By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
EVERY SPOKEN WORD
70 min read · 13,644 words- 0:00 – 2:45
Intro
- SSSteph Smith
Hey, everyone. Welcome to the latest episode of the a16z Podcast. This is your host, Steph Smith. In today's episode, I am thrilled to bring on Karen Cheng. Karen has been a creator for a long time, but she specifically has a knack for integrating technology into her creations. And you can only imagine, with the rise of some of these AI tools, like DALL-E, Midjourney, and Stable Diffusion, what she's been able to achieve. Now, if you missed AI summer and its impact on creators, you might have been sleeping under a rock, but you're not too late. It's truly just the beginning. In fact, I interviewed Karen back in September, and with every week, there was something that popped up where I thought, "Man, I wish I could have asked her about this." And yet still, in this interview, we cover a lot of ground. We'll chat with Karen about how she originally found her niche, how she manages to grow her social following, with a nod towards being additive instead of reductive. You'll also get to hear the behind-the-scenes of her many viral works, including a video of her becoming a lawnmower. Yes, truly a lawnmower. Her AI-generated Cosmo magazine cover, the first ever, her DALL-E fashion show, her transforming iconic art into 3D museums that people can explore, and much, much more. And by the end of this episode, I think the audience will come away with some really key things, including an understanding of the new tools that are available, that are literally at our fingertips, how AI can indeed enhance the creative process, and maybe even some of the second-order effects of these innovations, like how creators are paid or how they might work with these AIs. And I should note that since we're just at the beginning of this AI era, this is truly just the beginning of our coverage. So I hope you enjoy this episode.
- SPSpeaker
[upbeat music] The content here is for informational purposes only, should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any a16z fund. For more details, please see a16z.com/disclosures.
- SSSteph Smith
[upbeat music] I am absolutely thrilled today to have Karen Cheng on the line. Karen is a digital creator. She has taken social by storm. Seriously, she's all over my feeds. I can almost guarantee you've seen one of her videos. And she's done everything from high-five Mark Zuckerberg in the Metaverse. She's partnered with several A-list celebrities. Uh, she also was commissioned by Cosmo recently to create their first AI-generated magazine cover. So super cool. We'll be talking about that. And in particular, we'll be talking about her foray into AI, how she's using AI to be more creative, and how that relationship between humans and AI is evolving, how it enhances our creativity, or how it doesn't. We'll see.
- 2:45 – 6:29
Karen’s journey to content creator
- SSSteph Smith
But Karen, welcome to the show.
- KCKaren X Cheng
Thank you so much for having me. Thank you for that amazing intro.
- SSSteph Smith
I'm talking to you, Karen, in 2022. You have millions of followers across various social media platforms. But I wanna flashback to Karen from 2013. So I'm gonna preface this with I was going down this YouTube rabbit hole. I was not researching for this episode. I happened to stumble upon a video with Paul Graham and Sam Altman from 2013, YC Office Hours.
- KCKaren X Cheng
[laughs]
- SSSteph Smith
You know where this is going.
- KCKaren X Cheng
I, I can't believe you found that. [laughs]
- SSSteph Smith
So-
- KCKaren X Cheng
That's kind of obscure.
- SSSteph Smith
Yeah, so I'm watching this video. Someone else familiar pops into it. Ryan Peterson from Flexport is in this video, and I'm like, "Okay, flashback." And then someone else is in this video, and it's you, and you're pitching this company, Give It 100, which is basically your startup at the time. And the startup was structured around people sharing their journey in learning skills, sharing their progress every single day for 100 days. You also then, I think around a year later, you did a TED Talk where you kind of used that framework of 100 days to go and learn dance, and you shared that, and it was fascinating. People can look that up. We'll have it in the show notes. Okay, so this is a broad question, but how do you go from that, right, 10 years ago almost, uh, where you're building this startup, to where you are today? When did you... Was there a point in time where you decided, "I wanna start sharing this information online. I wanna start being a creator"?
- KCKaren X Cheng
Okay, so yeah, this was back in my, like, failed startup days. And so I, like, really badly, very badly wanted to get into YC, as, like, everyone did. And [laughs] um, managed to get on Stage Office Hours. Spoiler alert, did not get into YC, was turned down, I think, multiple times. Um, was interviewed on stage, um, and answered their questions. That startup eventually failed like a year later. Um, I think it was really tough on me at the time because up until then, I had like, you know, good academic achievements in school, and it was my first... Like the, I think your first startup failure is, like, your first definitive failure. It's like-
- SSSteph Smith
Yeah
- KCKaren X Cheng
... there's no [laughs] it's like, you failed. [laughs]
- SSSteph Smith
[laughs]
- KCKaren X Cheng
There's no like-
- SSSteph Smith
Like it's over
- KCKaren X Cheng
... it's like, well-
- SSSteph Smith
Yeah.
- KCKaren X Cheng
Exactly. Like, right? And so I sort of had to like, like, come to terms with that. But, um, honestly, like, going through that startup, I learned so much about video that I then applied to, like, doing my own stuff. And actually, I, I got this really interesting feedback when I was, uh, trying to, like, get business for my startup, and I was talking to someone. I was trying to, like, get them to invest in my company, and they were like, "We're not interested in this company." Like, they, they could see that it wasn't a scalable business. But they were like, "But we actually just wanted to hire you to make some videos." And I was like, "I don't wanna do that." Like, but then actually, it turned out, you know, years later, I'm like, yeah, that person was actually right. And, um, I, I think it just took, it took me trying every single career to realize, like I tr- I've, I've tried working for a huge company. I've tried working at a startup. I tried being a founder of my own startup. I tried running an agency. I tried, like, I've tried everything, and I've, I've come to the realization that for me, actually, like, being an influencer is the best thing for me, which is crazy because I, I always swore I would never be an influencer. [laughs]
- SSSteph Smith
[laughs] Well, I know exactly what you mean because I feel like, um, I also create things on the side, and for the longest time people would associate me with certain things. They'd be like, "Steph, you're like the writing girl." And I was like, "No. Please, please, please, please, I'm not the writing girl." [laughs] That is not how I picture myself. But ultimately, sometimes you just have to lean into what other people see in you.
- 6:29 – 9:30
Creative unlocks for viral content
- SSSteph Smith
And so you started creating these videos. Um, you've been doing it for years. I think you originally, um, uploaded your first Musical.ly in like 2017, and even in 2019, I think I saw you say in a video you had around 10,000 followers on Instagram. So that might seem like a lot to some people, but compared to today, where you're just smashing it, you have over a million followers, everything you create almost goes viral, um, was there a specific unlock where you found along that journey that like you noticed something that really worked or you, you honed in on something that people really wanted?
- KCKaren X Cheng
Yes, absolutely there was. So, um, for a very long time I struggled to grow my following. Um, first I resisted. First I was like, "Well, I don't even need to grow my following," so I was just making like viral one-off videos. Um, that was back in the days when you could make things go viral in a completely different way that doesn't work anymore. So back in like 2012 up until around 2016, you could make videos go viral just by coming up with something clever that would be a good headline, cold emailing like 300 reporters, getting them to write about it, and then it would almost guaranteed go viral. Um, that does not work anymore because Facebook ate like all the reporters' jobs. So, um, and now videos don't spread through reporters, they spread through algorithms. And so it is actually so, so competitive now. It's so much more competitive now than it was like 10 years ago. Uh, it is h- it is much harder to make a viral, a video just go viral. And so the way to do it now is you, you kinda have to build a following. And so it's not, it's not so much about making everything go viral, making things go viral, but it's more about just like building an established follower base so that your baseline of like the number of people who see your videos or your work just steadily gets higher. And I really struggled to grow my following. Um, I remember when I first tried growing my Instagram following, I was like, "Why is this so hard?" You know? I would get things to go viral, and it would be very hard to go, to, um, to get any followers out of it. Uh, the, the unlock for me was actually sharing my behind-the-scenes, and that was it. And so I would like sh- I, before I would just show these cool shots, and I'd be like, "Look at this cool shot," you know, and then maybe I would describe how I did in the caption. And like nothing, crickets. If I showed the same shot, the exact same shot, but showed the behind-the-scenes with it, all of a sudden it was going viral. And I've done that over, I've done that so many times now that I am confident that that was the, like the defining thing. There were lots of other things that I learned along the way to grow my followers, but if there was one thing, it would be that.
- SSSteph Smith
That's really interesting because a lot of people participate in these online worlds. A lot of people are creators, and a lot of the stuff is very similar. So sometimes all it takes is like one iteration, one like s- adjustment in what you're doing. You're still sharing the same kinds of videos, but then just giving that bes- behind-the-scenes look is completely changing the game for you. I also wanna just quickly get your take on something you mentioned, which is that social media is
- 9:30 – 12:21
The changing social media landscape
- SSSteph Smith
not the same, right? Like we, we use the same term from 10 years ago, but even if you think about social as like a word, social media isn't not really social anymore in a way. It's kind of gone from the social graph, where you followed your friends and your family, and then now it's an interest graph, uh, where you follow just things that this algorithm is serving you. Have there been any other changes that you've noticed in terms of just like bird's-eye view how social has changed that really matter to creators? Like is there a way that we're th- a lot of creators are thinking about this art incorrectly?
- KCKaren X Cheng
I think it's gotten a lot more competitive. I think people are a lot more aware of the downsides of social media now, so we're no longer in like the honeymoon phase of it for sure, [laughs] you know? It, it's kind of interesting that like for most of history, I feel like humans have prayed to God, you know? But now I see that like content creators, we pray to the algorithm. It's almost like a superstition. It's like, "What does the algorithm think? My video's not doing well because of the algorithm." It's like, "How can I optimize my video for the algorithm?" And it's like we're no longer, it's like we're almost no longer designing ourselves for like human taste, but we're designing ourselves for like the algorithm, and I do think there are quite a few downsides to doing that. Um, and there are good reasons to resist that as well.
- SSSteph Smith
Yeah, yeah. I think you're right about that, and, and I think increasingly those algorithms are becoming black boxes where, um, you think, you know, one variable is going to make your videos or your work go viral, and then you play into that. But you don't really know, right? Like you're not there seeing the, the different levers that control the algorithm, and so it really can consume you.
- KCKaren X Cheng
Yeah, the black box is actually what it makes it feel like a cult or a religion because it's like you don't know. It's like a mystical force.
- SSSteph Smith
Yes, exactly. And even some people who work at these companies say that they don't know exactly what plays or not so much what plays into the algorithm but the output of it. One of the ways that you've stood out is, at least recently, playing around with AI, right? I, I think you're one of the few creators that I've seen that has really mastered the art of partnership with AI as it relates to social. Um, and there'll be many more, of course, right? But, um, how did you start getting into that? What, what led you towards those kind of creations?
- KCKaren X Cheng
Um, you know, I was really just feeling quite stuck in my work. I felt like I was doing... Basically, like almost all one million of my followers are from the same video. I mean, it's a few videos, but it's the same kind of video. Um, it's my low budget-
- SSSteph Smith
Can I guess which one it is?
- KCKaren X Cheng
Yeah, go for it.
- SSSteph Smith
Is it, is it the phone matrix one?
- KCKaren X Cheng
Yeah, it's like the low-budget camera tricks.
- 12:21 – 15:23
Innovating with AI
- KCKaren X Cheng
So, uh, there was one video actually where I attached, like, a phone to a ceiling fan, and then made that into, like, a Matrix bullet time effect, and I got 300,000 followers from that one video. So e- every time I got a surge of followers, it was the same thing. It was like I would, I would attach a phone to some household object, find, like, some, like, hack to do a camera trick with it, and, um, people would follow me for that. And I was just like, I'm, like, out of household objects. I'm out of ideas. Um, and I just felt like I was... I just felt like I was running out of ideas, and I was like, I have to find new, like, tools to... New toys to play with. And in the meantime, I was, I was starting to see some of, like, the really insane AI stuff coming out in white papers. But it was all, like, two-minute papers, like, research videos of researchers, like, shooting whatever they had in their bedroom. And of course, they're so focused on the research, it's not their job to think about the storytelling or the cinematics of it. And I was like, "Huh, that's interesting. I don't think anyone is actually doing that yet, so I could try doing that." And so I started experimenting with, like, well, let's take the research from white papers and make it a social media-friendly video, you know? Um, and I started doing that at the beginning of this year, and it was just so fun. It was, it was... I think I'm s- um, I'm a, like, a very novelty-seeking personality, and, like, my greatest motivation is like, can I try to do something that hasn't quite been done before? And sh- like, yeah, sure, like, probably someone has done it, but can I find my own, like, original twist on it? And so the fact that, like, these white papers, no one was really, like, using this technology, no one was really using it in creative ways. So I could be like, wow, I really am, like, the first to do something with this. That was very exciting for me, whereas trying to come up with, like, the latest iPhone transition, all the ideas are taken. It is very, very hard to think of something clever or new with an iPhone camera trick. TikTok transition, very, very, very hard. But with the AI stuff, it's like this is just free for all right now.
- SSSteph Smith
Yeah. Yeah, and, and I know that a lot of these tools are becoming more democratized, more available, but there was, I think, a little bit of a technical gap too, right, in order for you to actually, like, start playing with these AIs relative to the average creator. Is that right?
- KCKaren X Cheng
Yes, there is, and I'm not that technical. Um, I took like, a coding class once in college, but I prefer not to get into that myself. So I actually work with really talented programmers who help me run it, or they'll teach me to set it up. Um, and so that's why I think, like, when DALL-E came out, I was like, oh, this is requiring no technical ability. Like, this is... If you can do a Google image search, you can use DALL-E, you know? And so that was really interesting for me.
- SSSteph Smith
Yeah, and I wanna talk about those, like, second, third-order effects of having that
- 15:23 – 18:27
Unique AI tools available today
- SSSteph Smith
fully democratized. But you mentioned DALL-E. I've also seen you use a ton of other tools. So I'm curious to hear from you, like, what are some of the other tools that maybe others haven't heard of or that you're playing around with? Or maybe, like, DALL-E is the first AI tool that you use, but I think you also use other ones to, like, clean up a video or to, like, iterate on the creation from DALL-E. So can you share maybe some of the tools that you're playing around with?
- KCKaren X Cheng
Yes. Uh, so there are a few. With the image generations ones, which are really popular right now, there's DALL-E, there's Midjourney, there's Stable Diffusion, um, there's Disco Diffusion. I think if you're actually watching the video version of this podcast, by the way, we'll have, like, visuals of all these-
- SSSteph Smith
Yes
- KCKaren X Cheng
... somewhere.
- SSSteph Smith
We'll bring all of them up. Yeah. [laughs]
- KCKaren X Cheng
Um, and so those are really popular right now. I also think there are other ones that are l- um, they require more technical knowledge, but they are, they enable kind of more, um, uh, techniques that are, that are less mainstream, and so therefore more unique if you can pull them off. So one of them is DAIN. Um, DAIN is really simple. All it does is it applies artificial slow motion to your video. Um, but I thought it w- could be used in really interesting creative ways by, like, actually shooting, like, a custom stop motion video for the purpose of it being, um, interpolated, um, to create m- these impossible movements. And you'll just... I can't describe. [laughs]
- SSSteph Smith
I think I saw one of yours. It was, like, uh, you as a lawnmower or something, right?
- KCKaren X Cheng
Yeah.
- SSSteph Smith
Like, you're, like, lying on the ground, and then you basically get dragged across, and that's-
- KCKaren X Cheng
Right
- SSSteph Smith
... is that using DAIN?
- KCKaren X Cheng
Right, exactly. So, um, if you're just listening to this, imagine, like, a stop motion of someone, like, on the ground, like, uh, planking basically, and then just, like, slithering across the ground in stop motion, right? But then with, with DAIN, you can make it perfectly smooth, and so it looks like I'm literally being dragged on the ground by, like, an invisible string or something. Um, but actually that was just a stop motion video that w- we used AI to smooth out. And so I did that with James Perlman. So he helped me out with, like, smoothing everything out. Um, other tools are NeRF. So NeRF has gotten pretty popular. NeRF is a way to use any camera, like your phone, to scan a scene, and then all of a sudden you have this beautiful 3D, like, scan of it. And the... what's interesting about it is it's different than photogrammetry because it can handle, like, light. It's basically constructing, like, a particle light field. And so when you look at things from different angles, the light changes, and so that's why it can handle mirrors, whereas, like, traditional photogrammetry can't. Then there are other tools too that we've played with. Um, I recently discovered EBSynth, which I actually learned is not an AI tool. Um, but I use that to basically figure out how to make DALL-E work for, um, video. So DALL-E is photo only, but using EBSynth, I was able to make DALL-E work for video.
- SSSteph Smith
Very cool. And, and we'll talk about this later, but you ended up doing, like, an AI fashion show, right?
- KCKaren X Cheng
Yes. Yes
- SSSteph Smith
... w- with the combination of many of these tools. I wanna hear from you,
- 18:27 – 22:03
How AI tools might differentiate
- SSSteph Smith
though- A lot of these tools are quite nascent. We're gonna see a lot more of them come to market. How do you see them all playing together? Because let's just use, like, the text-to-image generators. So you have DALL-E, you have Stable Diffusion, you have Midjourney. There'll be many more. A lot of them, um, are similar, but also what I'm imagining if we, like, extrapolate many years, like, how will they differentiate? How will you as a creator, or how will consumers-
- KCKaren X Cheng
Yeah
- SSSteph Smith
... decide I'm gonna use DALL-E or I'm gonna use Stable Diffusion? Will they be different business models, or how do you think that's gonna evolve?
- KCKaren X Cheng
Um, I can speak from the creator side maybe more from the business side, but, um, I think right now, AI in the, in the m- in the mass media, there's sort of this con- misconception or perception that AI is just this, like, all-powerful thing that's gonna replace humans. Um, um, but I think a more interesting way to think of it is that they are, they are tools, and each one is a specific tool for a specific purpose, just like I would have, like, a calculator or a compass or something else. S- and you can combine... And, like, what, where you get interesting things is where you combine the tools, right? So for example, um, you can generate an image in DALL-E, and then you can use CapCut, which is a program, which is, like, an, an app on your phone that will literally turn any image into, like, a 3D image. Then there's another problem, which is, uh, like DALL-E and Midjourney that, and all, I think all of these image synthesizers have, which is that they make faces hu- they don't do human faces very well. There will be, like, artifacts in the eyes. They'll just look kind of droopy or a little messed up. Well, guess how you can fix them? You can run them through Facetune, the app that all influencers use, and that'll make, that'll make, that'll fix it. And so I think-
- SSSteph Smith
And does it actually, Facetune, can fix it to that degree? 'Cause I've seen-
- KCKaren X Cheng
Yeah
- SSSteph Smith
... some of the outputs from DALL-E as an example, and they're really kind of wonky.
- KCKaren X Cheng
It'll fix them to be, it'll fix them to be flawless.
- SSSteph Smith
Wow. Okay.
- KCKaren X Cheng
I mean, it depends on the image, but, like-
- SSSteph Smith
Yes
- KCKaren X Cheng
... I have seen, I have seen it do it flawlessly.
- SSSteph Smith
Okay.
- KCKaren X Cheng
It's like, you know, you could, you could be the AI researcher who's, like, working really, really hard to, you know... It, it gets harder and harder to reach perfection, right? So you could be the AI researcher trying, like, really, really hard to get that last 1%, or, like, a creator could just be like, "I'm just gonna run this through Facetune." Bam.
- SSSteph Smith
Yeah. Yeah, yeah.
- KCKaren X Cheng
And it's, it's-
- SSSteph Smith
I also wonder whether some of these tools become more specialized, because these text-to-image generators are quite broad, right? They're being trained on, like, all of the internet, and they're being used for use cases of, like, literally, again, all of the internet. And because there's so, there's going to be so many of them, I do wonder whether they become specialized. Like, maybe some of them become specialized for just, like, hyperrealistic human photography, and other ones become really specialized for really, like, generative, um, illustrations or something like that. Um, do you think that might happen, or do you think they'll continue to stay really broad?
- KCKaren X Cheng
You know, different companies will have different strengths, and then they'll continue to be able to differentiate by sort of specializing in that. But I do imagine that some of these companies are gonna buy other companies, and then so they'll, like, you know, kind of have more verticals, kind of like how, you know, Meta bought, like, WhatsApp and Instagram. I see that happening with AI tools too.
- SSSteph Smith
Right. Yeah, so, like, if we use Meta as an example, they basically tried to own messaging, and so you might see, like, a similar phenomena with
- 22:03 – 25:53
Cosmo’s first AI-generated magazine cover
- SSSteph Smith
AI tools.
- KCKaren X Cheng
Mm-hmm.
- SSSteph Smith
Let's jump into a really cool example of how you deployed AI, which is something I mentioned before. You got commissioned by Cosmo magazine to create their first AI-generated magazine cover. We'll throw it up on the screen. It looks awesome. Um, and you've also done, as you talked about, a behind-the-scenes explainer. But for those who are not familiar with this project of yours, can you share a little bit more about how it went down?
- KCKaren X Cheng
Uh, yes. So, um, this was kind of shortly after DALL-E was released. The people at OpenAI gave me a call, and they were like, "Hey, we have this really cool potential opportunity where Cosmo wants to make the first ever AI-generated magazine cover." And they were like, "And we want you to generate it." I'm like, [laughs] "What? That's crazy." Um, and so, I mean, my first reaction was, like, uh, worried, I guess. Like, I hope I can do a good job. Like, um, because keep in mind, like, it's not... I, I can't emphasize enough how much, like, the human is needed in this process, you know? Like, yeah, they could have just... Y- you can't actually have a fully AI-generated magazine cover, 'cause you have to give a prompt. And I guess you could theoretically ask, like, an AI to generate the prompt, but still the human needs to, like, a human needs to give some input, right? And so for, for our process, actually, we went through so many different rounds of it- of iterations. And, um, eventually we landed on a creative direction of, of woman astronaut. And this was actually a, I thought, a really clever workaround by OpenAI to not show a human face. Because at that time, human faces were not allowed in DALL-E. In the early days, human faces were not allowed for DALL-E. And so we... And we also were like, well, what race is this woman gonna be too? Like, we didn't want to make a race statement. And so by making it a woman astronaut, it could be any ethnicity as well. Um, and so, so they had that direction, and then they sh- they showed me a reference image, uh, that, that was generated in DALL-E. And I was like, damn, this is so good. I don't know that I can even top this image. And I was just like, I, I don't know that I can make a better image than this. And so I spent hours trying to make a better image than that, and I could not. And I was like, ugh. Like, they hired me to do this. I obviously have to do a better job than, like, what they quickly threw together. So I just kept going for, like, a few more hours. And then eventually I, like... I was like, "Okay, I, I, I kept on troubleshooting it." I was like, "Okay, I want this woman to, like, walk with swagger." And it took me... I kept on saying, like, walking confidently, walking proudly, like, walking energetically, and none of those prompts were working out. And then eventually I was like, walk with swagger, and then boom. Then I got, like, the h- then I got the hip, like, thr- thrust. So, like, it was like... You know, it was like I, I, I kept on asking for walking confidently a- and that didn't work, but then when I a- asked for walk with swagger, it indeed gave me swagger. And then the other thing I was thinking of, like, was, um, there's, like, an influencer tip that's popular on social media for taking photos, and it's like the, the hack is that you, you bring the camera to the ground and shoot from the ground up to make it look really dramatic. And I was like, "Oh, let me try that." So then I was like, how about wide angle shot from below? And then boom, it made it way more dramatic. And then the image that I got, actually, its face was sort of messed up, so then I did inpainting and stuff to, like, fix the face, and, and then I did inpainting to, like, extend the boots. So I did, I did quite a bit of, like, uh, fixing up of the original image. It was... If you look at the original image, it was, it was quite a bit of a diamond in the rough. Um, so if you look at my behind-the-scenes video, there was, uh, a l- a significant amount of human work to make it happen.
- 25:53 – 34:06
Inpainting, outpainting, and the fungibility of artists
- SSSteph Smith
Yeah. And we'll share that. When you say inpainting, just for people who aren't familiar, you're literally... You're still using AI for that, right? But you're basically selecting a portion of the outputted image. You're saying, "Okay, I like a lot of this, but I don't like this area. So, like, can you, can you inpaint, can you replace that specific region?"
- KCKaren X Cheng
Inpainting is like magic Photoshop. Like, if Photoshop worked magically, that's what it would be. So imagine there's, like, a photo, and let's say you don't like what you're wearing or you don't like the person in the background or you don't like anything about it. You erase that, you type in what you want, and then boom, you get three options.
- SSSteph Smith
Which is so incredible. And that's what you also used partially for your fashion show. So again, we'll, we'll talk about that later. But I wanna get a sense of, you talked about the fact that humans need to be involved and that it took several iterations. Like, how many iterations are we talking in this, in this spectrum or this, this route to the final image?
- KCKaren X Cheng
It will depend a lot from artist to artist. You know, I think some people, they'll share their first prompt, and that's, um, that's interesting. Some people will spend, you know, maybe five to 10... They'll type five to 10 prompts. And then for some people, where they're really trying to achieve a very specific result for a specific project, I mean, will be in there for, like, hours and hours, um, running, like, dozens or hundreds of prompts. And in the early days of DALL-E, it was unlimited and free, so [laughs] it was a little easier to do back then.
- SSSteph Smith
Yeah, and I think it's... I, I don't know if they... I think they might have changed their business model recently or their pricing, but I saw an article, um, that this, this woman was basically creating, like, a llama dunking a basketball. Um, so we'll pull that up, too. But I think that one took 100 iterations, and I think it was, like, around $13, um, with that pricing. So that gives you an idea. And as you said, it, it'll depend. But I think the, the maybe more interesting question is how fungible was Karen in that process? And I don't want to dismiss you-
- KCKaren X Cheng
Yes, how replaceable am I? Completely-
- SSSteph Smith
But if you gave average Joe the same project and the same tool-
- KCKaren X Cheng
Right
- SSSteph Smith
... do you think they would come up with anything comparable?
- KCKaren X Cheng
I think a lot of people would've been able to make a really good Cosmo cover. Um, I think those people would have been... The people who would've done it best are people who have a good artistic eye, have the kind of the patience and the motivation to keep going and keep refining on the prompt, and who can, like, describe what they see in their head. And so a lot of these people are not necessarily traditional artists. You know, I think a lot of people could have made a really, really good result. I, there's actually a, an example that I have that we can show up, that we can show on screen. I, a while ago, wanted to expand The Girl with a Pearl Earring. So you may have seen one thing that you can do, uh, that was actually just announced, is you can expand paintings in DALL-E. Um, and so you can basically, like, take a famous painting and then imagine what was all around it. So you can take, like, the Mona Lisa and imagine, like, where was she? You can take The Girl with the Pearl Earring and imagine, like, everything that was around her. And so I made a video a while back where I imagined her, like, in a, like, library, and she was holding a book. And I was like, I want her to be, like, an educated woman, you know? Um, and so I made that. Um, and then I actually, um, am working with a project, uh, with OpenAI actually right now. It's one that they commissioned. Uh, it will... It's not out yet, but it will probably be out by the time this podcast releases. So if you go on my Instagram, uh, on KarenXCheng, you'll be able to go and see this filter. We made this Instagram filter where we expanded these famous paintings. And, um, you can actually go inside through them and see these famous paintings. And I looked at my Girl with the Pearl Earring, and I was like, [lips smack] "This can be done better." And so I hired August Kamp, who is, um, the person who taught me about this method, and she is, she's has spent way more time in DALL-E than I have. She's so artistic and talented. And then when she showed me, I hired her to do this Pearl Earring, and when she showed me hers, I was just, like, moved to tears almost, like jaw drop. I was like, "That is what you made?" And so I felt like she was irreplaceable in this. I mean, I'm sure... I, I could have hired 10 different artists to do this Pearl Earring and gotten 10 very different results. And what I like about this m- example is that everyone is starting with the same source image. And so it, it almost establishes, like, a little bit of, like, a control, you know, for it. And so, um- I think this is the perfect case study for like, hey, different humans get very different results with AI.
- SSSteph Smith
Yeah. It's, it's a tool, as you said. But the reason I asked if you were fungible is because I just, I really wonder how this progresses in terms of we know it's gonna be somewhat of a democratizing force, because now a bunch of people who, like, couldn't paint or couldn't do Photoshop in certain ways, like, can now do it and create all of this art or, or, or these outputs. But I do wonder then, does that make a lot of people really successful or a lot of people really capable, or does it still surface the very, very best to the top, right? Where, where there's a different filter and now the filter is prompt engineering versus painting or drawing or singing or whatever it might have been in the past. Um, so do you think that's still the case, where you're still gonna see these, like, outliers who are just so much better than the rest? Or do you think it'll be more of a level playing field than what we were seeing before?
- KCKaren X Cheng
Okay. So I think that, um, what this is gonna do, what AI art is gonna do, is it's going to significantly lower the barrier to entry to become an artist. So the, to be an artist right now, you have to have a lot of time, um, a lot of training, um, sometimes the monetary financial need, means to be able to do that, or the willingness to be, like, a starving artist to do it, or it's a hobby on the side. Um, but there are kinda some... There's definitely a barrier to being an artist and, and now it's like everyone can do it, and it's almost like... I almost liken these image synthesizers like DALL-E or Midjourney to, like, a Peter Pan. But instead of stealing from the rich and giving to the poor, it takes the artistic skill of artists and it's, like, gives it to everyone. Like, here you go, you know? And so I do think that a lot more people are gonna be empowered to be artistic or to be artists because they didn't necessarily have the patience to go, like, learn oil painting. But now, but they actually do have the talent to be able to describe what's in their imagination and continue to refine until they get the result they want. So a lot more people are gonna be empowered to be artists. There will still be the standouts and the extraordinary people because they're gonna be the ones who are finding, like, different or creative or innovative ways to do it. There's, you know, as much innovation as there is, there's always gonna be the ability to combine things in new ways or, you know, find the latest. So I, I definitely think that, um, there's gonna be a lot more artists, um, a lot of unhappy artists, um, a lot of happy artists, and also still standout artists.
- SSSteph Smith
Yeah. Yeah, I mean, it's, it's almost just reshaping the, like, skill ecosystem. And, like, some people who have invested in skills for many decades, like if, if they were just, like, an A+ oil painter. Unfortunately for them, that skill is somewhat being replaced by this technology. Um, and it reminds me a little bit of, you know, if we use, like, Uber as an example. Like, some of those taxi drivers had invested in their taxi license for many, many years. And so there's kind of a parallel there. Doesn't mean it's bad or good, but it is just a fact that these AIs are, like, as you said, almost like cherry-picking skills and, like, dropping them on the rest of the population, and it's this, like, great experiment to see what comes out of that.
- 34:06 – 34:53
AI enhancing the creative process
- SSSteph Smith
Um, I also kinda wanna hear from you. As you worked with these AIs, was there any, like, aha moments? Or, like, did you feel like it was truly, like, a back-and-forth process, or did you feel like it was unidirectional? And let me explain what I mean by that. If it was unidirectional, you're basically, like, giving the AI prompts, you're seeing what comes back, and you're like, "No, no, no, that's not what I want." You have a vision. But if it's more bidirectional, did you get a sense of, like, gave it a prompt, you got something back, and you go, "Whoa, I, I didn't even think of that, but, like, I like that. I like..." So, so was it more unidirectional or was it more bidirectional?
- KCKaren X Cheng
It's 100% that second one. Um, it does feel like a collaboration. And so you get, you oftentimes get things you don't expect, and you're like, "Oh, I didn't think of that. L- let me go down that rabbit hole
- 34:53 – 41:01
The virality of optimism vs pessimism
- KCKaren X Cheng
now more."
- SSSteph Smith
Mm-hmm. So, I mean, that, that reminds me of something I wanna discuss with you, which, um, we've actually talked about before in, uh, a call we had, which is the perception of AI. So there's almost, like, two framings. And I know this is overly simplistic, but on one side there's, like, human versus AI, or, like, human versus machine. This is, like, an age-old trope. And then there's another, which is, like, human plus AI, right? Which is more optimistic. It's like, whoa, like, what can we do with this partnership? And you've told me that you feel almost, like, a responsibility to portray this technology, these tools, in a certain way. Maybe even, you know, I don't wanna put words in your, uh, mouth, but more of this optimistic versus pessimistic doomsday lens. So can you just speak to, like, why you feel that responsibility or almost, like, how you think, um, AI should be portrayed or how you want to portray it?
- KCKaren X Cheng
I've had to unlearn a lot of my bad habits in terms of, like, I am a trained viral video creator, and so in many ways I'm trained very similar to the way a reporter would be, which means that I'm rewarded for making clickbait headlines. And so when all this AI stuff came out, I knew that, like, my, my first instinct was like, ooh, let me make a bunch of, like, human versus AI. Like, let's, let's have the AI do something and a human do something, and let's have, like, judges critique and say which one's better. And, like, that was my first instinct to do that. And then I realized, like, well, what is that just gonna do? That's just gonna... It's just gonna freak people out, you know? Um, and I do think, you know, people are, there, there are reasons to be worried for AI, about AI. I think there are very many legitimate ones, and it will negatively impact some people more than others. Um, but AI is a technology. It can be used for good, it can be used for bad. And I just, I think it clicked that, like, th- there's gonna be... Every gravitational force for the media is gonna push towards the bad, because that is what's gonna get more clicks and engagements and views. And what's, what's not the natural gravitational force is to actually think of ways to use AI for good. And so I was also sort of in this turning point in my career where I had just spent, like, two years optimizing for growing followers, and I finally just sort of let go of that, and I was like, "What would, what would it look like if I stopped trying to go my- grow my followers? What would it look like if I actively stopped trying and just made, like, what I wanted to make?" And I was like, "Well, since I no longer care about growing my followers," which was, which was very freeing for me, I was like, "Well, I actually don't have to do this clickbait thing. I actually don't have to try to get the most views by doing this controversial AI stuff. I actually can just make, like, wholesome, positive, find creative uses of it." Because if it is a technology that can be used for good or for bad, like, why not? You know, it's more rewarding for me to make it for good, even if it's less viral. Oh, well. Doesn't matter.
- SSSteph Smith
Have you found that things have been less viral when you go that more optimistic route?
- KCKaren X Cheng
Um, I think on a one-off basis, each thing is less viral. I, I, there are, there are many pieces where I was like, if I had written the headline in a more controversial way, it would've gotten bigger, it would've gone bigger. But I definitely think long term, it's a better strategy for me to consistently put out positive stuff so people can, people associate that, you know, with the content I'm putting out.
- SSSteph Smith
Yeah. 'Cause I, I was just curious, I mean, you, you mentioned news as an example. Pessimism sells, right? Clickbait sells. It, it captures your attention. Fear is, like, a very natural human emotion. Um, but I do wonder if, to your point, like, more optimistic creations endure, right? So, like, people follow you because they're like, "Karen is this inspiring person. She's learned all these technologies. She's creative." So they view you as truly, like, this, this creator that they more so, like, want to be like instead of, I think pessimism sometimes is quite surface level, right? It does capture your attention, but you're not, like, inching to go back to, like, hear from that creator, right? It's very, like, ch- transient.
- KCKaren X Cheng
Yes. I think that's a great insight. It's like you can make something go viral, but it's a one-off, and people aren't necessarily gonna be inspired to continue to hear from you. Or you can continue to put out positive content, and each piece of positive content isn't gonna get as many views, but you will kind of form a more long-term relationship, uh, in people's minds. And there was actually this really incredible quote that I saw a while ago, and it actually, like, changed a lot of how I think about things as a content creator, uh, in doing social media, and I don't remember who said it, but the quote is, "Seek respect, not attention. It lasts longer." And I was like, wow. Because I was on this hamster wheel of chasing likes and views, and as anyone who has made a viral video knows, it, it, it, it's over so quick. You can have, like, a million views one day and then, like, it's, it completely dies down by the next. And so people get on this hamster wheel of addiction of, like, continuing to churn out content to try to get viral videos, and I was on that hamster wheel for years and just being like, "When does this end?" Like, "Why am I doing this? Why am I conti-" you know? And it's, it doesn't feel good, it's addicting. And then I saw that quote and I was like, oh. If I actually just put more effort into making more memorable, good content, I don't have to put out content as often. And yeah, I would much rather have the respect of a few people who I respect rather than the attention of m- the shallow attention of millions who
- 41:01 – 46:23
Turning attention into business
- KCKaren X Cheng
forget about it by the next day.
- SSSteph Smith
I mean, even on that point, if, if we even remove the, like, feel good, I feel respected, et cetera, part of it, it's also good business. And, and what I mean-
- KCKaren X Cheng
It is
- SSSteph Smith
... by that is, like, a lot of people view attention as, like, the end goal. Like, I want a million followers, I want this, uh, this number of people to pay attention to me. But ultimately, like, you need to convert that attention if you want this to be, like, a long-term game, whether it's, like, to find clients or to, like, sell a course or to, like, start a rolling fund, right? There's many, like, versions of turning attention into business, uh, or commerce, but that second part of conversion does not work if you're just doing these, like, viral things. And I, I operate more on Twitter, but as an example of this, like, um, something I've been pretty outspoken about is just, like, these tweet threads which are just, like, so overdone and, like, so templated, so, um, they work, but they're, they're just really, really surface level. And, um, ultimately, what I find is i- I see a lot of creators, like, feed into that, and they're like, "Okay. I'm gonna go all in on this thing that I don't believe in, I don't feel good about, et cetera. But it's gonna bring me hundreds of thousands of followers." But then again, coming back to the point, it's like, well, try converting those followers into something, you know, that, that generates you money or that, that is long-lasting and, and it's very hard, right? Because those followers don't actually love your content. They don't buy into you. They just happen to be like, "Oh, okay. Another thread." Like, they're, it's almost like a habit for them to engage with that kind of work.
- KCKaren X Cheng
I am so glad you said that because I, I, I feel the same way, and ultimately, this is my job.
- SSSteph Smith
Mm-hmm.
- KCKaren X Cheng
The millions of followers are useless to me if I can't make a penny off of it. Like, I'm not gonna put in all this effort. It's my job, right? I need to make money o- out of it. And I remember that I used to feel- ... really bitter when I would see, like, these TikTok videos blow up and get millions and millions of views, and they would be the most inane things. Like, for example, I'll give you an example of a TikTok video that can go really viral. Um, it's a video where someone's like, "Try to get the number of hearts and the number of comments to be the same." Right? And so everyone's just, like, it gets m- like, tens of millions of views because it's, like, a very smart engagement bait, right? It's a very clever engagement bait. And I'll be mad at that. Or I w- or a- you know what? There's, there's other examples that are less clever. There's, there's just straight-up garbage on TikTok that is getting millions and millions of views, and that is so demotivating for creators, VFX creators, artists who are putting, like, hundreds of hours into their craft to make something look really beautiful, and then it completely flops and doesn't get very many views on social media. And I would, I used to be so bitter about that. And then o- like, someone told me, and then I learned over time that this was actually very much true, is that the, the people who are putting out low-effort content all the time, getting millions of views, they're not the ones getting hired by, like, major respected brands. You know, they're not actually getting great opportunities off of their follower count. Um, whereas the ones who are ... I know people who make incredible content. They only have, like, 20,000 followers. They're making bank, you know?
- SSSteph Smith
Yeah. Exactly. No, I mean, I just have to say, I resonate with what you said so much there as a creator myself. I, again, I operate more with Twitter, but same thing. I'll see these threads and I'm like, oh my God, like [laughs] what is going on? Why are people liking and retweeting this? And why is this one account blown up? It's, it's really low quality compared to some other accounts that I follow where I'm like, this is gold. This is gold. Why is no one paying attention? Um, but yes, I think, I think it's not productive to focus on that because ultimately the people who are putting the best stuff out there are the ones who, if we take the whole value chain, right? Like, the whole process, um, and consider that it's a business, those are the successful quote unquote "businesses."
- KCKaren X Cheng
People put so much value on the number of likes and the number of views because it is the only number we can really see and judge ourselves by. And so it's like we've been trained since, like, kindergarten to put value in, like, the number judgment. You know, when you get, like, a 97 on a test. Wow, that's so good. If you get, like, a 60, that's a D. That's bad. And so we are, we are so trained to take that feedback of numbers as the truth. And then so when it come- social media comes along and you get lots of views versus a little, it's so easy to say, well, a video was great because it got tons of views, and it's not because it got very little views. And it took me years to learn that that is, it is such a bad metric, but it's the only, like, a, quote unquote, "objective one" we have.
- SSSteph Smith
I know. I know. But it's hard sometimes, as we talked about, it's hard to just, like, to check out of that ecosystem, to be like, oh, I'm actually gonna focus on the things that I wanna put out there, the things that I'm proud of, instead of feeding into, as you said, s- what some people might argue is an objective metric of, like, this is what people want. So you also have to accept that.
- KCKaren X Cheng
You have to go
- 46:23 – 49:54
IP and ownership in the age of AI
- KCKaren X Cheng
against so many natural human instincts.
- SSSteph Smith
Yes. Um, I wanna talk about another topic which I've seen everywhere. It's coming up in basically every conversation I have around creators, around AI, and it's the idea of IP, right? So we talked about, like, being hired for things, but also, like, um, owning your creations is something that's mattered a lot for not just independent creators, but if you think about, like, all the way up, films, entertainment, like, what matters is who owns the IP. And AI introduces a very, very interesting filter on this. And, and honestly, I don't know who, if anyone, has the answers, but, but what I wanna walk through is where is the value am- along the AI creation process? So, like, there's the, there's the information that the AI is trained on, which you could argue the output wouldn't exist without that. So maybe that's where the IP is. There's also the argument that the creators of the AI tools, right? OpenAI, Midjourney, et cetera. Well, the outputs also wouldn't exist if the AI tool didn't exist. And then, of course, there's the prompt, which you could also argue if someone doesn't enter a very, very specific prompt, the output also does not exist. And I've even seen some, a, a company yesterday, um, that was selling prompts because-
- KCKaren X Cheng
Yes
- SSSteph Smith
... it is valuable-
- KCKaren X Cheng
Yes
- SSSteph Smith
... to have the right prompts.
- KCKaren X Cheng
I, I've bought a prompt before.
- SSSteph Smith
There you go.
- KCKaren X Cheng
Yes.
- SSSteph Smith
So there's, there's probably gonna be a whole economy around this. So I wanna hear from you. Obviously, there's maybe no right or wrong answer, but how do you think about that? Like, where is the IP, and how on earth d- will we think about that as creators moving forward?
- KCKaren X Cheng
Yeah, I, I don't know. Um, I actually, I've thought about this, but I don't [laughs] I don't have the answers. I'm not a lawyer. I know that also, like, lawyers don't even have the answers. Like, C- Congress doesn't have an answer to this yet. It's all so new. Um, and I think it does, ugh. Yeah, I don't know. I think, um-
- SSSteph Smith
Well, let me rephrase the question to something that I think you can comment on, which is how creators get paid, how creators charge. So something that we talked about, um, before, was that for your projects, whether it was with Cosmo or we were actually talking about a project, you said that, um, you feel or you felt like it was really important that those projects were paid for, right? And that might sound obvious, but, like, as we talked about before, AI, in certain cases, might not cost that much. It might cost, like, $2 to, to run the prompts you're, you're looking for. Um, but Some of those projects historically, like I don't, I don't know how much Cosmo would typically pay for a magazine cover, but it's not $2, right? So how do you think about the way that you charge as a creator in this new ecosystem?
- KCKaren X Cheng
Yeah, I, I think it, you know, one of the things that when I work with clients I'll, I'll literally say to them, like, "Hey," like, you know, "if, even if you don't hire me..." Um, and sometimes it doesn't work out, and so I'm like, "Hey, you know, I'm, you know, my schedule doesn't work out for this project, but whoever you hire, like please pay them. Please allocate this budget." Because, um, we are at like a tr- we are at like a moment in history that's very vulnerable where like certain standards can be established, and I would like to see the standard established that if you hire a human to use an AI as a tool,
- 49:54 – 51:07
The downward pricing pressure of AI
- KCKaren X Cheng
then you pay the human.
- SSSteph Smith
Do you think that holds up though? And I ask that because I think, I think like many industries, the very best of the best, if there re- really is like a 1% or like a 10X prompt engineer, uh, they will get paid very well. We've seen that happen in many fields. But then you also see when things get democratized, like it's very hard for those price economics to not change, right? Because, 'cause you imagine, for example, that at the beginning maybe, yeah, maybe certain creators are, you know, holding their ground. They're saying, "We must be paid X." But then you're gonna have insert creator here that is like, "Actually, I'll take half of that, or maybe I'll take a hundredth of that price." I don't know. But-
- KCKaren X Cheng
I think that's a really good point. I think there's, um, incredible downward pressure. Um, and so what's probably gonna end up happening unfortunately is that there will be like the more well-known people who can still charge higher rates, but then for the vast majority of people, like there's such insane downwards pressure on it, and I do worry about that, um, for creators, and I don't really know how that is all gonna play out.
- SSSteph Smith
Yeah. It, it's really hard to predict, but
- 51:07 – 57:50
Is generative AI still a gimmick?
- SSSteph Smith
yeah, I, I wonder how some of these fields change. I mean, I wanna hear from you because you've started to work with some of these companies. At this point in 2022, is AI, or at least the tools that people are using, are they seen as more gimmicks or, or do you see companies like really investing in this and saying, "Oh, actually we're gonna, like we are gonna replace our photographer with AI"? Or, you know, u- to use Cosmo as an example, was working with you just something where they could like put their flag in the ground and say, "Hey, like we're innovative. We're, we're working with AI"? Or do you see them actually going and like doing every fifth cover with AI? How, how are companies thinking about this?
- KCKaren X Cheng
In 2022, it is most certainly a gimmick and a headline grabber. It is like, "Let's be innovative. Let's be first. Let's be on top of this." I would imagine in 2023 the answer to that will be probably be different. Um, people will actually s- start using it as their workflow, um, because it makes sense and not as a headline grabber. Really interesting, I saw that there was someone, an, a, an, a reporter who used a Midjourney art for his article, I think about Alex Jones, and then he went viral in a bad way in that he got like majorly attacked on Twitter for using Midjourney art for, um, a magazine article. Um, and he was attacked because he didn't hire like a human artist instead, you know? Like, um, it was I think The Atlantic. Uh, correct me if I'm wrong. Um, but I think what, what people didn't have context was was that de- the department he was in, the image that he would've used instead would've been like a Getty Images or like some stock photo. It wouldn't have been like a official ar- commission, right? So that was what he was replacing, but he still got attacked for it. So I think, you know, the first few, you know, the, the leaders go, uh, the arrows attack the leaders, right? So I think the first few people who are gonna start using AI are going to get backlash, but it's going to get less and less and less and less until it's accepted. I honestly discussing some of these downsides is very uncomfortable for me because I don't know what's gonna happen, and I do think that, um, it could be bad for a lot of people. And I personally try to, do try to stay optimistic and try to see the optimistic side, but like there's no denying that like it is going to be bad for a lot of people and it is gonna be good for a lot of people.
- SSSteph Smith
It is a game-changing technology. I've heard some people say that it is like the technology of our generation in terms of how it shapes things. I also heard, um, Sam Altman say, which I thought was like an interesting frame, um, a lot of people will talk about like, and we even did this at the beginning of our call. They'll be like, "When? When is this gonna like change things fundamentally?" Um, he didn't give, at least in the interview I saw, a definitive timeline-
- KCKaren X Cheng
Yeah
- SSSteph Smith
... because I think no one can predict that.
- KCKaren X Cheng
Probably smart. [laughs]
- SSSteph Smith
Yeah.
- KCKaren X Cheng
I shouldn't have given any numbers either.
- SSSteph Smith
No, no, no, but we were just having fun. But, uh, I think his framing was really interesting, which is that he believes that, uh, again, it's, it's like the game-changing innovation, and if you look at humans in like the spectrum of time that we've been on this earth, if we're talking 5 years or we're talking 50 years, like that is still just like a drop in the hat, right? Like it's, it's such a short period of time where it's like almost inevitable. I think many people in this space will say that AI within 50 years is just like going to really, really fundamentally shape or change our world. And so again, like coming back to his point, it was like if it's in 5 months, 5 years, 50 years, in the spectrum or the span of human civilization it's, it's nothing, right? And, and it, it doesn't seem like nothing within our lifetimes, but, um, yeah, I mean, coming back to the idea of- It helping people, it hurting other people. How do you think about, like, what creators or, or people should be focused on? Should they be focused on, like, learning prompt engineering? Should they be focused on a field that they think won't be touched with it, by AI, if that exists? Like, w- what are you doing as a creator to say, like, "You know what? I'm, I don't wanna be left behind, so I'm gonna be learning X, Y, and Z"?
- KCKaren X Cheng
Um, well, I don't think the answer is focus on prompt engineering, because I wouldn't be surprised if, um, in a very short amount of time, AI gets extremely good at prompt engineering, and so those people are less, you know, relevant. Um, in fact, I was actually just talking to someone who was using GPT-3 to write prompts to run through Stable Diffusion. So, um, I would say that the best thing to do is not to learn a specific skill, but to... 'Cause that's gonna change, um, probably by the time this podcast comes out, you know? Um, I think the best thing is to adopt a mindset that you have to always be learning, and to accept that the model that humans had for much of humanity, which is that you could kinda choose a career and then have that career for life, that's gone. We are at the, like, the last dying, grasping breaths of those days. I think the sooner you can accept and, and embrace the mindset that, like, the world is always changing, the skills you have to learn always need to change, the better off you'll be.
- SSSteph Smith
I don't know if we've lived in a time that's been so exponential where these technologies are changing, but humans have adapted, so we'll see how we adapt. I mean, some, some interesting examples you pointed me towards in, um, an article from Nathan Baschez, where he talked about even, like, simple things like the introduction of Unsplash completely changed the, like, design architecture or th- or almost, like, the design. How companies stand out with their design changed, because all of a sudden, stock photography, like hyperrealistic pictures, became widely available. And so then you saw, like, what I think people call Memphis design come up, and you saw all these, like, very similar looking illustrations. And so I do wonder if there's some version of that, right? Where, like, we have access to unlimited 2D art, so maybe 3D art is important, just as an example, but then of course AI will eat that. So I don't know. I, I think it's interesting to kind of think through how humans might evolve,
- 57:50 – 1:03:17
The adaptiveness of humans
- SSSteph Smith
or do you think that this stuff is just moving so quickly that-
- KCKaren X Cheng
Um, I think humans are remarkably adaptable. Um, and so we will adapt. I do think that, um, you know, already the world is moving much faster than human evolution can keep up with. So for example, like, we weren't really evolved to sit in, like, office chairs all day. We weren't evolved to hear the comments of thousands of people on social media. We were evolved for, like, 30-person tribes, right? And so you see all of these, like, mental health and physical health problems that we have. A lot of those are because human evolution hasn't caught up to how much culture has changed. And so I, I say two things. One, humans are remarkably adaptable, but two, there is a limit to how much humans can actually evolve from, like, you know, a Charles Darwinian evolutionary perspective, uh, to keep up with some of this stuff. So we are entering a completely unknown time. I'm just, I'm trying to stay positive [laughs] and try to highlight as much of the positivity as possible, but, um, yeah, there is a lot of uncertainty.
- SSSteph Smith
It is such a sea change and, uh, to your point, like, even as two people who believe fundamentally that technology is good, or at least I don't wanna put words in your mouth, that's something I believe, it is sometimes hard to just, like, fathom what the hell we do and who's gonna be disrupted and h- a- and, and it's impossible to say that there aren't negative externalities to something like this.
- KCKaren X Cheng
Yeah. I, I would say I think technology is neither good nor bad. I think I, I like to view it as neutral. It's like the neutral canvas. And so I find that to be the most empowering, because if it's neutral, then we have the ability to influence it-
- SSSteph Smith
I like that
- KCKaren X Cheng
... in which direction.
- SSSteph Smith
And I've heard other people use, like, it's like a mirror, right? It, it, like, it mirrors back to humans-
- KCKaren X Cheng
Yeah
- SSSteph Smith
... ultimately certain things that maybe are below the surface sometimes.
- KCKaren X Cheng
Yeah.
- SSSteph Smith
Um-
- KCKaren X Cheng
That's good. I like that.
- SSSteph Smith
But taking a little bit more of an optimistic lens, I've seen AI used in a ton of different ways. Like, I, I saw something recently which was, like, an AI that helps write Excel formulas, which I thought was, like, quirky and neat. Um, I also saw one that, like, takes legal jargon and simplifies it for the end user.
- KCKaren X Cheng
Oh, interesting.
- SSSteph Smith
Um, so there's tons of, like-
- KCKaren X Cheng
Did it work well?
- SSSteph Smith
I tested it only a little bit. It did, at least the prompts I played around with did, but I didn't play around for very long. Um, curious to know from you, do you have, like, a wish list? Are there things that AI has not touched yet where you're like, "Man, I would really love this"? And while you think about it, I'll just quickly share one. Super simple, but we're doing these podcast recordings with video, and every so often, I would just love to re-say something, right? And things like Descript allow you to re- re-say something with audio very easily. But I want that with video, because if we're airing the video versions of these, I wanna just be able to be like, remove my filler words and, you know, fix up my performance. But obviously it'll look ridiculous if the video portion ... isn't equivalent. So that's my little wish list for anyone at OpenAI or otherwise to, to do. I know they're probably coming out with video at some point, but to sync the two.
- KCKaren X Cheng
What I have wanted for AI is video tools for creators that are AI-powered. That's why I've been doing so much hacking, trying to get, like, these image tools to work for video, 'cause it's not quite there yet, but I keep on, like, trying to make it work. Once it does though, like, I don't even know what's gonna happen. [laughs] So I'm almost like, be careful what you wish for, you know?
- SSSteph Smith
Yeah. I mean, it's coming, right?
- KCKaren X Cheng
Right. It's-
- SSSteph Smith
It-
- KCKaren X Cheng
It's for sure coming. Um, it's hard for me to say a wish list, I guess, 'cause I'm just trying to keep up with what's already coming out.
- SSSteph Smith
Oh, totally. Something we talked about before we jumped on, which is, like, a neat little idea, um, just to get the wheels spinning for any listeners, is we were talking about audio AI.
- KCKaren X Cheng
Mm-hmm.
- SSSteph Smith
Have you actually seen, um, any AI ASMR? I wonder if that's a thing.
- KCKaren X Cheng
Oh, no. I've never heard of that.
- SSSteph Smith
No, no, I haven't either, but I was just-
- KCKaren X Cheng
You know what would be interesting with ASMR is, like, if they could, like, uh, connect to your senses and measure, like, your response to it.
- SSSteph Smith
Oh, yeah.
- KCKaren X Cheng
And then, like, in real time, like, adjust-
- 1:03:17 – 1:12:10
0 The importance of AI ethics
- KCKaren X Cheng
My wish list is actually for all humans to take the ethics of it very seriously. Um, and what I mean by that is, like, my wish list is that everyone working on AI be held to a standard to be using it for positive forces rather than negative, and that we somehow develop an insane amount of cultural pressure that anyone who's not using for positive s- sources is, like, not ostracized from society, but, but penalized in a way that it is severely dissentivized.
- SSSteph Smith
How do you think about what is good and bad though in that context?
- KCKaren X Cheng
Okay. So I mean, it's very hard to be the arbiter of that, but for example, like, deception is bad, right? So when you're altering things, you need to disclose what you altered and how.
- SSSteph Smith
Do you think that we will need to somehow label things as AI-generated or human-generated? Because, you know, a simple example, uh, you brought up deception, but people do these deepfakes, right, of, of people without their consent. They're not quite there yet, but they're very close to being-
- KCKaren X Cheng
Right. Well, have you seen the Tom Cruise ones? They are actually fully-
- SSSteph Smith
Like, you can't tell the difference
- KCKaren X Cheng
... the Tom Cruise deepfakes. I, I, I can't tell. I, I think most people can't tell.
- SSSteph Smith
Okay.
- KCKaren X Cheng
Um-
- SSSteph Smith
So, so we're already there. It's not even a future, uh-
- KCKaren X Cheng
We're already there on edge cases, and we'll pretty soon be there for-
- SSSteph Smith
Yeah
- KCKaren X Cheng
... mass cases. I think that will be developed. I think, I think, um, it will be necessary, and it will happen, and I think it'll be very similar to, like, nutrition facts on, like, all the food. Like, it's a universal standard that we have nutrition facts. And so there will need to be some sort of system that, like, shows, you know, this is a video. This was produced in this way.
- SSSteph Smith
I've heard people talk about this, and it sounds like, I might be wrong, 'cause I've just, you know, just listening to another podcast, but it likely would have to be at the hardware level, as in run by, like, Apple or Google, et cetera, because, um, if it's not at the device level, that's, like, really easy to hack or to, to abuse, right? And so it literally would have to be like, okay, this AI is running on this machine, and this machine is plugged into Stable Diffusion or whatever, insert AI here, um, and the output in the metadata, like, and it, it would have to be, like, a, a case where you could not remove it, has that information.
- KCKaren X Cheng
It would have to be the hardware level, and I think it will be first adapted in, like, really high-stakes scenarios where the veracity of something is extremely important. So for example, in, like, political messaging, maybe they'll have, like, a very specific kind of camera or hardware where it cannot be altered, and then that is, you know. So for certain types of, like, really important messaging where it's really important that it not be faked, whereas, like, you know, the, like, someone's social media influencer video, like, it's, doesn't matter if something was, like, adjusted or tweaked, you know?
- SSSteph Smith
So there's, like, standards, or that's part of the regulation, that in certain fields you must use certain hardware, or at least you must use, like, a suite of hardware that, that plugs into these?
- KCKaren X Cheng
I mean, regulation will catch up, but I actually think it will be somewhat self-regulated in the sense that some people will need the des- will have the need to show that what they do is true and not altered. And then there will be companies that pop up that say, "Okay, this is the technology that does that," and then there will be people who then buy that technology and use it of their own free will and accord, not because of regulatory requirements.
- SSSteph Smith
Mm-hmm.
- KCKaren X Cheng
Uh, 'cause it'll take a while for the regulation to catch up.
- SSSteph Smith
Yeah, but I do wonder i- in that case, like, yes, the good actors will use the technology that has that functionality, but then bad actors likely won't, right? Unless there's regulation.
- KCKaren X Cheng
That's, that's why I'm, that's why I say that, like, somehow I want, my wish list is for the culture to change. Because humans are a self-enforcing society. You know, we can- Are very naturally tribal, and we behave well largely because that's what's expected of us from other humans. It's not like someone's watching us at every single moment, but we still in general, you know, wanna behave well. And a lot of that is the cues that you're getting from culture. And so I would love to see a culture developed where, hey, as, as humans, as part of being a good human, um, we use technology responsibly and ethically. I don't know how that's gonna happen. [laughs] But that's my, that's why it's my wishlist.
- SSSteph Smith
I love going down these rabbit holes, 'cause it does just get the wheels spinning on like, man, there is, there's so much to figure out. And when you're dealing with an exponential technology, policy, regulation, IP, a lot of the stuff takes time, right? And so it... Like, I also wonder whether we create regulation or policy, let's say, in 2023, and then by 2025, like, it's, it's already just like, you know, it, it's completely different.
- KCKaren X Cheng
I actually think culture can change a lot faster than regulation. Much, much, much faster. And I'll give you an example. The fastest I've ever seen culture change was the handshake during COVID. You know? How many, like, h- how many thousands of years or hundreds of years has the handshake been the golden standard, and then all of a sudden, like, socially unacceptable. So many things became socially unacceptable almost overnight because of COVID. And I was like, wow, human culture can change like that. And then now that the pandemic is better, then, like, we're back to doing it again. So, um, I do think that, uh, we're gonna need a lot of cultural forces, um, to kind of sprint ahead of regulation, 'cause regulation will take time.
- SSSteph Smith
Yeah. Well, I have no idea how this will progress. It's been fascinating to hear how you think it might progress. As someone who has played around with these tools, and we'll be sharing a lot of your projects as we talk about them throughout this episode. Um, but I wanna end with something really fun, which is that you've worked with tons of creators. Uh, creators that people know, creators that people don't know. Um, do you have a, a wishlist, or is there one creator that you would just love to partner with? I love throwing this stuff out into the ether, 'cause you never know. You never know who might make this happen.
- KCKaren X Cheng
You know, it's interesting, 'cause in the past I would've had, like, a dream client or a dream, like, something. But I've actually learned over time that it's not about, like, which celebrity you can get or which brand you can get, but it's, for me, I've learned it's actually about doing my best ideas, my best work. And so the way I do things is I try to match-make, like, my ideas to a client who will accept them unaltered, essentially, or with the least alterations possible. So I don't, I actually don't, won't answer that question, or don't even ha- actually, nothing actually comes to mind, because it's, it's mostly about, like, uh, making the ideas that I wanna make.
- SSSteph Smith
Yeah. Yeah, I love that. It used to be, or at least for me, when I was younger, it used to be like, I wanna work with these people, or I want these companies to hire me. And I think you're right. There is, like, an evolution, like we talked about before, where you're almost, like, transcending attention, and you've gotten to a place where actually your focus is, how do I make something really great?
- KCKaren X Cheng
Yes. It's almost like, it's, like, almost like the Maslow's hierarchy of needs, and it's like, oh, I'm actually, I'm, I'm actually just, I'm actually focused on, like, doing the ideas that I wanna do. Whereas in the past, like I know 10 years ago, I would've been like, I wanna dance in a Coca-Cola commercial. Like, I, I'm not, I'm good. I don't need to do that now. And then, like, even two years ago, I was like, I really, really, really, really wanna make a commercial for the Apple iPhone. And now I'm like, I, I'm good. Like, I, I don't actually need to do that anymore. So, um, uh, so yeah. I, I don't have, I don't have one, 'cause it's actually about the idea, and it actually is so cheesy. It sounds so cheesy, but it actually is about enjoying making the idea.
- SSSteph Smith
Well, that's a wonderful place to end things off. So Karen, thank you so much, uh, for talking with us today. This was so fun to just think through, you know, the optimistic, the pessimistic, the dystopian, utopian futures that we might live in. But at the end of the day, I think one of the biggest takeaways for me is that AI is a tool, and that tool is being democratized. So that in itself, I think is exciting for many people to be able to participate in that future.
- KCKaren X Cheng
Yeah. Thank you so much. Thank you for having me, and thank you for these great questions. [upbeat music]
- SSSteph Smith
Thanks for listening to the a16z podcast. If you liked this episode, don't forget to subscribe here on YouTube to get our exclusive video content. We'll see you next time. [upbeat music]
Episode duration: 1:12:10
In the uListen app, the transcript lights up word by word as you listen. Get uListen
Transcript of episode 45kg85J3MPs
