37:26realism to it. I think so much of what happens
- 37:28is just if it's possible, it will happen and that's how I
- 37:31yeah I think sort of Of assumption.
- 37:34Underlying some of this conversation.
- 37:35Is that like the learn to code crowd is going to win the word
- 37:40cells. Writers Among Us are the ones
- 37:42who are going to be screwed. Do you actually agree with that
- 37:45premise? I mean, we're seeing these like
- 37:48programs, do some coding or like, is it clear to you like
- 37:52which human skill sets? Come out on top.
- 37:55If this all happens, is you see it?
- 37:58And which ones are the losers? Yeah, that's a good question.
- 38:01I think I don't have a full answer.
- 38:03Whereas because I think a lot has the yet be like, discover,
- 38:06and like understood for like World models can actually go.
- 38:10I think, with every kind of like previous moment in time in the
- 38:13past for something, like I said happen.
- 38:15Like there's like jobs and disciplines unlike things to
- 38:19just disappear, right? Like you don't need the the
- 38:22market for like people could do analog film editing has like
- 38:26possibly being reduced, right? No one is just cutting films
- 38:29like with Caesars anymore, right?
- 38:30You can't find these people, right?
- 38:32And the reason is that those folks have to adjust their jobs
- 38:35and like, there's has two for a diesel age and they have to
- 38:38learn a new task and understand the limits, and the kind of like
- 38:41all directions of using that I think.
- 38:44Overall, my hunch that we'll start to see that like
- 38:46professions and jobs will radically change right software
- 38:50engineer for, I think will become more about again, having
- 38:52an idea and helping a system like execute the idea in a
- 38:55secure way, right? So if you want to write right
- 38:58now you can just like something like copilot and just write a
- 39:01function and have like the copilot like Complete the whole
- 39:04function for you, right? So you're not going into the
- 39:06Commendation to try to understand apis and then going
- 39:09back just like it works really really well, right.
- 39:11But you still need to give guidance and you need to the
- 39:13human need to understand your intention, right?
- 39:15And you're like having the system just help you along the
- 39:18way on that. Right?
- 39:19So I think that that not sure like software Engineers are like
- 39:23job itself is like not. It's still going to be here,
- 39:26right? Just radically going to be
- 39:27different. It's going to be sort of
- 39:28function differently. It's gonna skills that might
- 39:31require going to be different. You might not need to No, this
- 39:33very obscure like the levels of, like, API documentation because
- 39:37now you have something that can do it for you.
- 39:39And for writers, I think it's the same, it's like pretty,
- 39:40like, think about like, how like computer.
- 39:43So you can typewriters change. The process of writing, you are
- 39:46able to like Grace things easier, they're able to like
- 39:49compose ideas more quickly. You're able to Like Remix thing
- 39:51because you have them saved in different ways, right?
- 39:53I think that just changes the nature of like writing and cool
- 39:56so you can work. And right if I'm not an English
- 39:59like, native speaker, like this algorithms can help me like
- 40:02significant, Writing skills. And now I kind of have a
- 40:05conversation with someone else in another language that I don't
- 40:07have and you can be exposed to because I don't have and that
- 40:09can kick-start something out. So I think it's more interesting
- 40:12to have a like, a very open mind and very kind of like learning
- 40:15mentality of like, okay. How is my job or they think I
- 40:18do? The augmented and how we need to
- 40:20be transform and change. And I think there's like that
- 40:24the hesitation on the, the word conflict component of is very
- 40:26natural like silent film was, like, the thing for 20 years
- 40:31like in the early 1900's, right? And when Audio came to be a
- 40:35feasible technology to have in cinemas.
- 40:38The first reaction from like everyone in the Hollywood
- 40:40industry is like we need to ban audio, right?
- 40:42This is going to destroy the like and this is true like
- 40:45Charles shut not want to be on that side.
- 40:48Are we still recording? Charles Chaplin was like binding
- 40:54audio like I will show you. I the adage is like the
- 40:56association for like film and audio in like the musicians from
- 41:00like Hollywood are bonding audio from movies, right?
- 41:03Because Who's gonna pay for like the orchestra's that are in the
- 41:05theaters, right? And of course, like something
- 41:09happening for the, for the group.
- 41:11A couple things you're saying that stick out to me, are one
- 41:15there's potential here that the proliferation of these tools
- 41:18that let's just focus on writing, like, help you write
- 41:21could actually raise the bar. For what type of writing were
- 41:24willing to an interested in consuming, because if everybody
- 41:27can, let's say in a year quote, right?
- 41:30Like to be T3, the bar is going to be Much higher for your
- 41:34average, I think, article, or newsletter, or whatever because
- 41:37you're going to have to be bringing something really
- 41:39special to the table in order for people to spend their time,
- 41:41reading it. Yeah, 100% maybe that's just me
- 41:43trying to be I just want you to know I've sort of been in
- 41:48Silicon Valley long enough and watch the government be slow
- 41:51that. I'm I'm like whatever is
- 41:52possible will happen and sort of nothing will put the cat back in
- 41:56the bag. So worried about it too much is
- 41:59you know, it's good to think through these things, but but
- 42:01yeah, the technology is going to progress.
- 42:03Yes. But I do think it's an open
- 42:05question. You know what this does to
- 42:08literacy and like all of a sudden people are just like
- 42:10Outsourcing you know their professional writing.
- 42:14You know, most writing is bad, most radius formulaic.
- 42:17Yeah. But people learning to think is
- 42:20important, I mean, I feel like there's been a lot of chatter
- 42:23about the sort of, I don't know, High School, cheating problem.
- 42:26I literally see it on teacher subreddits, you know, worrying
- 42:29about the cheating problem. I don't know.
- 42:32Are you worried about Out it, like, one solution.
- 42:34I was saying is I assume the AI systems will just get better and
- 42:38better over time. So to become easier and easier
- 42:40to identify old cheaters. So, it seems like a big risk to
- 42:45cheat when we're going to be able to figure it out, cheaters
- 42:47more. I don't know.
- 42:48Do you worry about that? I mean, I don't worry about it
- 42:52too much. I think like it might help us
- 42:54like, really understand like from perhaps from are more first
- 42:57principle like thinking perspective, like while they go
- 43:00love, like schools are like, it's not to memorize like words
- 43:02like you need to Have critical thinking and like I'll learn
- 43:05you, I'll teach you how to think and like do the world and
- 43:08process it and come up with ideas, right?
- 43:10And that's what we'll go look. Like, going to school to like
- 43:13anything. Right?
- 43:13If you want to learn something, it's not about memorizing stuff.
- 43:16It's not about managing understanding how to use a
- 43:19complex process and our workflow or whatever it is about.
- 43:22Like look at the wall in a critical way having like a way
- 43:26of understanding things, right? And I think if like if I learn
- 43:29how to code with something like a language model that can Teach
- 43:33me constantly that can fit me, that give me examples and, and
- 43:36they can do a good software idea or I can execute something by
- 43:40doing that. That's fantastic.
- 43:41Right. Did I have to go through like a
- 43:43formal training of reading this like books?
- 43:46And this traditional process? Maybe it didn't, then the didn't
- 43:49matter, right? I think like that might happen
- 43:51and of course, it's normal. There's like push back at the
- 43:54beginning, right? Because it's just it's
- 43:55different. Right?
- 43:56It's changed. And we're humiliating researcher
- 43:58on dates. Yeah, we're very reluctant to
- 44:01change. Like we're very reluctant to
- 44:03change. And then the thing is like we
- 44:04get used to change a lot. I was actually trying one of the
- 44:06self-driving cars UNICEF a couple of weeks ago and the
- 44:10first like ten seconds is just very scary.
- 44:13You're like I want to get out just like yeah yeah just like
- 44:15where's the driver? Like and then after a minute
- 44:18you're like okay I'll just take my phone and just relax, right?
- 44:20It's right for a mini relate, very reluctant to change and
- 44:23then you've tried it and then you kind of like a Sumo K, their
- 44:26works. And in like doesn't like crash.
- 44:28I'm not dying. It can move on.
- 44:29And like, it works really well and the human psychological
- 44:32Journey on I'm embracing new technologies, a little sad.
- 44:35Right? When something is exciting, were
- 44:37very fearful about it and then like you're saying, we just sort
- 44:40of embrace it. And then now, you know, when I'm
- 44:4310 minutes late, because of the subway, I'm like Furious, even
- 44:46though, you know, prior world's, we're getting across New York
- 44:49City, could have taken me like days or, right?
- 44:52And required a horse, right. Right there weren't two Bridges
- 44:55to do it all, you know, like, so it's sad how much we take
- 44:59existing technology for granted and then spend all our time.
- 45:03Session over, what's right over the horizon.
- 45:05But right that's that's Humanity.
- 45:07I agree. Your optimism for this.
- 45:09Technology is truly contagious because I really am.
- 45:11I am thinking about it in less scary - cynical ways.
- 45:15I'm thinking about it in more positive ways.
- 45:17As what could it do to like you, use the word augment or Eric.
- 45:20Maybe you did. Like, how can we look at these
- 45:22as tools that will help augment creativity and not replace it?
- 45:26Yeah, in terms of particular companies, obviously you're
- 45:29excited about your own but what are The project sort of using
- 45:35generative AI that you're most excited about right now or what
- 45:38do you think people should be watching for?
- 45:41Did a lot of things I've seen? There are some I think there's
- 45:44some movie particular that I think for me, exemplifies, a lot
- 45:47of other will continue to Syria over the next couple of years.
- 45:49I don't know, you guys to watch it.
- 45:51Everything everywhere all at once I love love.
- 45:55Yeah. Okay.
- 45:55They're all kind of kind of I know.
- 46:02So there. Beautiful movie, right?
- 46:04He's having watching the you're listening to these specific.
- 46:05Whoa, watch it. It's like phenomenal movie,
- 46:07right? And I think are quicker.
- 46:08I seriously got the million guys.
- 46:09I don't know if you guys remember is it has a lot of
- 46:11visual effects, right? It's very intense usually thing.
- 46:13That's right. The most interesting fact for
- 46:16me, when I learned is and it was just like a Spanish about, it
- 46:19was like the majority of those effects were made by five
- 46:23people. Five people.
- 46:25Right. Just five people, the didn't.
- 46:28That's you think that didn't have previous experience
- 46:30building like feature film complexities and Future?
- 46:33Film effects like that were able to put together that kind of
- 46:37level of like quality and like they're like very talented
- 46:39people for sure, right. But it just fight people that
- 46:42used to take hundreds, if not thousands of like wounds and
- 46:46when I did, I was like wow you just like this is just insane
- 46:49who are these people these are like super humans, right?
- 46:51What's going on? I searched for those five folks,
- 46:53a search for the like MDB. Provide proof of the movie, got
- 46:57the five, kind of pfx, people behind the movie.
- 47:00And, of course, like wanting to run away and see if they
- 47:03Delegate counts, right? And all of them had accounts,
- 47:06right? As I, that's phenomenal ways to
- 47:09reach out to of all of them is like, yeah, we'll chat and we
- 47:12chatted and they eat. They used to use parts of Runway
- 47:15to edit a movie, right? It was just a small part, but
- 47:18yeah, very validating. Yeah.
- 47:22And and then my essentials like, well, if y wanted to like, I
- 47:26guess what, I had the inside of like, I will search for them in
- 47:28our database is because if you think about creating a movie
- 47:30like that, the complexities of the movie already, That's right.
- 47:34But you to be able to execute that you had to automate it, you
- 47:37have to simplify it in a way, right?
- 47:39And my assumptions like maybe they came across the runway
- 47:41because it basically what we try to offer, right?
- 47:44And so they did and we wrote a case study about it.
- 47:46Now, some of them are already using it for other phones.
- 47:48I'll send you the case study. And I think like, what I'm
- 47:50really excited about is those kind of things right?
- 47:53There are very highly creative teams with a lot of motivation.
- 47:56A lot of, like, great to just understand how to get their
- 47:59ideas out without having to have this budget restrictions or
- 48:04There are no other set of limitations than just execute it
- 48:07and it worked. It worked really well.
- 48:08It's like the blue is going to probably going to win like a
- 48:10dozen of dozens of awards. And so the journey would say,
- 48:13I'm really excited about those teams that are embracing
- 48:16technology like this that are like, experimenting with it,
- 48:19that are using that are trying to push it in ways that I
- 48:21haven't even thought of like, no one else I thought.
- 48:23It's just like, you need to, like, expose it to more people
- 48:26and I'm also excited about if you think about that, same
- 48:29approach taken to product building, right?
- 48:31So I don't have the research in the Space has been mostly led by
- 48:34like domain experts. Right?
- 48:36The researchers were like very deep into the way.
- 48:38It's like phds who are like incentives to like make
- 48:41benchmarks in X Y, or Z metric better.
- 48:44But I think the outcomes and the results of using these models
- 48:47are going to impact way more than the researchers, right?
- 48:50And the insights are going to come from like multidisciplinary
- 48:53teams. You need hackers in the Artis
- 48:54need like more people jump in here.
- 48:56So the companies that are more excited or the products and the
- 48:59products, I'm excited about our products that can blend those
- 49:02things. So you have We go, folks.
- 49:03But I think that you have artists like speaking and seated
- 49:06at the same table and be like, okay, here's what's possible,
- 49:09Right? Would that be useful without the
- 49:11not useful? How I'm going to how can I help
- 49:13you out meant something, right? I think that's the second way
- 49:16that I think I was telling you guys are early on like 20.
- 49:18First wave of a I was pretty like these new wave of ideas
- 49:21have been out. There have been a few
- 49:22historically but 2015 to 2020. One was the first one in 2022
- 49:26onwards. The next one is going to be
- 49:27about figuring out how to collaborate with humans, right?
- 49:32And how to take this algorithm From research, domains and silos
- 49:36into real world examples. But part of what I'm taking you
- 49:38to say is that, you know, there's a lot of emphasis on the
- 49:43generalized intelligence and part of what you're saying is
- 49:46that actually you need people focusing on specific use cases
- 49:51and how they play out or is that the right contrast to be zeroing
- 49:55in on. Yeah, I care more about humans
- 49:59and I think they'll adjust has been the purpose of humans,
- 50:01right? So we're helping you You to
- 50:03something I think when we might lose that like, narrative or the
- 50:06other direction, when you think too much about the analogy for
- 50:09the sake of technology and that's the technology as a way
- 50:11of serving humans in a particular way.
- 50:13Wow, I know a lot of people who could hear that.
- 50:20Not gonna say who this is a safe space.
- 50:24I guess you can do it super safe space.
- 50:28Do you mean like open-air? I mean, they're obviously the
- 50:31elephant in the room Microsoft is reportedly like can invest
- 50:35ten billion dollars. I mean, are they arrival of
- 50:38yours is their success good or bad or neutral for you or like
- 50:41what's your view on open eye and how do they relate to what
- 50:44you're doing? I think opening, ISO research
- 50:48driven, Station right there mostly about like, pushing the
- 50:51limits of like research, right? And they've done an incredible
- 50:54job at that. They've released multiple like
- 50:55breakthroughs over the last 10 years, right?
- 50:58They started much more focus on reinforcement learning and then
- 51:01move because they discover new things and that's just how
- 51:03research works. And I think that there are
- 51:05pretty interesting unlike organization.
- 51:08That's done, really breakthroughs in research, but I
- 51:11think research is not enough for products.
- 51:13If you think about products products about again, helping
- 51:16serve human needs and helping People to achieve or summer
- 51:20problem, specific way and part of it.
- 51:23In this case, it's research because you can build the
- 51:25fundamental Technologies but part of it is also.
- 51:27How do people interact with it, right?
- 51:29And there's a lot more that has to be built around it.
- 51:32I think I don't think I was open as a competitor to the NS.
- 51:34I think they're just research organization that's been do.
- 51:37Use their apis or anything, we don't know.
- 51:40No product. We build our own everybody, not
- 51:42only might at some point. I'm sure, you know, more about
- 51:45this than I do. But like, Jasper, whatever
- 51:46superhot company in the tech space was built.
- 51:49On top of open a.i. and then you know open it I comes out with
- 51:54this chat gbt that ends up competing with if I don't write.
- 51:58Am I crazy? I mean there are lots of
- 51:59questions about like in terms of companies who's actually going
- 52:03to be able to win sort of the war here.
- 52:06I mean our bad is, you need to own your stack, you need to own
- 52:10your technology, right? If you're building something you
- 52:12needed to be able to like understand every piece of it.
- 52:15Right? Because if not, someone else can
- 52:17just take it off your right and four.
- 52:19First kind of like Focus as a company has been just building
- 52:22the full stack component, right? That's very interesting in a
- 52:25way, it's sort of like asking the question is generative AI
- 52:29more comparable to the technology that is the
- 52:32smartphone or is it more similar to an operating system in that?
- 52:37Will it matter? You know what I mean by that?
- 52:39Like, is it, is it the technology that things will be
- 52:41built on top of like, in the case of Runway, or are?
- 52:44We talking, which is more, like, I guess you could consider it
- 52:47like more commodified and less like Like owned or are we
- 52:50talking about it as like oh it's equivalent to like a fully owned
- 52:54and operated operating system or something like that?
- 52:57I think it's pretty big. My hunch is more close to a
- 52:59smartphone to be honest. Like, when you have I mean,
- 53:01Apple has radically made that like the case, right?
- 53:05You've you own your Hardware, you build everything from the
- 53:08harbor itself to the software to the platforms where people build
- 53:12on top, right? And that's a way stronger
- 53:14business case or a business like argument than having like slide.
- 53:19The gracious mostly because again, technology moves really
- 53:21fast, right? And so you can consider anything
- 53:24or you can give anything for granted, if things are moving so
- 53:26fast and they're becoming obsolete so fast, we can let you
- 53:30go or Alexis you have any final thoughts?
- 53:33No, I think I really I really grown on this podcast.
- 53:38I'm glad what do you think? I think cuz I think I started a
- 53:41little more fearful and now I think I'm more open minded.
- 53:44And also I am terrified of the fact that 20 years from now
- 53:48we're going to listen Back to this and I don't want to sound
- 53:51like the person that got interviewed by the internet in
- 53:53the 90s and was like You guys heard of this internet thing.
- 53:55Thanks. I'm really trying not to be.
- 53:58I've been clear in my I think it's huge.
- 54:00I mean I was super skeptical of self-driving cars because they
- 54:03need to be complete. Whereas this side you know, it's
- 54:06about the human interaction. I'm already seen you know like
- 54:10professional type. People generating questions and
- 54:14drafting emails and I feel like even if you're plenty smart to
- 54:18write a good piece of It's just like solving it.
- 54:21And then the point that you've made sort of very clear on this
- 54:23podcast, that nothing is more valuable than what a technology
- 54:27saves human beings time. And so, even if you don't view
- 54:30it as like, okay, it's going to be more creative than us in
- 54:33certain ways. It's like, if it's saving us
- 54:35time to do creative tasks, that's sort of the core value of
- 54:39technology and people will always want to save time.
- 54:42So I'm, I'm super bullish on that.
- 54:44I I still think. I mean we saw like lenses there,
- 54:47you know, there's still a hype cycle going.
- 54:49I-i'm. No, I mean, people get so
- 54:51excited, something it's cool and then it falls off.
- 54:54So it's about sustaining that enthusiasm.
- 54:58Yeah, I agree that it's sustaining.
- 55:00The work is sustaining do to set some underdevelopment more than
- 55:03like specific spikes of hype. And so here for the for the long
- 55:06run to make that happen, I think sometimes get things get better
- 55:08after the hype cycle dies down a little bit actually because
- 55:11there's less like fervor and like hand waving in the field,
- 55:15you know? And the people who are genuinely
- 55:18able to and passionate about Those tools like I think of the
- 55:20Creator. Economy is a great example.
- 55:21Remember the five minutes when like everybody, pretended to
- 55:23care about the creative economy. And now money percent of those
- 55:27people are not focused on that, but the ones that are are the
- 55:30folks that are actually passionate actually competent in
- 55:33that space. And I think that ultimately, the
- 55:34tools is a result will be much better for it and I think
- 55:37something like this and self-driving.
- 55:39Cars are probably also parallel examples.
- 55:42Totally cool. Well thank you both for coming
- 55:45on. Listen to Alexis on
- 55:46non-technical Chris Runway check Get out.
- 55:49Alright, thank you very much for coming.
- 55:51Of course, thank you for inviting me.
- 55:53Thanks, Chris, goodbye. Goodbye.
- 56:06Goodbye. Goodbye.
- 56:08Goodbye, goodbye. Goodbye.
- 55:49Get out. Alright, thank you very much for
- 55:50coming. Of course, thank you for
- 55:52inviting me. Thanks, Chris, goodbye.
- 56:06Goodbye. Goodbye.
- 56:07Goodbye. Goodbye, goodbye.
- 56:09Goodbye.