How to Unite Reality With Imagination
My episode with Bilawal Sidhu
As a proud owner of the exceedingly rare “two-digit” designation on CompuServe (the internet’s precursor), I’ve always been an early adopter of new tech.
I was, therefore, particularly excited to speak to Bilawal Sidhu, a one-man corporation whose prolific output outpaces many creators put together.
Since he was 11, Bilawal has been passionate about using cutting-edge tech to create videos that unite reality with his imagination.
He made his bones as a product manager at Google, but after his short-form videos generated millions of views on social media, he decided to go all-in on his creative projects, including this video showing ‘shadow aliens’ invading Miami Mall, which racked up 11M views in 24 hours (!) on TikTok.
Bilawal is also an adept rune-reader in the tech industry — his Creative Digest newsletter and YouTube channel offer insights and analysis on tech and market developments, while his TEDAI podcast broke the recent Helen Toner x OpenAI story.
If you’re a creator curious about the opportunities presented by cutting-edge tech, you’re gonna love this episode…
Here’s the full transcript. As always, if you like what you hear/read, please leave a comment or drop us a review on your provider of choice.
Jim O’Shaughnessy:
Well, hello, everyone. It's Jim O'Shaughnessy with another Infinite Loops. Today my guest is Bilawal Sidhu who just honestly freaks me out. You are a one man corporation as I said to you before we started creating. Your CV is amazing. You are a creator, an engineer. You worked at Google and then you decided, "You know what, I'm just going to... I'm going to become an entrepreneur and I'm going to start playing with AR, VR, AI, and just make kickass videos." We chatted about the one that I saw and if I'm seeing it, man, that is a viral video because I don't have to TikTok. And what's even better is I had a bunch of friends send that to me.
Bilawal Sidhu:
That's amazing.
Jim O’Shaughnessy:
The video that I'm speaking about was the Aliens at the Miami Mall. And again, I immediately... All my friends sent it to me and I was like, "This is so cool." And it immediately made me think of Orson Welles's War of the Worlds. You familiar with it?
Bilawal Sidhu:
Oh, absolutely. I mean, it's funny because War of the Worlds had a similar effect in the sense that back in the radio days when it would air on radio airwaves, a lot of people assumed it was real and it caused a bit of a mass panic in society.
Jim O’Shaughnessy:
And so I guess that's kind of my first question because as I was going through your stuff and listening to several of your podcasts and watching your work, are you worried about this? Are you worried about people...? Like normie. So I've been interested in AI for a long, long time and love it. And yet some of my friends who don't know too much about and fellow olds... By the way, I'm 64. And they're like, "This is scary." Do you worry? I'm a big fan of a book called The Fifth Science, and in it, he's got this great quote about narrative collapse. And basically what he says is that when a society gets hyper-connected and when the tech gets so good that it's very difficult to distinguish a deep fake from a real video that all truths sort of collapse into a sameness. What do you think?
Bilawal Sidhu:
I think I am worried about it, but I'm also excited and optimistic about all of this stuff because the fundamental technologies that we're talking about to create and then distribute this content, I think is dual like use most other technology. And it is interesting, even if we leave deep fakes and generative AI and sort of the ease of creation that's obviously dropping over time, right? Visual effects... Being able to create convincing deep fakes or content that's indistinguishable from reality has historically been the domain of the visual effects houses and eventually the individual effects artists still required a lot of effort. And now what even took me... Otherwise, would've taken me like a week or so to make, in the case of this Miami Mall video, I was able to crank out in a couple hours on the weekend. And so obviously that creates... On the production side, things are getting easier and easier.
Bilawal Sidhu:
But on the distribution side, this trend has been playing out. The phones in our pocket, the screens that we look at all day are already mediating our perception of reality. They're the windows through which we see the world and there is... Whether it's shallow fakes and misinformation on Facebook and WhatsApp groups, and now this new generation of deep fakes, I think that capability has already been there. But I think at the same time, people are getting more and more aware and sophisticated and sort of cultivating a discerning intellect to be able to tell or at least stop and assume like, "Nah, this looks fake." Are we going to reach a point where you really can't tell what's real or fake and truth loses all meaning? I don't know. I don't think so, but that's certainly something I do worry about.
Jim O’Shaughnessy:
So I was chairman of Stability AI for a couple of years and watching it in real time just astounded me. Literally day-to-day changes that were exponentially better. And one of the things that we batted around was, "Maybe we should try to do a prize for a deep fake detection software." And if we got close enough, the idea was we would just release it for free and we'd give it to all the news organizations, we'd put it on GitHub or anywhere anyone could download it. But then it becomes kind of like that arms race. You're young, so you probably don't remember when cops used radar on your car.
Bilawal Sidhu:
Totally, yeah, yeah.
Jim O’Shaughnessy:
Okay. So what happened was... It's apocryphal, but it is a great story because there was a little conspiracy... Nothing like today. Back then, there were very few conspiracies, but there was a little conspiracy that it was the same company selling the radar to the cops and the fuzzbusters to the people and they kept just updating it. And so we'd see something similar. So I guess with these emerging technologies, it's kind of like fire, right? Fire created our prefrontal cortex. When we started cooking food, that's when we were like, "Aha, wait, I'm conscious. I can understand the past and future and I can make plans." And we didn't ban fire even though arguably it's one of the most dangerous technologies, if we're talking about technology, in the world. Instead of banning it, we created fire departments and fire alarms and fire drills. Do you see that happening with all of the technologies that are emerging today?
Bilawal Sidhu:
I think that's the direction it has to go. You're totally right that it's turned into a game of whack-a-mole, right? You create a new discriminator that can tell what's generated, what's not, and then if you open source it, bad actors can use that discriminator to train a new generator to bypass it and on and on that goes. So to me it feels like the analogy that comes to mind is sort of like... And this is perhaps funny to talk about given that we just had this global Microsoft outage over the last couple days. It's the new version of sort of penetration testing and vulnerability management. We're clearly using software that isn't bulletproof. Best efforts are made to sort of make it rock solid, but there's an expectation that people are going to find a way around that and we're going to have to keep up and it's going to be this ongoing arms race or a game of whack-a-mole.
Bilawal Sidhu:
And so I think we're seeing that play out where already in common... The telltale sign du jour for generative AI imagery has historically been, "Oh my God, it has six or seven fingers." It doesn't get the appendages perfectly right or whatever. Or if you zoom in and look at some of the high-frequency detail, you can kind of tell... And that's already outdated. You can generate immaculate still images that, I think, shatter the visual Turing test completely. And I think we're now on the cusp of video sort of hitting that standpoint. And I think going back to good old-fashioned ways of, "Are there different sources that have the same video clip? Do we have multiple observations of this thing to at least rule out the possibility or reduce the probability that this isn't completely fake?" And I think people just employing tactics like that, we're seeing that happen honestly in the light of the presidential assassination attempt.
Bilawal Sidhu:
So much content circulated, some of which was weird, shallow fakes that weren't exactly true. And you saw the community coming up with enough evidence to rule out certain things or discredit certain sources of evidence. And I think that'll keep playing out. And so between efforts that Adobe and companies like it are taking with the C2PA sort of this content authenticity initiative of embedding the right sort of metadata, that's one layer. But on the other hand, I think just public awareness. I like to say if you can't solve this at the point of generation, you have to solve this at the point of distribution. And I think the large tech companies that own the distribution platforms right now have all the incentives to solve it. Yet I will say I see this happening, but it's slower than we expect because again, shallow fakes are still predominantly what proliferate on these platforms. And those are sensibly easier to find because you can tell, "Oh, this exact photo of this politician was actually sourced from this other time and place," because you've got some way to do a reverse image search because this image was repurposed.
Bilawal Sidhu:
That's obviously harder to do with generative AI, but I think the right incentives are there for these platforms to at least not let content that is reaching hundreds of millions of users while being blatantly fake. I'm sure that'll happen, but I assume that game of whack-a-mole will only get better.
Jim O’Shaughnessy:
Yeah. I remember way back when when they faked a letter about George W. Bush and the reason it was discovered to be a fake was the community. Literally in a day, people were posting on the various platforms of that era, "Hey, wait a minute. This letter uses a font that wasn't invented at the date of the letter." And it's like one of the reasons why I'm such a proponent of open source as opposed to closed because the world's creatives and the world's people like you... I would certainly love if something showed up and it looked very, very real. One of the first places I'd check in now, it would be on your channels to see what you had to say about it. So I think that that's right. I think that this is an emergent phenomenon and that we are really undergoing something I'm incredibly excited by. Let's get that out of the way.
Jim O’Shaughnessy:
I absolutely think that this technology is probably the most powerful technology that's ever been invented during my lifetime and its uses also are pretty extraordinary, which we'll get into in a minute, but I don't want to bury the lede here too much. You single-handedly got 360 million views for your art. And I'm calling it art because that's what I really look at it as. You're very creative, it's one of your goals. You're obsessed with uniting reality with imagination, one of my obsessions as well.
Bilawal Sidhu:
Cool.
Jim O’Shaughnessy:
Tell everyone listening or watching, "Okay, how do I do that too?"
Bilawal Sidhu:
Honestly, it's easier than ever before. I like to joke that I got into CG and VFX at an early age at 11. I was making these dorky flash animations back then, and I happen to see... I'd just gotten access to the internet two years before that and I was able to access all these websites that showed me 3D Studio Max. And there was this amazing TV show called Mega Movie Magic that was showing the interface for... This is back... In the pre Maya days, all the tools that existed back then and some of the earliest alias tools. And I saw, "Holy crap, you can use this computer to create visuals that are indistinguishable from reality that can totally..." You can turn your mind inside out and take real world footage and put something fantastical in it and it feels like it's there.
Bilawal Sidhu:
And back then it was like, "Holy crap, I have to buy these massive 3ds Max Bible books that had CD ROMs at the end of it with sample projects and there was no YouTube and all this other..." There's a lot of hoops to jump through to build even baseline expertise to be like, "Oh, finally I can make that lightsaber battle with my homies in the backyard as the clothes are drying on the clothes line there." Or, "I can have a spaceship and make it look like climbing down in reverse from a slide is actually coming down a spaceship." And now it's like the phone in your pocket is capable of that and then some. It's kind of wild. This is back still right as digital video is popping off. So it's still in the mini DV era. I started off in the high 8 era and then I was still digitizing cassettes with FireWire cables.
Bilawal Sidhu:
And so now you've got a 4K camera in your pocket, sometimes 8K, especially if you have an Android these days, it's got a bunch of compute on it, does computer vision and augmented and virtual reality, which AR to me, augmented reality to me is just the visual effects pipeline in real time on the phone in your pocket. So you've got a bunch of apps in your phone that are capable of doing these things, not to mention a myriad of web apps and desktop applications that you can do cool things. So a couple of applications... If you're coming into this, let's say as a complete beginner and you want to do cool things, there's a few things I would recommend. For your video editor, just go CapCut. Mobile or desktop, up to you. If you feel a way about ByteDance and the CCP or whatever, maybe go get Descript.
Bilawal Sidhu:
That's also a really cool, easy to edit application. And if you really want to get fancy, there's the Adobe tools, DaVinci Resolve, but I'd wait to get there. And then on the visual effects side, I would say two or three applications come to mind. One is an app that I've been loving recently called Skyglass VFX. It's this mobile app. It's a bunch of these ex-Apple cats that have basically made a mobile app where you can put yourself in a virtual environment. So if we break down the whole landscape of visual effects that you can do, to saying like we want to have digital characters in the world. Number two, we want to have digital sets. We want to put real characters in virtual environments. And number three being we want to add all miscellaneous other effects, simulations, explosions, all the cool stuff that you want to think about on top of that.
Bilawal Sidhu:
Let's put that as the third bucket. For putting yourself into a virtual environments, level one is just use CapCut or TikTok and use the green screening filter. And there's some fun filters in there where you can put a 360 image so you can get something from Street View or take a Pano or use a bunch of AI generators online to make a 360 Pano. And suddenly you can kind of orbit around a virtual environment. Skyglass takes that to the next level where it's literally pixel streaming Unreal Engine to your phone. So it's using all the AR data off your phone to track you. It'll segment you out or key out your green screen if you've got one, and it'll put you in this virtual environment. And then finally it'll go a step further to relight you believably so it looks like you fit in that environment.
Bilawal Sidhu:
And at the end of it, you just get the video file that you want or you could get a couple of layers like your foreground, background, and the Alpha channel that you can toss into a CapCut or Descript like tool to add more tweaks on top of that. And that tool is just amazing because taking what Mandalorian and high-end virtual production workflows are doing with Unreal Engine and track cameras and letting you do that with the phone in your pocket. And that to me is absolutely wild. And if I spend a second here just imagining if the 11, 12-year-old me had access to stuff like this, I'd lose my mind. So kids don't know how lucky they are these days or adults that are getting into this stuff.
Jim O’Shaughnessy:
Yeah, I remember one of the first viral videos way before YouTube was the Dancing Baby. You remember that one? Or have you seen it?
Bilawal Sidhu:
I also remember seeing it in Ally McBeal weirdly or something like that. Is that true or am I misremembering that?
Jim O’Shaughnessy:
I am not sure about that, but I remember when I saw it, I just got ridiculously excited because at the time I was running an asset management firm and I sold that. And one of the first verticals that we included in O'Shaughnessy Ventures was Infinite Films. Because of all of these technologies, we think that literally you're going to be able to make really watchable and fun films almost exclusively using AI. Now, I also believe that most of the people who just do that, just have the AI do it, I think most of that's going to suck.
Bilawal Sidhu:
A hundred percent.
Jim O’Shaughnessy:
I think that marrying human creativity with it, the so-called Centaur model, man-human plus machine, that's going to be the big winner here. Your work demonstrates that beautifully. And so what do you think? Are we heading for a time when, for example, you could be just your own production company and making videos?
Bilawal Sidhu:
I am of two minds about this one. I think this could be a fun discussion because I did read about Infinite Media and Infinite Films, and implicit in that is sort of the idea that there is going to be an unlimited amount of content that's created. And so let me paint the part bit that I'm excited about and the bit that I'm apprehensive about. What I'm excited about is a hundred percent already... The Skyglass thing I talked about and layer in something like Wonder Dynamics, what lets you do the inverse of that to swap out real humans with a CG character, right? A controllable CG character. And then you've got a bunch of AI video tools that you can use to do other cool things.
Bilawal Sidhu:
That stuff was things that would've been like James Cameron's wet dream in 2009, and that is ubiquitously available now for a very cheap monthly subscription to anyone with an internet connection to the point where you don't even need a fancy computer. You just use the compute in the cloud. And so I think we are already seeing this trend play out where it was sort of like... Let's say the first wave of democratization was YouTube where a percent of the people... The percent of the monthly active users on YouTube were actually creating content. And they're learning these tools like non-linear editing tools like Adobe Premiere, after effects, learning DSLRs, learning lighting, packaging, all these other skills that you need and making some very cool things. And perhaps Mr. Beast is the highest exemplification of that wave. And then you've got the TikTok-ification that's happening, where it's like, "Oh, you don't need any of that. The phone in your pocket is the movie studio. You can do it all on your phone." And so instead of a percent of your monthly actives, you've got 15, 20, 30, 40% of people are creating content, not to mention the shift that we've seen in distribution. And I do want to talk about this in terms of both creation and distribution because you have to look at those things together, in the sense that you could be a no name accounts, have 1,000 followers on Instagram reels, YouTube shorts, or TikTok, and suddenly gain viral success. And you did it all with your phone in this hokey fashion. And I see some kids these days that are recording things in one app, and then they save it to their camera roll, and they take it to another app.
Bilawal Sidhu:
And I just sound like a cranky person. I'm like, "What about generational loss and the compression?" They just look at me like, "What the frick's wrong with you? Look at how many views my videos are getting. Nobody cares about some compression artifacts, you millennial." And so this democratization from YouTube to TikTok's been happening. But what also happened with TikTok is more people could create because it was easier and the blank canvas problem was solved, where you're not thinking about the 10, 15-minute video to make, there's a theme that people jump on. And you're seeing that play out now with the third wave, which I call this generative AI wave, which is supercharging the YouTube world. It's supercharging the short form video world, and it is certainly supercharging the OTT, the Netflix, the traditional studio and the new streaming studio models too.
Bilawal Sidhu:
And all this to say with each of those jumps, distribution becomes more, and breaking through the noise becomes more and more challenging too because there's just more content being generated. So I definitely think more content will be made that wouldn't have been otherwise, that would have never been green-lit. All sorts of content too, like passive video content, interactive video stuff, like gaming experiences, all of these pseudo-interactive things like rap an NPC in some sort of a narrative experience and turn that into something cool. The spectrum from passive to interactive, a bunch of cool things are going to happen.
Bilawal Sidhu:
But on the other hand, I worry that are we creating an exhausting treadmill of our own design. Because the way generative AI is manifesting into creators' workflows right now is ideation and repurposing of content. I would say at the top of the funnel, what do you end up making, and at the end of your creation funnel, how do you distribute this for max impact everywhere? And so I do worry about this thing where, and I'll talk about it from the indie creator standpoint and then even the OTT YouTube standpoint. On the indie creator standpoint, " Oh, so-and-so is putting out 10 videos a week? Holy crap, I need to do 15. Let me figure out how to use AI to do 15 a week." And then it becomes this weird arms race of trying to put out more and more content.
Bilawal Sidhu:
And we're losing sight of what got us into creation in the first place because one would think would happen is more of a Tim Ferriss four-hour work week type story, where you're like, "Ah, yes, let's automate everything and spend our time doing the really creative tasks." But how it's playing out right now is this more volume game, and that's maybe playing to the distribution algorithms. You want to have more at-bats, so you're putting out more content. You don't know what's going to go viral in this sort of interest-based graph.
Bilawal Sidhu:
On the OTT and YouTube end, there's also interesting things playing out, and a lot of that manifested in the tension with the writers' strike that happened recently, especially around generative AI. But really that started off with the change in business model that's happened there. If you were a writer on Seinfeld back in the day, you're probably getting residuals for months and years to come as your content was syndicated in other geographies and whatnot. And now with OTTs like Netflix and Apple, Apple+ moving towards this cost-plus model, as it's called, where here's a flat fee, we're going to pay you a million. You made this. This is our content. Then the platform gets to milk it.
Bilawal Sidhu:
What you're seeing is even a downward pressure on those budgets, where we're on the YouTube side, you've got Mr. Beast increasing his budgets, creating content that's rivaling a Netflix show, especially if you take the Squid Games example, which apparently Netflix had the chance to sponsor and decided to pass on it, and then obviously that thing totally blew up on YouTube and has really high quality visual effects, by the way, like virtual production, unreal engine virtual sets and all this cool stuff. What's Netflix going to do when 25% of people's time on a TV even, which is the recent Nielsen numbers, is spent watching YouTube? And that's all UGC. If Mr. Beast goes out 10 more will jump in playing out that hunger game of more volume.
Bilawal Sidhu:
And so in response, it seems what the OTTs are doing, they're like, "Holy crap, we can't take too many huge swings on crazy tent pole pieces of content that cost 15, 20, $50 million. Let's push that money further and do a bunch of 800 K or $1 million pieces of content, but basically green light a more diverse slate globally so that there's greater chances of some sort of crossover hit happening and hitting a bunch of demographics."
Bilawal Sidhu:
And so I think that pressure, it makes me worried about this race to the bottom that's effectively happening. And so on one hand, to summarize, I'm excited. I think this is a kid in a basement is going to be able to do that stuff that Hollywood directors only dreamt of. On the other hand, I think it's going to turn into this insane, exhausting treadmill, perhaps of our own creation, where we're just trying to crank out such an insane amount of content to break through the noise that I think that's the only outcome I see happening unless, I don't know, distribution is drastically re-imagined in this world of plenty.
Jim O’Shaughnessy:
We have very similar views on that. Part of our thesis is that as this tsunami of AI-generated content comes, the idea that having good taste, that's a very difficult word to define, of course. It's like the Supreme Court justice. When they were trying a case on pornography, he said, "Well, I can't define it, but I know it when I see it."
Bilawal Sidhu:
100%. I'm familiar with the those lines. Perfectly said.
Jim O’Shaughnessy:
So good taste and curation. We think that what's happening now will lead to the emergence of competing platforms that are competing on good taste and good curation, and that those will be distinguishing factors as distribution models. Now, it's not going to replace YouTube, it's not going to replace the streamers, but I definitely think, again, with the advancement of the technology itself, those kinds of platforms will become not trivially easy, but easy if you've got adequate funding.
Jim O’Shaughnessy:
And so we are doing a lot of analysis of the various business models that make the most sense with those as parts of our hypothesis. And I think that in many ways it's where Occam's razoring it. In other words, all of this is going on over here. What's the simplest solution that might emerge from this? And that can take many different paths. It doesn't have to be an actual platform. It could be a channel on YouTube. It could be any of those things. So it could get out into the distribution network effectively, as long as you're willing to understand the impact of it over here or over here, etc.
Jim O’Shaughnessy:
But I definitely share your concern that with the ease with many of these innovations allow, literally the race to the bottom, it's going to be really wild. And I think cultural change is really interesting to me because it's sped up. There's no question about that. All of my life, I was an early adopter. I was on one of the first internet, didn't even call it the internet then, they were all walled gardens, so CompuServe. And it was a real prestige thing. You didn't go by Jim O'Shaughnessy. You went by a number. And the big prestige thing was it had, I can't remember, six digits, a comma, and then if you had a two-digit number, man, you were rad.
Bilawal Sidhu:
The real OG.
Jim O’Shaughnessy:
You were the real OG. And I was so proud of that.
Bilawal Sidhu:
That's awesome.
Jim O’Shaughnessy:
But the fact is that I definitely think we don't live in a static world. We live in a very dynamic world. And when there are opportunities as well as profits available in the sector, capitalism finds a way. All movements start out as passions or whatever. They quickly turn it into businesses and then often descend into being rackets, right?
Bilawal Sidhu:
It's cyclical. It's a cycle, and then it starts all over.
Jim O’Shaughnessy:
It truly is. Let's talk a little bit more about long form. Do you think that there will be some breakthrough film, piece of art, etc. that uses all of these incredible technologies and also is an amazing movie?
Bilawal Sidhu:
100%. I think it's a question of not if but when. And we could get into the technical limitations of the tools to get there, but honestly even before that, if I may just slightly react to the point you brought up previously, if that's all right.
Jim O’Shaughnessy:
Sure.
Bilawal Sidhu:
It's interesting what you're saying about it's not clear on the distribution side, the layer of abstraction that'll make most sense to make sense of distill down, find the signal in an ocean of noise. Is it going to be islands of craters that follow the channel? Like the equivalent of going to Marques Brownlee, that's my favorite tech consumer tech person. And so if Marques has a take on, I'll go watch that. And so that becomes the filter through which I make sense of consumer tech noise or new platforms. And I keep wondering, and this relates to your video question, which is in the future, how is user behavior and preferences going to shift? And what is going to be the breakup of people putting a huge premium on fully organic human-generated content? Is it going to be the split between buying organic and whole foods and it's like a small percentage of the population pays that premium to have something that's total farm to table, super organic, authentic, artisanally crafted, etc? And then on the other hand, most people are totally cool with Costco Target, Target and Mountain Dew straight to the vein like generated content. And I keep wondering, I keep asking myself what that split is going to be, and I worry the split's actually going to be very much in the favor of this fast food content, if you will, because that's what short form is and it works really well. And so in that context, I think we're not that far away, and we are seeing examples of short form content that captures people's imagination, that has a decent story to it. Maybe the highest manifestation of it being montage sci-fi trailers has become the thing du jour lately. And again, this is cyclical. It solves the blank canvas problem for the next set of creators that'll make a bunch of it. Then everyone's like, if I see one more Wes Anderson theme blah blah blah thing, people get over it quickly.
Bilawal Sidhu:
And so long form I've seen recently there was this Terminator 2 remake, I don't know if you saw this, that was made with a bunch of generative AI tools. It's a team of five or six artists that got together just using all of the commercial and open source tools. And it was really bad, but it was funny. It had moments in it, but it was audacious enough to try and do the full thing. I think the moment these tools allow people to do consistent ... To take a snapshot of where we're at right now, there's impressive consistency within a generation, whether that's 60 seconds, a minute, two minutes, but that cross- generation consistency of spaces, places, things like consistent entities, if you will, isn't there. I haven't seen anyone crack it. There are brute force ways of doing it. But that needs to be solved.
Bilawal Sidhu:
And then the iterative control thing needs to be solved too, which is I love everything about this generation except this one thing. Can I change this one thing? Somewhat solved in image. It's trivial to do with a static image, with a static 2D set of pixel. The moment you got the temporal aspect, the video more complicated. And so all this to say, I think even before these technical needs for fitting into the traditional creative process where I have an idea and I would like to trans this idea into something concrete, before that's solved, I think we will see somebody who's very good at co-creating with these models, as I like to call it, which is sort of like, I'm not trying to translate my script into a movie. I have a loose idea for what I want to do, but I'm open to happy accidents.
Bilawal Sidhu:
I'm open to discovering new possibilities and in real time updating your model for what that story ought to be based on where you end and trolloping and strolling through latent space. And I think from that, you'll see something creative that captures people's imagination, that will be narrative, that won't be an exact waterfall creation, if you will, but will be more of this weird agile co-creation thing. And so it's sad that the people that have most of that muscle of telling good stories are at least publicly averse to talking about it. So you're seeing most of those folks use it earlier in their production life cycle to do mood boards and explore different concepts before they write the screenplay manually or whatever. But I think as that stigma reduces, we will see some really cool things happening.
Jim O’Shaughnessy:
If you study new technologies from the past and you look at how they were initially adopted, it's really freaky. Because for example, when they introduced movie cameras, people were like, "Well, I guess we should film plays." That's what they did. They took the camera and filmed a play because that was their context. That's what they ... Let's do that.
Bilawal Sidhu:
The closest proxy.
Jim O’Shaughnessy:
Exactly. That's a great way of putting it. The closest proxy that they're familiar with. And you also saw exactly the same reticence and pushback from the earlier users of that. For example, when records became popular, symphony orchestras took out full page ads in big newspapers saying, "This is an abomination. This is not the way you listen to music. You've got to see music live." And so you saw that with radio. That played on the fear of people like, "Radio? There are waves." And if you go back into the way, way back machine, you'll see there were a lot of newspaper stories showing dead birds around radio towers. And so the thing that you can always, in my opinion, rely on is human OS, human operating system, doesn't change very often. And so if you know that and you study it and you study how new technologies get adopted, that's always part of it, right?
Bilawal Sidhu:
Totally.
Jim O’Shaughnessy:
I don't know whether you saw the Larry David ad in the Super Bowl where they're bringing him all these different things. We call it the wheel, and then Larry David got that face. It's like, "That'll never work." But there is a pattern, if you will, to the way human beings respond. And I suspect that one of the things we're trying to do with Infinite Films and Infinite Media is make it creative-friendly. And it's so interesting to watch people. I have a lot of creative friends, and many of them were on the phone to me, "This AI stuff is horrible. This is awful. It's just going to ruin everything>" And rather than argue with them, I would say, "You know what? How about ..." We had a graphic artist and we have a company called Wand, and we also are venture investors. And I'm like, "How'd you like to just iterate on your own work?" And they're like, "Oh, I can do that?" And I'm like, "Yeah, you can. Let me show you this thing." And then just let them do it. Don't lecture them. Don't tell. Show, right?
Bilawal Sidhu:
100%.
Jim O’Shaughnessy:
That's a good part of good storytelling as well. And I got secretly really excited when my wife, the photographer who I told you about earlier, as you might guess, is not terribly pro-AI, despite what her husband does. She finally came in and goes, "Is there a tool that I could just put all of my own work on and see how to ...?" And I'm like, "Yes, of course there is. I've had it waiting for you right here all the time." So I definitely think that's going to be part of this process as it evolves. And cultural evolution is a real thing. It takes a while. It's like the adoption curve of any kind of tech.
Bilawal Sidhu:
100%.
Jim O’Shaughnessy:
And you're obviously way over here on the early side.
Bilawal Sidhu:
100%.
Jim O’Shaughnessy:
But I do think that ultimately there will be some piece of art, some movie, some story, some series that people are just like, "Holy shit, this is really good." It is my thesis that will be the tech being used by a really cool, creative human being or a group of human beings. And we also have a publishing division. I think when you think about TV shows, they all have writer rooms. For the most part, they don't have a single writer. The West Wing, I guess was best when Aaron Sorkin himself was the main writer for it. But Seinfeld, I love Seinfeld and it had a big writers room, and that's why I also embrace the term you used a moment ago, co-creation. I'm a big believer in co-creation. I think that if Larry David wasn't in that room, Seinfeld would not have been the huge series that it was.
Jim O’Shaughnessy:
I love Jerry. You watch him do his stand-up. He's really, really funny. But that show had its bite, and you'd be hard-pressed to find anything that wasn't addressed first either by the Simpsons or by Jerry Seinfeld. You got to put yourself back in the '90s where they were doing really funny things with what at the time were taboo subjects. And a classic Seinfeld episode is where they have the anti-masturbation bet with the group. And the funny GIF that people always use is Kramer coming in and slapping his money on the counter and saying, "I'm out," at day one.
Bilawal Sidhu:
That's amazing. I need to go down the Seinfeld rabbit hole. A lot of friends have recommended it. I've never gone down it.
Jim O’Shaughnessy:
Oh, I think you'd love it. I think you'd love it. Because the thing that I find it cool retrospectively is, man, they were taking on everything, everything that was at that point in time considered rather taboo.
Bilawal Sidhu:
Breaking new ground.
Jim O’Shaughnessy:
Breaking new ground. And I think that we're going to see the same thing happening here, just a lot faster.
Bilawal Sidhu:
I agree. I think it's definitely the federation of creatives wielding these tools are going to do something amazing. I think maybe the trendy way or pithy way of putting it is everyone's waiting for that Toy Story moment, what Toy Story was for classical computer graphics and 3D animation, where all the Disney cel-shaded traditional animators were also like, "Holy crap, at least let's start doing some hybrid stuff. Let's do the background sets and 3D, and then the foreground can be done the old-fashioned way and then full CG." And it is totally an adoption curve.
Bilawal Sidhu:
It's funny how we always have that initial aversion to it. Maybe we should, after this, talk because I'm curious to get your take on the whole ethics of training data quandary, which seems to be a huge concern. The other one is the seating control concern, which I think is a nonissue because you've got researchers making breakthroughs, and then you've got creation tool builders building creation tools, and both are approaching each other to create the right user experience to empower the creative rather than making them feel like they're being replaced. And I think we've already seen huge strides towards that. And I love the whole control net wave for that, by the way because it was way richer than just text. And immediately a classical artist could see, whether it's their outline scans or their 3D models with a death map, all that stuff could be used to tame the chaos of AI image generation and guide the generation there.
Bilawal Sidhu:
But gosh, these shifts are so funny. I remember when mini DV was a thing, and then digital video became a thing. And I remember George Lucas filmed one of the Star Wars on digital, and everyone was just up in arms, and CG was used in the 2000s rendition. Everyone was like, "What the hell? You're going away from practical. What are you talking about?" And it's funny, that narrative, weirdly, I don't know if you've double-clicked on this, seems to live on in a sense, where I made this video about Top Gun where there's 2,400 visual effects shots in Top Gun. It's nominated for an Academy Award, yet the advertising all, it's like, "Ah, nah. It was just all practical. It was just like in jets."
Bilawal Sidhu:
It's like, no, no, no. Actually every shot has three CG jets, and there's one real one. And by the way, that 3D jet was immaculately 3D scanned, and there's all these other invisible VFX where everything but the actor is replaced. And so I'm like, there's this romanticism for doing things the old-fashioned way that seems to persist. And you, in a way, see that playing out with what I would say is the most vocal community against AI image generation, which is people who are doing digital art essentially with a Wacom tablet in Photoshop or an Apple tablet and Procreate. And it's interesting to see that happen. It happens in more innocuous ways too. I love playing the guitar and everyone's like, "Oh, a tube amplifier. You'll never get there with solid state. You need to have a Marshall or a Mesa Boogie tube amplifier to really get that down." I don't know. Digital modeling is pretty close, so I'm going to go direct into a computer and not have a room and a cabinet full of amplifiers and piss off my neighbors to record some music.
Jim O’Shaughnessy:
I think that, again, that's just another part of human OS. I think that the governing body men in our operating system, if you just look evolutionarily, is fear, especially fear of novel things. And I always joke that if we found ourselves, and maybe we will when we're in VR, maybe you and I will be able to stroll around the savannas of Africa-
Bilawal Sidhu:
That'd be sick.
Jim O’Shaughnessy:
... back when we humans were just emerging. But the thing is, you and me and everyone, or for the most part everyone, almost everyone, never be all or nothing. Again, the idea of deterministic thinking is, I think, going to try to keep that flag hoisted, but it's going away in terms of probabilistic thinking, which is a better way of doing it. But if we saw a bush shaking, we are the descendants of the humanoids who ran away.
Bilawal Sidhu:
Totally.
Jim O’Shaughnessy:
And so that is deep in our genetic structure. And then you see it playing out with all of these new technologies just like it played out in the past. Ooh, what's that? I'm afraid. Douglas Adams, the guy who wrote his Hitchhiker's Guide to the Galaxy, I'm a huge fan of his. And I think that if you think about him as a philosopher and not a sci-fi writer, it makes a lot more sense. He had this great quote, which is, "Any tech that you came upon when you were a kid is a natural order of things, and it's the way things should be." And he goes, "Then any tech that emerges past the age of" whatever, 40, I can't remember what he said, but "is the devil's workshop."
Bilawal Sidhu:
It's an abomination. Yeah, totally.
Jim O’Shaughnessy:
But I think that ultimately one of our primary concerns at OSV is we are here for creatives. And what we are trying to do is in no way trying to replace them, trying to make them work for less or any of those things. It is to find really, really creative people like you and work together, the whole idea of co-creation. And I think that for artists, it can be hard because again, it's something that's very, very consistent.
Jim O’Shaughnessy:
When the Impressionists burst onto the scene, that name was given to them as an insult by the French Academy. They literally would not let them display their work in the salons of the era. And if you go back and you look at what was the dominant art at the time that the Impressionists were beginning to be seen, it was done. It had become so stultified and so just bad. And here come the Impressionists and everything that involves we humans, there was a hierarchy, and the people who ran the French salons at the art academy were at the top of it, and they were threatened when they saw this. So literally they banned them from exhibiting in the fashionable salons. They named them, "This work, this is not real art. It looks like a child finger painting with just impressions." And so what did they do? They did what creators are doing today. They found different distribution channels. They found different audiences. And the rest is history. So I definitely think that we're going to see something similar play out with all of the work of today.
Jim O’Shaughnessy:
And I don't know that there's a formula. I don't think there's an algorithm that you can use to say, "This one's going to really knock everyone's socks off." We're looking, by the way, I would be lying if I told you that we weren't. But you know what an urpattern is? So an urpattern, it's really cool. Howard Bloom is a genius in my opinion, he's written many books that I have devoured, and one of them posits an urpattern, which it's like a primitive in AI models, it is a pattern from which that is very simple, and yet, like a Penrose tile, from this very simple pattern, very complex and very interesting...
Bilawal Sidhu:
Complexity emerges.
Jim O’Shaughnessy:
Complexity emerges, right.
Jim O’Shaughnessy:
And so, one of the things we're doing is hunting for urpatterns. Because in literature, the hero's journey is a well-known urpattern. And it's hard to deny that when you look at the most popular stories, not just movies, but books, and books from 300 years ago, the ones that really clicked with humans were the ones that, for the most part, follow the hero's journey.
Bilawal Sidhu:
The Hero with a Thousand Faces It's the same story over and over again since time immemorial. There's something interesting about that framework, the fact that it's become this conduit to, I don't know, encode generational knowledge or just stories that are important to the human experience.
Jim O’Shaughnessy:
And I think that's exactly what it is. I think that that is the pattern is do these stories speak to many humans in a way that it really resonates with them? Now, I don't think that that's going to preclude... We've gone from a normal distribution pattern, a typical Bayesian distribution pattern, where everything was aimed at that 68.5% in the middle. And we've gone from that into a world that has a chaotic, normal distribution pattern popularized by Mandelbrot, with a very peaky middle and very long tails. And so I think that there's gold in them there tails.
Jim O’Shaughnessy:
You'll be able to have a career that is incredibly renumerative, that you love, that you're obsessed with, in areas that under the old distribution and the old aiming at that middle...
Bilawal Sidhu:
Niche.
Jim O’Shaughnessy:
... Wasn't possible.
Jim O’Shaughnessy:
Exactly. And so it doesn't mean power laws go away. They won't, but there'll just be many different, smaller hills. And so I think that there is room for both, and that's one of the things that gets me really excited about all of the various tools, and creatives using those tools, they can conquer niches that literally were unconquerable because they wouldn't pay off. Now, there's so many different ways that people could succeed creatively, and we think, "Why isn't there room for both of these?"
Jim O’Shaughnessy:
One of the things that fascinates me about you is, what's your process for spotting a trend that attracts you, like, "ooh, that's very cool and then moving from that to one of your finished products?
Bilawal Sidhu:
Interesting. Deciding to make something?
Bilawal Sidhu:
I could use the Miami one as a great example.
Jim O’Shaughnessy:
Please.
Bilawal Sidhu:
It's interesting. So I would say the theme for my content, is it real? I want to leave people asking that question of is it real or is it fake? My goal is to blend reality and imagination. And I noticed when things are obviously saturating social media... So a lot of people are talking about a thing. I remember making a previous rendition of this that helped me blow up my TikTok account, back when Raid Area 51 became a meme, I don't know if you remember this.
Jim O’Shaughnessy:
I do.
Bilawal Sidhu:
People were like, "Oh, there's going to be World War Three. And by the way, there's this Facebook group that's going to show up to a military installation, try to Naruto run inside the base."
Bilawal Sidhu:
And so when I see things like that jumping into public conversation, public consciousness, and there's... Especially with this Miami Mall thing, people were circulating the most grainy, insane helicopter footage and trying to 30X to be like, "Look, if you look at those four dots, there's alien over there."
Bilawal Sidhu:
And I'm like, "What if there was actually 10, 12 foot tall aliens? What would that actually look like? Can we visualize that?" With the idea being, obviously it was so over-the-top and absurd that most people that see it are like, “obviously this isn't fake.” But I'll get into this gear where I'm like, "There's attention around this topic. I can make a reasonable facsimile representation of what that perceived narrative might actually manifest as, and then I go and create it."
Bilawal Sidhu:
And honestly, I try to time box these things. This is what I've always done for my Shorts. The reason I love Shorts over making long-form content, personally, is because it's the type of thing where over the course of four, eight hours, maybe two days like Saturday and Sunday, and two four-hour shifts, you can basically go from idea to completion, and still have enough time to explore the creative opportunity space at what you're going to make, especially with these AI tools.
Bilawal Sidhu:
And so now what I've started doing is previously what I used to do is I do this the old-fashioned way that I'd go capture real footage that was intended for it, now I can just do it all in Midjourney, storyboard it out, and throw it into Runway, Pika, or Luma, or whatever the best image to video thing de jour is, and then finally bring it back into After Effects to make something fun with it.
Bilawal Sidhu:
And then the way I add this final layer of believability and anchoring in it is definitely the audio, I think, makes a huge difference for these type of things. And adding that very subtle signature little bit of handheld camera shake, things like almost artificially decimating the video a bit more to make it look more authentic, put that all together with the helicopter, or the sirens and all of that, it feels anchored and embodied in that event that happened. And then most people saw it as like, "Oh this is funny. This is if Attack on Titan happened in Miami." And then a bunch of people were like, "Oh my God, I knew it! It's real!" And I'm like, "Oh, that's where I get a little worried."
Jim O’Shaughnessy:
But that's where... You're making precisely the point that I think is important here, and that that's where your creativity kicked in. And you were able to just go, Oh, wouldn't it be cool?" And another one that you did that I loved is where you took Tiffany Gomez, the woman on the plane, and you turned her into an alien. But the payoff of that was, "That motherfucker isn't real!" And then you shift to Scooby Doo in a [inaudible 01:05:50]. I can see very easily why that went viral, because...
Bilawal Sidhu:
It's memes.
Bilawal Sidhu:
And what's also funny about that one is there's this co-creation aspect with the community, too, which is a lot of fun. And I actually picked up the... I won't take credit for the really fun panning over part. I saw that in another video and I was like, "Wow, that's cool." And it lines up perfectly with that shot from the latest Scooby Doo movie where it feels totally embodied in the scene. So I was able to put those things together, which is a testament to how awesome remix culture is, and you see that play out every day on YouTube, TikTok, and certainly Twitter and all these platforms, too.
Jim O’Shaughnessy:
But again, as somebody who's been around for a long time, remixes are really not all that new.
Bilawal Sidhu:
Oh, amen.
Jim O’Shaughnessy:
They have been part of art, part of music, part of film, part of everything for a long, long time. And you mentioned earlier, how do I feel about training the AI models on data? And my view is have you ever been to a museum? What do you see in museums? You see a ton of art students. What are they doing? They're trying to paint that Rembrandt, but in their own way, they're trying to learn. How do they learn? They learn by going to the museum and looking at great artists.
Jim O’Shaughnessy:
And so look, I am very sensitive to the idea that you should not have the ability to just hijack some other creative's work and present it as your own. There should be red zones on, "Okay, this is a remix, this is an homage."
Jim O’Shaughnessy:
But you made a remark about AI, it's trained on us. And so to expect that it's going to be suddenly incredibly ethical, and incredibly fair, and all of those things, I don't know, have you read human history? But also I always try to think, what can I root for? It's really easy rooting against things. It doesn't take much intelligence to be, "I'm against that!" That's bad.
Bilawal Sidhu:
Especially against change, to your earlier point of things that induce uncertainty in people that just has that, we have this visceral reaction to it, to push back against it.
Jim O’Shaughnessy:
And the illusion of control, if you really want to fuck with a human being, take away their illusion of control. And why do people get put into solitary confinement? If you know the literature, removing a human being or another animal for that matter, from their community is really bad. And there's two examples.
Jim O’Shaughnessy:
They did a study of looking at people who were put in solitary confinement, and many instances it was as little as 18 hours before they began hallucinating.
Bilawal Sidhu:
Oh, my God!
Jim O’Shaughnessy:
And why are all those prophets... Why do they spend all those 40 days in the desert and start seeing burning bushes? I have an idea.
Bilawal Sidhu:
You go deep into meditation in a cave for years, you're going to start having some crazy extra sensory experiences, I'm sure.
Jim O’Shaughnessy:
Exactly. And that doesn't make meditation bad. I think meditation is really, really... I love sensory deprivation tanks.
Jim O’Shaughnessy:
But the point is, people... We were talking a little earlier, deterministic versus probabilistic, things are not, yes, no, things are not zero 100. And so we have this deterministic thought pattern that thinks that we are in control, that does the illusion of control, and it becomes very, very difficult for many people to get out of that box, that thinking box.
Jim O’Shaughnessy:
But back to the person in solitary, guess what? If you take a rat and you put it in a box, and this is a very famous test, one of the drinking things is water the other one is water laced with cocaine. And if you've got a single rat in that box, guess what he does? He drinks the cocaine. He does the cocaine water until he literally dies.
Bilawal Sidhu:
Oh my gosh.
Jim O’Shaughnessy:
I can't recall the name of the researcher. We'll have it in the show notes, and we'll include a link to the study.
Jim O’Shaughnessy:
This guy thought, " Wait a minute." So they tried to prove, "Look at how addicting this cocaine is, and this is why we can't allow it anywhere, anytime." And he thought, "Wait a minute. Maybe they've got this wrong. Maybe it's the isolation that is causing this rat to feel like such despair and such depression that hey, why not do the cocaine?"
Jim O’Shaughnessy:
And so what he did was he created rat parks, and did the exact same thing. One was just water, one was cocaine-laced water. And guess what happened where it was a community of rats, not a single rat?
Bilawal Sidhu:
That's amazing.
Jim O’Shaughnessy:
The use of the cocaine water plunged.
Jim O’Shaughnessy:
So I think that if we're mindful about all of these things that rather than saying, "That's bad, let's ban it. No, you can't do that. Let's make math illegal." Rather than that, let's be cognizant of these types of occurrences. Let's try to figure out things that can ameliorate them and then move forward. It's like AI. I know that you have done some...
Jim O’Shaughnessy:
By the way, one of the secondary effects of getting ready for you is I've got to reserve 10 hours of my time to listen to all of your podcasts and stuff, because you seem to be interested in exactly the same things I'm interested in, like AI companions, AI, mental therapy.
Bilawal Sidhu:
That's where my head was going with that example.
Jim O’Shaughnessy:
I have that, on a use case thing I wrote several years ago, they were on there. And I showed to some people who were not into AI. The visceral reaction against was so overwhelming. I'm like, " Listen, we're caring for my nearly 98 year old mother-in-law, and we're very lucky because we get to have her live with us, give her companionship, all that stuff."
Jim O’Shaughnessy:
But there was a period where I was at a rehab place for elderly people, and I got to tell you, man, I do not want that fate, because they're alone in their room. They're like... I'm really not exaggerating. When I said I was walking to one room to visit somebody, and the cries said, "Will somebody please come help me? Will somebody talk to me?"
Bilawal Sidhu:
Oh, my god. That's heartbreaking.
Jim O’Shaughnessy:
And one of the nurses said to me that the hottest item on that floor was a radio or a TV. They didn't have enough for everyone. And so my thesis is, of course, if you have the ability to take care of an elderly person, then, by all means, do that.
Jim O’Shaughnessy:
But life isn't perfect, and I tell you, if I couldn't be taking care, for example, of my mother-in-law with my wife, I would want her to have companionship chat. How cool would it be if we could animate all of her grandchildren and children and she could have conversations with them? And then I get it, I get the uncanny valley. "But they're not real." Okay, the first versions of these might not be that great, but mental health, I think that a lot of people actually might be better served by AI, large language models, trained...
Bilawal Sidhu:
Infinite patience.
Jim O’Shaughnessy:
Exactly. Infinite patience, no judgment, all of the things that people...
Bilawal Sidhu:
Amazing memory.
Jim O’Shaughnessy:
Amazing, all of those things.
Jim O’Shaughnessy:
And so I look at those things and I'm like, "Where do I sign up, man? This just makes a ton of sense to me." And yet this pushback of, "No. It has to be a human therapist." And for those listening, I'm not arguing against human therapists. I think here, too, this centaur model that's half human, half AI will emerge as revolutionary in mental health discussions and things like that. But I think that this idea that no, it's got to be this or nothing, it's never that. And so am I just completely off base here or what do you think?
Bilawal Sidhu:
I think you're hitting on all the good points. I would say I always view these things as two possible futures, and it's fun to talk about them as dichotomies and polarities, because they frame the contrasting outcomes a bunch. But in reality, it's always a mix of both simultaneously.
Bilawal Sidhu:
People do good stuff happens, bad stuff happens. People are going to, with the NPCs being infused with far more intelligence to the point where they're no longer non-playable, and they're super engaging AI characters that you can talk to, I think there's going to be huge benefits to that. You mentioned certainly if you're an elderly person, and this chatbot has all the context of your life, presumably because you have a huge digital footprint, literally you've been creating your digital twin, and it can keep you anchored and tethered in reality, and engage with you on the topics that you care about, and make you feel heard, listened to, that's amazing.
Bilawal Sidhu:
And on the other hand, there's going to be the whole AI, girlfriend/boyfriend AI relationship craze period. And maybe people are going to use it unhealthy to the point of going back to a Cartman and the Mountain Dew in the Vein analogy of the World of Warcraft Cartman episode where his mom's coming in swapping out the poop tray and giving him the new whatever, hot pockets. And "Here you go. Here's the new Amazon Prime just droned in your new Mountain Dew that you just jack straight in." Both will happen, but that's just how it's a free country, at least in the West. That's how it ought to play out. But it's also binary in other ways people painted.
Bilawal Sidhu:
And you brought up a great example of like, "Oh, well, we need to have real teachers, or we need to have real therapists." And nobody's saying all of these folks need to be replaced. I think people are so quick to fast-forward to the replacement scenario. The same thing with creators to, it's like we're all just replaced by digital twins. Hey, maybe that will happen for some long tail of content that maybe should have been automated anyway. But I think in the scenario of these one-to-one human expert roles that we have, I think you can scale that relationship so much more by having, as a teacher, this came up with in an episode with Sal Khan, you've got your infinite, always-on teaching assistant that's available for the students when you're not there so the students' not bothering you on Sunday, where students that are perhaps reaching a different level of proficiency with the curriculum, and maybe are struggling in other aspects to round that out so everyone completes the curriculum together. And it's the same thing with chatbots. Where, like talking to Alison Darcy, the Woebot CEO, it's interesting. We talked to one of their patients, too, and this kid, Gen Z kid basically is like, "Yo, I can talk to this thing whenever I want. If I'm starting my day..." It's almost like I felt like this kid was using it as positive affirmations during the day and then like a diary whenever stress or anxiety popped up too high. And to be able to do that in the moment you need it the most, I think, is just so powerful. And I think it's a yes and, versus an either or, I think you probably still want to go have those calls in person.
Bilawal Sidhu:
That's where it can get interesting. I was talking to a CEO of one of these coaching platforms who will remain unnamed, which is basically like you're saying that they empowered their coaches with all these perception capabilities, too. So you're pointed at a person, you could look at, what is their facial expression, what is the likelihood that they're feeling excited? And it was like Terminator vision for a coaching call, and it was almost like too much data for them. They were like, "Holy crap, this is weird."
Bilawal Sidhu:
So, I think, to your point, maybe it'll take a younger generation that is far more acclimatized to these type of capabilities to start taking advantage of those signals and even include them in practices. Because as you mentioned earlier, which I think is such a beautiful depiction of what always plays out, people who are at the top of the hierarchy, the experts are always the most averse to adopting this tech. Of course, the Harvard, Stanford MDs that are making a money, a boatload of money off of therapy, let's say in this case, are going to be averse about what some second-rung. "Therapists can now augment themselves with AI and offer this thing? It's going to be competitive against mine?"
Bilawal Sidhu:
But I think we'll see that always play out where the people that are the underdogs or the people that are new to this are far more cool with the tech. And so chatbots all end on this as thing is another, they're been very fascinated by Character.ai, which is obviously created by some of the founders of the Seminole Transformer paper. But you see this, it's really popular amongst Gen Z and younger. And you see the Reddit pages about this stuff, and you see the "AI characters" or NPCs that folks are authoring, and it's amazing. And they're spending millions and millions... There are millions and millions of chats that are created with all of these folks, because it's second nature, it feels like, duh, to them because they're growing up in video games, seeing these NPCs evolve. Whereas for us, it feels like it came out of nowhere and as this abomination, we're way more averse to it as this thing.
Bilawal Sidhu:
Whereas, I don't know, this just feels like Siri on steroids and it's just now Siri has a bunch of different personas, and based on the context, now, our AI assistants can be embodied as a talkable, maybe visually-embodied agent, maybe then full 3D-embodied agent, and this is a more intuitive way of interfacing with them.
Jim O’Shaughnessy:
And one of the things I think I've been lucky in that I've always been incredibly, voraciously curious about, I love new tech, I love all kinds of new tech. Now, does all of it work? No. A lot of it doesn't.
Bilawal Sidhu:
Most of it is...
Jim O’Shaughnessy:
Most of it doesn't.
Jim O’Shaughnessy:
And so I ended up running up quite a bill on, "I have the DVD holder with a hundred, you come and we can watch one of a hundred movies!" And then streaming [inaudible 01:22:18] really bad investment with that piece of hardware.
Jim O’Shaughnessy:
But I definitely think that one of the things you need to do to humanize... It's like, I love techno music, and so I can't use that for most members of my generation, but I was listening to this great song and it's like, "Come with me and you'll never be alone. Etc." And I had this image for how do you humanize... How do you get people to be really excited about digital twins? Again, well, you got to go to their emotions. So what about a two-minute video that has some really evocative music playing at the same time that starts with, and this is harsh, but you got to go for those emotions, that starts with a doctor coming out, looking at a pair of parents with the look on his face of, "I'm so sorry?" And those parents and those parents just breaking down and then showing, creating a digital twin of that child who had that disease, and finding a cure. Two different pictures emerge. And to me, you've got to emphasize, being 100% cognizant of all of the problems that present themselves, because I am not going to try to be panglossian in here and say, "Oh, it's going to solve everything!" No! It's going to actually cause a lot of new problems, but better problems in my opinion. And if you get to choose, do I want to live in a world where my grandchildren's children can stay alive because they had a digital twin and the AI was able to do search a hundred million vectors and find the four that was going to be the right cure for them? Hell yes, I do!
Bilawal Sidhu:
It's also interesting, because in a place where all this data exists. We've got these islands of data that are sharded and sitting around, and we're not really even putting it together in interesting ways. So I am very excited about that.
Bilawal Sidhu:
I am curious what you think about... This is one of the positions I've evolved a little bit on, because previously I was in the, everything's a remix camp pretty much, which you're describing. And I've had some empathy for independent creators in this, which is the following that the playbook that the large tech players are playing out, let's say OpenAI, to pick on them for a second, is like, "You train on everything. You scrape the public web..." And there's of course this funny euphemism of publicly-available data, which is like anything you can access with a browser and internet connection.
Bilawal Sidhu:
And of course it seems like YouTube ain't going to do anything about it because they're like, "Wait, it would be hypocritical of us even though our terms of service let us train." So it's interesting, you got this place where everyone's training on publicly-available data. And so that, is in a weird way, like a distillation of all human knowledge and creativity, documented human knowledge and creativity. And it's like holy crap, in a sense Web 1.0 and Web 2.0 was like when you had the ability to create text and image content, we were creating this digital twin. We were co-creating this digital twin as a hive mind, and now from that digital twin, we're extracting and distilling intelligence, whatever, however you want to frame that.
Bilawal Sidhu:
But I think the problem in all of that is the playbook that the Open AIs of the world are setting is like, well, you go train on everything, you raise a bunch of venture capital, and then you go and retroactively cut deal with the largest IP holders, the people that have legal teams that can come after you. And so, "Let's go get the Atlantic down. Let's go talk to Shutterstock and not get rid of these watermarks in our Dolly data set, or whatever and fix all these things." And the people left holding the bag in that world is like the no name, YouTube, insert art station, whatever, creator.
Bilawal Sidhu:
And that's where I feel like I know from technical standpoint, attribution is very challenging. And how can you really tell provenance? Removing duplicates from Laion-5b itself is a huge technical undertaking, but I can't help but feel some empathy for folks that are like, "That's cool, but all the journos..." Maybe with the exception of New York Times that, let's say, has enough of a war chest to sue OpenAI, everyone else is taking their pennies on the dollar right now to be like, "Oh, shit! Elvis has left the building. Cat's out of the bag. Let's fricking take the money while we can right now as the sort of ad-funded search traffic model is on a downward trajectory. And the people that get screwed into this are the indie creators. What do you have to say about that? I have yet to see a business model where people are like "Well, we just need to create a new marketplace that renumerates creators." And I'm like "And give them what? A fraction of a fraction of a cent." I haven't seen an economic model that sort of makes sense there.
Bilawal Sidhu:
And perhaps it's fitting that even in the video space, which is sort of this higher level of abstraction of creating content, really you can make passive money on YouTube. And maybe to an extent Twitter I'll get 100 bucks every two weeks. I guess that's cool but that's not a living, right? How do you feel about that where sort of the indies are left holding the bag and the people that have ... That are sort of higher up the hierarchy that you described, that are maybe not on the top of it but are closer up, are still getting benefits of this transition that's taking place even though they're criticizing it the most publicly?
Jim O’Shaughnessy:
I have, like you, tremendous empathy with the indie creators. We have not yet found a business model that works perfectly. It does not stop us from looking because we think that it is a major thing that needs to be addressed. If we're picking sides here I'm on the side of that indie creator.
Bilawal Sidhu:
Okay.
Jim O’Shaughnessy:
History in the world is a movie it's not a snapshot, right? Let's not forget that Steve Jobs and Wozniak we're in a fucking garage and look at Apple now. There were others like IBM, DEC. There were huge companies that controlled computing back then, right?
Bilawal Sidhu:
Huge. Totally.
Jim O’Shaughnessy:
And what were they trying to do? They were doing everything they possibly could to make sure that Apple didn't work, didn't win.
Bilawal Sidhu:
Totally.
Jim O’Shaughnessy:
Schumpeter, an economist, called it creative destruction. That's what markets are. Markets are perpetual creative destruction machines.
Bilawal Sidhu:
Totally.
Jim O’Shaughnessy:
People who try to freeze winners and losers in place don't understand the way that markets work.
Bilawal Sidhu:
That's called calcification, right? That's going to shatter.
Jim O’Shaughnessy:
Yes. It's stasis, it's debt. And it's also incredibly intellectually arrogant. We know everything there is to know now and we're freezing it. These guys at the top are going to stay at the top. These guys here, no you can't create, no you can't innovate, right?
Bilawal Sidhu:
Yeah.
Jim O’Shaughnessy:
I always remind people of the idea that the monster companies of today were not monster companies 20 years ago.
Bilawal Sidhu:
Totally, yeah.
Jim O’Shaughnessy:
And the idea that the monster companies of 20 years from now are going to be the same companies of today, I don't think so, I wouldn't bet that way, right?
Bilawal Sidhu:
Yeah, yeah.
Jim O’Shaughnessy:
As long as there is a relatively free market, as long as there are relative rules of law in place ... And will they get murky and will they have to be redone? Yes, they will, right? If you look at laws that were passed in, I don't know ... Let's make it even current, 1950. What percentage of those laws are antiquated, and make no sense, and are actually retardants of change?
Bilawal Sidhu:
Totally.
Jim O’Shaughnessy:
A lot of them, a lot of them. I definitely think that yes, it's a problem but I am optimistic that the markets will figure out a way to make sure that indie creator ultimately is paid in some form. I don't have that model yet but it doesn't mean we're not looking for it. It doesn't mean that there won't be a lot of other people that'll finally figure that model out. I mean, I don't know. I've gotten pitched on Ethereum, and chain of ownership, and provenance, and all that stuff. I'm not smart enough to know what the technical difficulties are there. With as many people thinking about it as possible, it will emerge. And interestingly enough as you ... Not to just pick on OpenAI, but the big companies, they're not trying to be your ally here. They're really trying to keep their advantage as ... By the way, as would I.
Bilawal Sidhu:
The incentives for them are to do that, right? Yeah, of course.
Jim O’Shaughnessy:
Exactly, right?
Bilawal Sidhu:
Yeah.
Jim O’Shaughnessy:
My argument is to say, there are other incentive structures. As long as those incentive structures are not thwarted by political capture, regulatory capture, all of the ... What do you see them all doing right now?
Bilawal Sidhu:
Totally, yeah.
Jim O’Shaughnessy:
They are trying their best to capture the regulators. And again, rather than shake our fingers and say, "That's bad, that's awful" ... It's the inevitable outcome. If you're a trillion-dollar corporation and want to stay a trillion-dollar corporation, of course, you're going to try to capture the regulators.
Bilawal Sidhu:
Mobilize your army of lawyers and go do some lobbying. Of course, yeah.
Jim O’Shaughnessy:
Exactly. Look at Napster, right? People forget Pandora's box, right? They always say, "Oh, we're opening Pandora's box." People who don't know the story of Pandora's box don't know that at the bottom of Pandora's box was what? Hope.
Bilawal Sidhu:
Cool.
Jim O’Shaughnessy:
That was what was at the bottom of Pandora's box. When VHS came out and was really big, the movie industry went batshit crazy trying to suppress it. Where do they make most of their money from? Or not now, but for a long period of time the movie companies were making all of their money from the video rentals and-
Bilawal Sidhu:
Oh, totally, yeah.
Jim O’Shaughnessy:
All of that. They spent armies of lawyers against those companies. They tried to suppress it every way they could. Same with the music industry. They did everything they could to suppress the new technologies.
Bilawal Sidhu:
I mean, recently an example that came up because his mother had shared it is ... And I had this post go insanely viral. I think it's actually now my most liked tweet which is weird because it's a heartfelt piece on Aaron Swartz who was so crazily prosecuted for, I don't know, scraping JSTOR articles which just feels like innocuous today given hey, let's just scrape the public internets and sell it for 20 bucks a month and everything's gravy. It's funny how the Overton window shifts. You're totally right, there has been that aversion.
Bilawal Sidhu:
And it seems like, especially in the case of creation, empowering more people to do something that previously few people could do especially when it comes to creative self-expression, I think on net on balance, that's going to be a disproportionately good thing. There's going to be some of these misinfo, disinfo thing. Everyone loves calling it slop these days. There's just all this AI slop that's coming to the internet. But I think we'll get good at identifying the slop and sort of cutting through the noise, and maybe there will be that cream that rises to the top. To your point, the power laws in these various niches playing out.
Bilawal Sidhu:
I think regardless of the ethical quandaries, and I'm sure there's solutions there ... Jon Barron, one of the NeRF folks at Google was saying, "It'd be cool to have a model that was purely trained on real world imagery, like a generative AI model, just to give the art community a sense of what can these models do if you just image observable things in the real-world." And I think we all might be very surprised with how much capability that already infuses in these models. It's like through the lens of our consciousness and then sort of our artistic endeavor we're sort of filtering reality and remixing it in interesting ways anyway.
Bilawal Sidhu:
I'm excited to see overall where that goes out. But yeah, it's funny to see the incumbents change. It's funny to see the rise of new entrants that are trying to dethrone the incumbents. It's funny to see open source and the folks that are championing it. Obviously, Stability huge in the space there. Meta would be the bastion of open source. I did not have that on my bingo card for last year. And it's been really cool to see that play out as well.
Jim O’Shaughnessy:
Yeah, I agree. I could not agree more wholeheartedly with more is better. Getting these tools in the hands of as many people as we can, I think, is a good thing. No matter how smart they are, no matter how creative they are, you can never make them create a list of things that would never occur to them.
Bilawal Sidhu:
Yeah, totally.
Jim O’Shaughnessy:
Right?
Bilawal Sidhu:
Exactly.
Jim O’Shaughnessy:
What doesn't occur to me might occur to you. That's how we build the human colossus. You can say the hive mind. I prefer the metaphor of the human colossus.
Bilawal Sidhu:
I like that, yeah.
Jim O’Shaughnessy:
And it's emerging already, and then AI is just supercharging it, right? You have to have cognitive diversity to make that actually work. Again, the arrogance of it just offends me. There is no cabal of people who know what's best for everyone, right?
Bilawal Sidhu:
Totally, yeah.
Jim O’Shaughnessy:
And the arrogance of thinking that leads to things like the guy closing the patent office in the late 19th century because everything's been invented, right? Okay, we're done. That actually happened, right?
Bilawal Sidhu:
That's amazing.
Jim O’Shaughnessy:
And David Deutsch is one of my favorite thinkers on this-
Bilawal Sidhu:
Cool.
Jim O’Shaughnessy:
Where he's talking about "Hey, what were people saying about the internet, and nuclear, and quantum physics back in 1900?" They weren't saying anything because no one was thinking about that particular ... Those particular things, right?
Bilawal Sidhu:
Yeah.
Jim O’Shaughnessy:
Again, human OS, right? We fall into this trap of thinking we know everything and we don't we're just scratching the surface. Humanity is really young. The first movie ever taken was a Pope of the Catholic Church, right?
Bilawal Sidhu:
Cool, yeah.
Jim O’Shaughnessy:
And the guy was born in 1810, okay? Okay. To put this into context, my great-grandfather was born in 1835. That's how close-
Bilawal Sidhu:
That's not that long ago.
Jim O’Shaughnessy:
No. We are babies, we are babies. This innovation curve that is going vertical, yeah, will it cause a lot of chaos and a lot of creative destruction? Yes, it will. I suspect that the more people that have these tools available to them, the more diverse the people who come at the world from different points of view. India should have their own version of AI. So should Bangladesh.
Bilawal Sidhu:
Oh, sovereign AI.
Jim O’Shaughnessy:
Yeah, sovereign AI, right? Different things are more important to different cultures. Okay, but what's wrong with that?
Bilawal Sidhu:
At the same time, isn't it becoming a very tempting form of exerting geopolitical control, right? Again, the window through which you see the world, if you can own the window through which you query and generate knowledge, oh my God, that sounds extremely strategically beneficial, right? Maybe an oligopoly of models that control the flow of this information and then a long tail of open source models. How do you see that playing out? Singapore is going to have their own sovereign AI thing and it's going to be just as good as, I don't know ... But maybe Germany decides to change ... Take some forked version of the US one because we're more ... A Venn diagram of overlap. I'm curious, what world do you see there?
Jim O’Shaughnessy:
I have a long history of being wrong, by the way, so let's get that out there.
Bilawal Sidhu:
So let's go.
Jim O’Shaughnessy:
But I definitely think, again, the most utility that a particular type model or particular arrangement yields that ... Those models usually win, right? Will some totalitarian systems use that to crush the people? Yeah, and that really sucks. But do I think that the cognitive diversity of a country that is free like the United States ... Do I think that that's the one I'm going to bet on? Yeah.
Bilawal Sidhu:
Hell yeah.
Jim O’Shaughnessy:
I'll definitely bet on that one, right? For example, what do you do if you're sitting in Beijing and you type in Tiananmen Square?
Bilawal Sidhu:
Oh, man.
Jim O’Shaughnessy:
Or you ask their version of ChatGPT, "Hey, tell me about Tiananmen Square." And it's like "No."
Bilawal Sidhu:
Totally. Oh, by the way, and you've been flagged. And, by the way, somebody's going to come to your door.
Jim O’Shaughnessy:
Exactly.
Bilawal Sidhu:
And by the way, your rating went down a certain number of points and now you're going to get a shittier-
Jim O’Shaughnessy:
Your social, credit score-
Bilawal Sidhu:
Rate on your credit card.
Jim O’Shaughnessy:
Just went into the basement.
Bilawal Sidhu:
Exactly.
Jim O’Shaughnessy:
My point of view is, for the most part, all of these innovations are neutral. They can be used for bad, right? Like a knife or-
Bilawal Sidhu:
Oh, totally.
Jim O’Shaughnessy:
For cutting food-
Bilawal Sidhu:
Slab of butter or stab somebody.
Jim O’Shaughnessy:
Or stab somebody, right? You forget that it's the way the human uses-
Bilawal Sidhu:
Yeah, wields it.
Jim O’Shaughnessy:
This particular tool that determines the outcome. Now I'm very long human creativity, human freedom. There's a guy, Julian Simon, who wrote a book called The Ultimate Resource and it was human creativity. And he actually did a bet with one of the Club of Rome doomers for ... Back in, I think, 1970, in which the doomer was saying, "We're running out of everything, we won't have any of these incredible things." Julian Simon was like "Dude, you are so wrong. You're forgetting about we humans. We're really good at innovating and figuring shit out out." And so they actually did a bet. And Simon let the other guy ... Ehrlich I think his name was. He let him pick the super-rare minerals. Simon said, "10 years from now on an inflation-adjusted basis all of them are going to be cheaper." And this other guy is like "You're out of your mind, man."
Bilawal Sidhu:
Fixed resource is going to get-
Jim O’Shaughnessy:
Simon won.
Bilawal Sidhu:
Of course, yeah.
Jim O’Shaughnessy:
Simon won, right? This idea of scarcity, it's like humans for most of our history lived in scarcity. That's in our programming too. That has conditioned us for millennia. And now we are getting into a situation where digital, at least Abundance, is a possibility. I call it the digital divide, right?
Bilawal Sidhu:
Totally, yeah.
Jim O’Shaughnessy:
The EU, to me, seems to be committing suicide. What's going on?
Bilawal Sidhu:
And they're empowering the people at the top of the totem pole, right? Who has the ability to go implement DMA and GDPR compliance? It's the large tech companies that have huge teams. And it's going to stifle innovation from the startups that ... We're seeing the cracks of cloud calcification already in some of these giants. You're preventing sort of the natural antibody process, or whatever you want to say it, from circulating in the free market and sort of dethroning and putting a better option forward. It's weird, yeah.
Jim O’Shaughnessy:
Exactly. That's why if you understand the underlying architecture of a system you can make reasonable assumptions about what's going to happen, right? Back in 1982 I was writing about ... And all my friends teased me and said I was a lunatic. I'm a big journal keeper, and this is the one that actually includes this prediction that I made. This is from 1982-
Bilawal Sidhu:
Nice.
Jim O’Shaughnessy:
When I was 22. And I was writing about the fact that the old Soviet Union had to collapse. And it had to collapse because the underlying system that maintained it was very fragile, right?
Bilawal Sidhu:
Yeah, yeah.
Jim O’Shaughnessy:
People get-
Bilawal Sidhu:
A lot of people were like "That could never happen." And then it-
Jim O’Shaughnessy:
Bingo.
Bilawal Sidhu:
They're like "Oh, you're too idealistic," right, "This can never happen."
Jim O’Shaughnessy:
And then it happens.
Bilawal Sidhu:
Then the rest is history.
Jim O’Shaughnessy:
Right?
Bilawal Sidhu:
Yeah, totally.
Jim O’Shaughnessy:
Yeah, exactly, exactly. Well, I could talk to you for hours, this has been so much fun.
Bilawal Sidhu:
Likewise.
Jim O’Shaughnessy:
I'm getting the hook from my nanny here. At the end of all of our podcasts we ask our guests ... We're going to make you the emperor of the world ... Except you're not one of those guys who can shut everything down. You can't have one powerful AI that you control and everyone only gets access to that. You can't put anyone in a re-education camp and you can't kill anybody. We're going to hand you a magical microphone and you can speak two things into it, two things.
Bilawal Sidhu:
Okay.
Jim O’Shaughnessy:
You are going to incept the entire population of the world. Everyone is going to wake up the next day with those two things in their minds and one big difference. They're going to say, "You know what? Unlike all the other times I'm actually going to act on these two things." What are you going to incept into the world's population?
Bilawal Sidhu:
One hell of a question. You've certainly eliminated the obvious ... My dictatorial tendencies will now need to be suppressed. Just kidding. Let me think about this for a second. So I know one of them. Probably two. If you are afraid of something just go and try it. I think, as you mentioned, we are very averse to change, we're very versed to uncertainty. And I think especially in this time where the rate of change is only increasing. And every couple months something new comes out that makes you re-question your fundamental assumptions about the world. Stop trying to predict what's going to happen and just go play with this stuff, whether you're a knowledge worker, creative.
Bilawal Sidhu:
And the moment you do that it will stop seeming like this Kaiju monster that's going to stop you from the hyperbole that even creators such as myself perhaps perpetuate, and certainly the media does. And you'll realize it's more like this chaotic golden retriever, I like to joke, that you can coax into doing cool stuff for you to forwards your agenda. So just go play with the tools. Especially if you're choosing to push back against this, maybe go dabble with it first. That would be one thing, especially for the folks that are very resistant to the changes that are happening.
Bilawal Sidhu:
I think the second thing I'd love to change, and this really is ... I feel like we need to have a way for people to actually own their data. We're taking shards and facets of ourselves and putting it into all of these amazing platforms, right? Google has a facet of my human activity and my thinking in it. Apple has a huge facet of it, right? All these social media platforms have different facets of what I'm doing. Twitter is sort of like my real-time journal. And I wish we could empower users to somehow take the aggregate of their digital twin, such that they own it, and they can query it and take it with them. I think the technology for that exists, the incentives do not, at the moment. And I wish the incentives did. And we can talk about why. We've either got the paid Apple model or we've got the ad-funded model, and nothing else seems to be lucrative at scale. That's another thing I could change.
Bilawal Sidhu:
Maybe the third piece if I could ... I know you said only two. But if I drop the first one. The third one I would say to platform creators, especially who own distribution. If they could wake up and realize this treadmill, exhausting treadmill that's slowly getting cranked up ... It's like boiling the frog, nobody notices it just yet. I think if they could find a way that optimizes, not for watch time, not for revenue but some sort of ... I don't know what the right metric is but creator wellness is being a proxy for it. Like satisfaction from creation where you don't feel like you're feeding the beast.
Bilawal Sidhu:
After breaking some scoop on this podcast, I've talked to a bunch of journos, and they all talk about "Oh, well, you know how press is, it's like feeding the beast." I'm like "That's an awful metaphor. Why do you have to keep feeding the beast? Can we make this something that's more symbiotic?" I think pathfinding and finding that's ... Model that keeps in mind advertisers, consumers, and creators, with them at the center of this multi-sided marketplace, would be very exciting. And I think the incentive right now is to keep the money printing machine going. And I hope some of these large companies will not be worried about short-form video competition or whatever and actually try to find an experiment to find something that really does ... That is good for the mental health of creators and consumers alike whilst giving advertisers what they want. Those would be my two if I had to refine the third one.
Jim O’Shaughnessy:
We are very simpatico because I agree with all three of those. On your first one ... Joseph Campbell, The Hero's Journey, had a great quote which is "The treasure you seek is in the cave you fear to enter."
Bilawal Sidhu:
Oh, yeah, totally. The obstacle is the way, in a way.
Jim O’Shaughnessy:
Exactly, exactly. Well, this has been amazingly fun I was really looking forward to it.
Bilawal Sidhu:
Thank you for having me.
Jim O’Shaughnessy:
You did not disappoint. Quickly, where can our listeners and viewers find your work and you?
Bilawal Sidhu:
Totally. Go to my website, Bilawal.in. that's B-I-L-A-W-A-L.in, I've got links to all my stuff there. If you're interested in being entertained check out my TikTok and YouTube channels, Billyfx, Billyfx. If you're interested in being educated about the cool new research, Primitives Creative Technology, that's coming out there. You can follow me on Twitter slash X, or LinkedIn. I've also got a newsletter called the Creative Tech Digest, you can check that out as well. And it's all neatly linked on my one-pager website, Bilawal.in.
Jim O’Shaughnessy:
Perfect. Thank you so much for all of your time, learned a lot from you.
Bilawal Sidhu:
Thank you for having me, it was a pleasure.