AI and the Creator Economy with Karen X Cheng

a16z2022-11-09

12K views|1 years ago

💫 Short Summary

Karen Chang, a digital creator, explores the impact of AI technology on creativity, emphasizing the evolving relationship between humans and AI. She reflects on her journey from startup failure to successful content creation, highlighting the transformative power of overcoming challenges. The video discusses the shift in social media algorithms and the importance of building a follower base over going viral. It delves into the use of AI tools in content creation, showcasing various techniques and capabilities. The discussion also touches on AI's impact on art creation, intellectual property challenges, and the evolving landscape of AI integration in creative industries.

✨ Highlights

📊 Transcript

✦

Karen Chang discusses her use of AI technology in her creations, including viral works like an AI-generated Cosmo magazine cover and a Dolly fashion show.

02:30

She explores the impact of AI tools on creativity and the evolving relationship between humans and AI.

The interview covers how AI can enhance the creative process and potential implications for creators in the future.

The discussion highlights the beginning of the AI era and the opportunities it presents for innovation and collaboration.

✦

Karen's journey from a failed startup to a successful content creator.

04:13

Despite being rejected by YC, she learned valuable lessons about resilience and adapting.

Her first startup failure was a pivotal turning point in her career.

Emphasis on the importance of learning from failures and using those lessons for future success.

Karen's story showcases the transformative power of overcoming challenges and leveraging setbacks for growth.

✦

Importance of Building a Following on Social Media.

07:17

The speaker initially focused on creating viral videos instead of growing their following.

Relying on reporters to spread videos is no longer effective due to changes in social media algorithms.

Building a following is now crucial for success in the current social media landscape.

Changes in algorithms have made it more competitive to go viral on social media platforms.

✦

Building an established follower base is key to growing a following on social media.

08:25

Sharing behind the scenes content was the key to viral success for one creator.

Social media has shifted from a social graph to an interest graph, changing how creators interact with their audience.

The social media landscape is more competitive, with creators recognizing the downsides and the impact of algorithms on content performance.

✦

Impact of algorithms on content creation.

10:59

Concerns raised about designing for algorithms rather than human taste, leading to uncertainty and potential overwhelm for creators.

Use of AI in content creation discussed, with speaker sharing successful partnership with AI for social media content.

Speaker's personal journey of feeling creatively stuck and exploring AI advancements in research papers to enhance their work.

✦

Creating social media friendly videos through experimentation with research from white papers.

13:33

The excitement of being at the forefront of using new technology creatively, such as AI tools, which are currently in a 'free-for-all' state.

Admitting to not being technically inclined, the speaker relies on talented programmers to assist with technical aspects.

Utilizing tools like Dolly and other image generation tools to enhance video creations.

✦

Tools for video editing such as Dane, Nerf, and EB synth are discussed in the segment.

16:33

Dane is used for creating custom stop motion videos, while Nerf is used for 3D scanning scenes.

EB synth was utilized to make Dolly work for videos, resulting in the creation of an AI fashion show.

These tools offer unique techniques and capabilities, including handling light changes and creating smooth movements.

The segment emphasizes the evolving nature of video editing tools and their potential for creative applications.

✦

Highlighting different AI tools like Dolly, stable diffusion, and mid-journey and their capabilities and shortcomings in generating images.

19:26

Addressing misconceptions about AI as solely replacing humans, emphasizing AI as tools for specific purposes.

Mentioning the potential of combining tools like facetune to enhance image quality for better results.

Contrasting AI researchers striving for perfection with creators using tools like facetune to showcase different approaches to achieving desired outcomes.

✦

Potential specialization of AI tools for specific use cases such as hyper-realistic human photography or generative illustrations.

21:51

Predictions of companies acquiring others to expand verticals, like Meta's acquisitions of WhatsApp and Instagram.

Creation of an AI-generated magazine cover for Cosmo magazine, highlighting the importance of human input.

Details on iterative rounds of development and the creative direction chosen to feature a woman astronaut due to restrictions on human faces in AI.

✦

The process of creating the final image involved multiple iterations and adjustments.

24:52

Swagger was added to the woman's walk and influencer tips were used for dramatic effect.

Extensive manual editing was required, including fixing the face and extending the boots.

The technique of 'in painting' was utilized to refine specific areas of the image.

The video showcases the detailed work of enhancing visual content through a combination of AI and human intervention.

✦

The process of creating digital art involves multiple iterations and prompts, with some artists spending hours refining their work.

27:41

The pricing model for digital art creation tools has evolved over time.

The key to achieving good results lies in having a good artistic eye, patience, motivation, and the ability to describe the vision clearly.

Traditional art skills are not always necessary, as many individuals can produce impressive work with the right tools and mindset.

There are examples of successful digital art expansions, such as enlarging famous paintings in Dali.

✦

Using AI to create art by expanding famous paintings like Mona Lisa and Girl with a Pearl Earring.

29:34

Hiring an expert to enhance the Pearl Earring painting and being moved by the results.

AI democratizes art creation, allowing non-artists to produce art.

Questioning whether AI will elevate many to success or only the very best artists.

Exploring whether AI will change the criteria for artistic talent and if outliers will still emerge as exceptionally skilled.

✦

Impact of AI on Art Creation.

31:40

AI art is making art more accessible by allowing more people to explore their creativity.

The technology is reshaping the skill ecosystem and may replace traditional skills like oil painting.

Similar to Uber disrupting the taxi industry, AI is changing the landscape of art creation.

Despite the impact on traditional artists, creativity and innovation remain important in art creation.

✦

Importance of portraying AI positively in media.

35:29

Emphasis on potential collaboration between humans and AI.

Responsibility to show optimistic side of AI and move away from sensationalized content.

Acknowledgment of legitimate concerns about AI's negative impacts.

Stress on the need for a balanced and nuanced portrayal of AI in the media.

✦

Shifting from Negative to Positive Content Creation.

37:47

The speaker discusses the trend of focusing on negative content for more clicks and views.

They share a personal turning point where they shifted to creating positive and wholesome content instead of clickbait.

Using AI for good purposes is emphasized, rather than controversial or negative topics.

The importance of inspiring and enduring creations over transient, attention-grabbing pieces is highlighted.

✦

Shifting focus from seeking attention to earning respect in content creation.

39:51

Importance of building relationships with a smaller, respected following rather than chasing shallow attention.

Converting attention into business opportunities for long-term success.

Drawbacks of relying solely on viral content.

Emphasizing authenticity and depth in content creation for sustained success.

✦

Importance of Quality Content Creation on Social Media Platforms.

44:01

Creators express frustration over low-quality content gaining millions of views on TikTok while their high-quality work receives less attention.

Low-effort content creators may not receive the same opportunities or recognition from major brands as those who produce high-quality content.

Examples are given of creators with fewer followers making substantial income from their work, emphasizing quality over quantity.

The discussion emphasizes the importance of focusing on creating valuable content rather than chasing large follower counts for success.

✦

The importance of metrics in measuring content creator success.

45:46

Social media has created a culture where validation is constantly sought through likes and views.

Intellectual property in the age of AI is a complex and significant topic of discussion.

Ownership of IP in AI-generated content raises questions about value and presents challenges regarding training data and creators of AI tools.

Balancing audience preferences and creative integrity is crucial for content creators.

✦

Highlights on the Value of Selling Prompts

49:16

Uncertainty exists around intellectual property rights and creator payment in the prompt-selling economy.

Creators must be compensated fairly as the ecosystem evolves.

Advocacy for setting standards in paying humans through AI tools is crucial.

Debate on payment distribution between top talent and democratized fields with fluctuating price dynamics.

✦

AI in creative fields like photography is currently seen as a gimmick and headline grabber, with companies using it more for innovation than practicality.

51:48

By 2023, AI is predicted to be integrated into workflows as it becomes more accepted and useful.

A reporter faced backlash for using AI-generated art in an article instead of hiring a human artist.

Despite the context of the situation, the reporter was criticized for his choice, highlighting the ongoing debate and challenges surrounding AI in creative industries.

✦

Impact of AI on Society.

53:43

Initial backlash against AI will decrease as acceptance grows over time.

Discussions about AI can be uncomfortable as it has both positive and negative impacts on people.

AI is seen as a transformative technology that will reshape the world in the next 50 years.

Creators should embrace and adapt to AI advancements rather than trying to avoid them, as prompt engineering may not offer long-term protection.

✦

The importance of adaptability in a rapidly changing world.

56:46

Emphasis on continuous learning and evolution to keep up with technology advancements like AI and stock photography.

Reflection on human adaptability and the challenges posed by the fast-paced evolution of society, leading to mental and physical health issues.

Questioning how humans will continue to evolve in response to technological advancements and societal changes.

✦

The evolution of culture surpasses human evolution, emphasizing adaptability and limitations.

58:56

Uncertainty and the importance of staying positive amidst technological advancements are highlighted.

Technology is viewed as neutral, empowering individuals to influence its direction.

Examples of AI applications are shared, including simplifying legal jargon and aiding in writing Excel formulas.

The speaker expresses a wish for AI to develop video tools for enhancing podcast recordings, enabling easy editing of audio and video content.

✦

Potential of AI-powered creators and advancements in audio AI.

01:03:14

AI ASMR and personalized content based on user prompts are discussed.

Importance of ethical considerations in AI development is emphasized.

Suggestions for labeling AI-generated content to distinguish from human-generated are made.

Calls for cultural pressure to incentivize ethical AI practices are made.

✦

Development of AI technology for detecting deepfakes in videos and ensuring authenticity.

01:04:55

Emphasis on hardware-level implementation for security purposes.

Potential adoption in high-stakes scenarios like political messaging.

Predictions on self-regulation within industries for verifying content authenticity through hardware standards.

Anticipation of companies providing technology solutions to guarantee unaltered content.

✦

Importance of culture in influencing human behavior and responsible use of technology.

01:07:20

Concern about pace of regulation compared to cultural changes, especially evident during the pandemic.

Emphasis on cultural forces leading ahead of regulation to effectively address technological advancements.

Significance of cultural evolution in shaping societal behavior and the need for responsible technology usage.

✦

Evolution of idea creation focus.

01:11:31

Emphasizes importance of quality work and aligning ideas with receptive clients.

Shift from pursuing celebrity clients to prioritizing personal fulfillment in idea creation.

Reflection on AI democratization as an exciting tool for broader participation in the future.

Optimistic view on AI democratization potential for positive impact.

00:00hey everyone welcome to the latest

00:01episode of the a16c podcast this is your

00:04host Steph Smith in today's episode I am

00:06thrilled to bring on Karen Chang Karen

00:09has been a creator for a long time but

00:11she specifically has a knack for

00:12integrating technology into her

00:14creations and you can only imagine with

00:17the rise of some of these AI tools like

00:18Dolly mid-journey and stable diffusion

00:21what she's been able to achieve now if

00:24you missed AI summer and its impact on

00:26creators you might have been sleeping

00:27under Rock but you're not too late it's

00:29truly just the beginning in fact I

00:31interviewed Karen back in September and

00:33with every week there was something that

00:35popped up where I thought man I wish I

00:37could have asked her about this and yet

00:39still in this interview we cover a lot

00:41of ground we'll chat with Karen about

00:43how she originally found her Niche how

00:45she manages to grow her social following

00:47with a nod towards being additive

00:49instead of reductive you'll also get to

00:51hear the behind the scenes of her many

00:52viral Works including a video of her

00:54becoming a lawnmower yes truly a

00:57lawnmower her AI generated Cosmo

00:59magazine cover the the first ever her

01:01dolly fashion show her transforming

01:04iconic art into 3D museums that people

01:06can explore and much much more and by

01:09the end of this episode I think the

01:11audience Will Come Away with some really

01:12key things including an understanding of

01:15the new tools that are available that

01:17are literally at our fingertips how AI

01:19can indeed enhance the creative process

01:21and maybe even some of the second order

01:23effects of these Innovations like how

01:25creators are paid or how they might work

01:27with these AIS and I should note that

01:29since we're just at the beginning of

01:31this AI ERA this is truly just the

01:33beginning of our coverage so I hope you

01:35enjoyed this episode the content here is

01:38for informational purposes only should

01:40not be taken as legal business tax or

01:42investment advice or be used to evaluate

01:44any investment or security and is not

01:46directed at any investors or potential

01:48investors in any a16z fund for more

01:51details BC a16z.com disclosures

01:55foreign

01:59[Music]

02:04I am absolutely thrilled today to have

02:07Karen Chang on the line Karen is a

02:09digital Creator she has taken social by

02:12storm seriously she's all over my feeds

02:14I can almost guarantee you've seen one

02:15of her videos and she's done everything

02:17from high five Mark Zuckerberg in the

02:19metaverse she's partnered with several

02:21A-list celebrities she also was

02:24commissioned by Cosmo recently to create

02:26their first AI generated magazine cover

02:28so super cool we'll be talking about

02:30that and in particular we'll be talking

02:32about her foray into AI how she's using

02:35AI to be more creative and how that

02:37relationship between humans in AI is

02:39evolving how it enhances our creativity

02:42or

02:43how it doesn't we'll see but Karen

02:45welcome to the show thank you so much

02:47for having me thank you for that amazing

02:49intro I'm talking to you Karen in 2022

02:51you have millions of followers across

02:53various social media platforms

02:55but I want to flash back to Karen from

02:582013. so I'm going to preface this with

03:01I was going down this YouTube rabbit

03:03hole I was not researching for this

03:05episode I happened to stumble upon a

03:08video with Paul Graham and Sam Altman

03:10from 2013 YC office hours

03:14you know where this is going I can't

03:16believe you found that

03:17so that's kind of obscure yeah so I'm

03:21watching this video Someone Else

03:23familiar pops into it Ryan Peterson from

03:25flexport is in this video and I'm like

03:27okay flashback and then someone else is

03:30in this video and it's you and you're

03:31pitching this company give it 100 which

03:33is basically your startup at the time

03:35and the startup was structured around

03:38people sharing their journey in learning

03:40skills sharing their progress every

03:42single day for 100 days

03:44you also then I think around a year

03:46later you did a TED Talk where you kind

03:49of use that framework of 100 days to go

03:51and learn dance and you shared that and

03:53it was Fascinating People can look that

03:55up we'll have it in the show notes okay

03:57so this is a broad question but how do

03:59you go from that right 10 years ago

04:01almost uh were you building the startup

04:03to where you are today when did you was

04:06there a point in time where you decided

04:07I want to start sharing this information

04:09online I want to start being a Creator

04:10okay so yeah this was back in my like

04:13failed startup days and so I like really

04:18badly very badly wanted to get into YC

04:21as like everyone did and

04:25um managed to get on stage office hours

04:27spoiler alert did not get into YC was

04:30turned down I think multiple times

04:33um was interviewed on stage

04:35um and answered their questions that

04:39startup eventually failed like a year

04:41later

04:43um I think it was really tough on me at

04:44the time because up until then I had

04:46like you know good academic achievements

04:48in school and it was my first like I

04:50think your first startup failure is like

04:51your first definitive failure it's like

04:55yeah there's no it's like you fails

04:58there's no like like it's over exactly

05:01like right and so I sort of had to like

05:03come to terms with that

05:05but

05:06um honestly like

05:08going through that startup

05:11I learned so much about video that I

05:13then applied to like doing my own stuff

05:16and actually I got this really

05:18interesting feedback when I was uh

05:20trying to like get business for my

05:22startup and I was talking to someone I

05:24was trying to like get them to invest in

05:26my company and they were like we're not

05:29interested in this company like they

05:31they could see that it wasn't a scalable

05:33business but they were like but we

05:35actually just wanted hire you to make

05:37some videos and I was like I don't want

05:39to do that like but then actually it

05:41turned out you know the years later I'm

05:42like yeah that person was actually right

05:43and

05:45um I I think it just took it took me

05:47trying every single career

05:50to realize like I TR I tried working for

05:52a huge company I've tried working at a

05:54startup I tried being a founder of my

05:56own startup I tried running an agency I

05:59tried like I've tried everything and

06:00I've I've come to the realization that

06:02for me actually like being an influencer

06:04is the best thing for me which is crazy

06:06because I I always swore I would never

06:08be an influencer

06:10well I know exactly what you mean

06:12because I feel like

06:13um I also create things on the side and

06:15for the longest time people would

06:16associate me with certain things they'd

06:18be like Steph you're like the writing

06:19girl and I was like no please please

06:20please please I'm not the writing girl

06:22that is not how I picture myself but

06:24ultimately sometimes you just have to

06:26lean into what other people see in you

06:29and so you started creating these videos

06:31you've been doing it for years I think

06:33you originally uploaded your first

06:35musically in like 2017 and even in 2019

06:39I think I saw you say in a video you had

06:41around 10 000 followers on Instagram so

06:43that might seem like a lot to some

06:45people but compared to today where

06:47you're just smashing it you have over a

06:48million followers everything you create

06:50almost goes viral

06:52um was there a specific unlock where you

06:54found along that Journey that like you

06:56noticed something that really worked or

06:58you you honed in on something that

06:59people really wanted yes absolutely

07:02there was so um for a very long time I

07:06struggled to grow my following

07:08um first I resisted first I was like

07:10well I don't even need to grow my

07:12following so I was just making like

07:13viral one-off videos

07:15um that was back in the days when you

07:17can make things go viral in a completely

07:18different way that doesn't work anymore

07:20so back in like

07:222012 up until around 2016 you could make

07:26videos go viral just by coming up with

07:28something clever that would be a good

07:29headline cold emailing like 300

07:31reporters getting them to write about it

07:33and then it would almost guaranteed go

07:34viral

07:36um that does not work anymore because

07:38Facebook eight like all the reporters

07:41jobs so

07:43um and now videos don't spread the

07:45reporters they spread through algorithms

07:46and so

07:48it is actually so so competitive now

07:51it's so much more competitive now than

07:53it was like 10 years ago uh it is it is

07:56much harder to make a viral a video just

07:57go viral and so the way to do it now is

08:00you kind of have to build a following

08:02and so it's not it's not so much about

08:03making everything go viral making things

08:05go viral but it's more about just like

08:07building an established follower base so

08:09that your Baseline of like the number of

08:11people who see your videos or your work

08:12just steadily gets higher and I really

08:14struggle to grow my following

08:17um I remember when I first tried growing

08:19my Instagram following I was like why is

08:20this so hard you know I would get things

08:22to go viral and it would be very hard to

08:24go to um to get any followers out of it

08:26uh the unlock for me was actually

08:29sharing my behind the scenes

08:31and that was it and so I would like

08:33before I would just show these cool

08:35shots and I'd be like look at this cool

08:37shot you know and then maybe I would

08:38describe how I did the caption I'm like

08:40nothing crickets

08:43if I showed the same shot the exact same

08:45shot but show the behind the scenes with

08:47it

08:48all of a sudden it was going viral

08:50and I've done that over I've done that

08:52so many times now that I am confident

08:55that that was the like the defining

08:57thing there are lots of other things

08:59that I learned along the way to grow my

09:00followers but if there was one thing it

09:01would be that

09:03that's really interesting because a lot

09:06of people participate in these online

09:08worlds a lot of people are creators and

09:10a lot of the stuff is very similar so

09:12sometimes all it takes is like one

09:14iteration one like adjustment in what

09:17you're doing you're still sharing the

09:19same kinds of videos but then just

09:21giving that behind the scenes look is

09:23completely changing the game for you I

09:25also want to just quickly get your take

09:26on something you mentioned which is that

09:29social media is not the same right like

09:31we use the same term from 10 years ago

09:33but even if you think about

09:35social as like a word social media isn't

09:38not really social anymore in a way it's

09:40kind of gone from the social graph where

09:42you followed your friends and your

09:43family and then now it's an interest

09:45graph where you follow just things that

09:47this algorithm is serving you have there

09:49been any other changes that you've

09:52noticed in terms of just like bird's eye

09:54view how social has changed that really

09:57matter to creators like is there a way

09:58that we're a lot of creators are

10:00thinking about this art incorrectly I

10:03think it's gotten a lot more competitive

10:05I think people are a lot more aware of

10:08the downsides of social media now so

10:10we're no longer in like the honeymoon

10:11phase of it for sure you know it's kind

10:15of interesting that like for most of

10:17History

10:18I feel like humans have prayed to God

10:21you know but now I see that like content

10:25creators we pray to the algorithm

10:27it's almost like a Superstition it's

10:29like

10:30what does the algorithm think my video

10:32is not doing well because of the

10:34algorithm it's like how can I optimize

10:36my video for the algorithm and it's like

10:37we're no longer it's like we're almost

10:39no longer designing ourselves for like

10:41human taste but we're designing

10:43ourselves for like the algorithm

10:45and I do think there are quite a few

10:48downsides to doing that

10:51um and there are good reasons to resist

10:54that as well yeah yeah I think you're

10:57right about that and and I think

10:59increasingly those algorithms are

11:01becoming black boxes where

11:03um you think you know one variable is

11:05going to make your videos or your work

11:07go viral and then you play into that but

11:10you don't really know right like you're

11:11not there seeing the the different

11:13levers that control the algorithm and so

11:16it really can consume you yeah the black

11:19box is actually what it makes it feel

11:20like a cult or a religion because it's

11:22like you don't know it's like a Mystical

11:24Force yes exactly and even some people

11:26who work at these companies say that

11:28they don't know exactly what plays or

11:31not so much what plays into the

11:32algorithm but the output of it one of

11:34the ways that you've stood out is at

11:37least recently playing around with AI

11:39right I I think you're one of the few

11:41craters that I've seen that has really

11:43mastered the art of partnership with AI

11:46as it relates to Social and there will

11:48be many more of course right but um how

11:51did you start getting into that what

11:53what led you towards those kind of

11:55Creations

11:56um you know I was really just feeling

11:58quite stuck

12:00in my work I felt like I was doing

12:04basically like almost all one million of

12:06my followers are from the same video I

12:09mean it's a few videos but it's the same

12:10kind of video

12:12um I guess which one it is yeah is it is

12:16it the phone Matrix one

12:18yeah it's like the low budget camera

12:20tricks so uh there was one video

12:22actually where I attached like a phone

12:24to a ceiling fan and then made that into

12:26like a matrix bullet time effect and I

12:28got 300 000 followers from that one

12:31video so at every time I got a surgeon

12:34followers it was the same thing it was

12:35like I would I would attach a phone to

12:36some household object find like some

12:38like hack to do a camera trick with it

12:42and um people are following me for that

12:45and I was just like I'm like out of a

12:47household objects I'm out of ideas

12:50um and I just felt like I was I just

12:51felt like I was running out of ideas and

12:53I was like I have to find new like tools

12:57to new toys to play with

13:00and in the meantime I was I was starting

13:02to see some of like the really insane

13:05AI stuff coming out in white papers but

13:08it was all like two minute papers like

13:11research videos of researchers like

13:14shooting whatever they had in their

13:15bedroom and of course they're so focused

13:17on the research it's not their job to

13:19think about the storytelling or the

13:21cinematics of it and I was like huh

13:23that's interesting I don't think anyone

13:25is actually doing that yet so I could

13:28try doing that and so I started

13:30experimenting with like well let's take

13:32the research from white papers and make

13:33it a social media friendly video you

13:36know

13:37um and I started doing that at the

13:38beginning of this year

13:40and it was just so fun it was it was I

13:43think I am

13:44um I'm like a very novelty seeking

13:47personality and like my greatest

13:49motivation is like can I try to do

13:51something that hasn't quite been done

13:52before and like yeah sure like probably

13:54someone has done it but can I find my

13:56own like original Twist on it and so the

14:00fact that like these white papers no one

14:02was really like using this technology no

14:04one was really using in creative ways so

14:06I could be like wow I really am like the

14:09first to do something with this that was

14:11very exciting for me whereas

14:13trying to come up with like the latest

14:15iPhone transition

14:17all the ideas are taken it is very very

14:20hard to think of something clever or new

14:22with an iPhone camera trick tick tock

14:24transition very very very hard but with

14:27the AI stuff it's like this is just

14:30free-for-all right now yeah yeah and and

14:33I know that a lot of these tools are

14:36becoming more democratized more

14:38available but there was I think a little

14:40bit of a technical Gap too right in

14:41order for you to actually like start

14:43playing with these AIS relative to the

14:45average Creator is that right yes there

14:47is and I'm not that technical

14:51um I took like a coding class once in

14:53college but I prefer not to get into

14:56that myself so I actually work with

14:58really talented programmers who help me

15:01run it or they'll teach me to set it up

15:04um and so that's why I think like when

15:06Dolly came out I was like oh this is

15:09requiring no technical ability like this

15:12is if you can do a Google image search

15:14you can use Dolly you know and so that

15:16was really interesting for me yeah and I

15:19want to talk about those like second

15:20third order effects of having that fully

15:23democratized but you mentioned Dolly

15:25I've also seen you used a ton of other

15:28tools so I'm curious to hear from you

15:31like what are some of the other tools

15:32that maybe others haven't heard of or

15:35that you're playing around with or maybe

15:36like dolly is the first AI tool that you

15:39use but I think you also use other ones

15:40to like clean up a video or to like

15:43iterate on the creation from Dolly so

15:45can you share maybe some of the tools

15:47that you're playing around with yes uh

15:50so there are a few with the image

15:52Generations ones which are really

15:53popular right now there's Dolly there's

15:55mid-journey there's stable diffusion

15:58um there's disco diffusion I think if

16:00you're actually watching the video

16:01version of this podcast by the way we'll

16:02have like visuals yes we'll bring all of

16:04them up yeah

16:06um and so those are really popular right

16:09now I also think there are other ones

16:12that are

16:14um they require more technical knowledge

16:15but they are they enable kind of more

16:19um

16:20techniques that are that are less

16:22mainstream and so therefore more unique

16:25if you can pull them off so one of them

16:27is Dane

16:28um Dane is really simple all it does is

16:30it applies artificial slow motion to

16:32your video

16:33um but I thought it would be used in

16:35really interesting creative ways by like

16:36actually

16:37shooting like a custom stop motion video

16:39for the purpose of it being interpolated

16:44um to create these impossible movements

16:46and you'll just I can't describe I think

16:49I saw one of yours it was like you as a

16:51lawnmower or something yeah like lying

16:54on the ground and then you basically get

16:56dragged across and that's is that using

16:57Dane right exactly so um if you're just

17:00listening to this imagine like a

17:01stop-motion of someone like on the

17:03ground like uh planking basically and

17:05then just like slithering across the

17:06ground in stop motion right but then

17:08with with Dane you can make it perfectly

17:11smooth and so it looks like I'm

17:12literally being dragged On The Ground by

17:13like an invisible string or something

17:16um but actually that was just a

17:17stop-motion video that we used AI to

17:19smooth out and so I did that with James

17:21Pearlman so he helped me out with like

17:23smoothing everything out um other tools

17:25are Nerf so Nerf has gotten pretty

17:28popular Nerf is a way to use any camera

17:31like your phone to scan a scene and then

17:33all of a sudden you have this beautiful

17:373D like

17:39scan of it and what's interesting about

17:42it is it's different than photogrammetry

17:44because it can handle like light it's

17:46basically constructing like a particle

17:49like field and so when you look at

17:51things from different angles the light

17:52changes and so that's why it can handle

17:54mirrors whereas like traditional

17:56photogrammetry can't then there are

17:58other tools too that we've played with

18:01um I recently discovered EB synth which

18:03I actually learned is not an AI tool

18:05um but I use that to basically figure

18:08out how to make Dolly work for

18:13um video so Dolly's photo only but using

18:15EB synth I was able to make Dolly work

18:17for video very cool and and we'll talk

18:20about this later but you ended up doing

18:21like an AI fashion show right yes with

18:24the combination of many of these tools I

18:26want to hear from you though a lot of

18:28these tools are quite nascent we're

18:30going to see a lot more of them come to

18:31Market how do you see them all playing

18:33together because let's just use like the

18:35text to image generators so you have

18:37Dolly you have stable diffusion you have

18:39mid-journey there will be many more

18:41a lot of them

18:43um

18:43are similar but also what I'm imagining

18:46if we like extrapolate many years like

18:48how will they differentiate how will you

18:50as a Creator or how will consumers

18:52decide I'm going to use dolly or I'm

18:54going to use stable diffusion will they

18:56be different business models or how do

18:58you think that's going to evolve

19:01um I can speak from the Creator side

19:02maybe more from the business side but

19:05um I think right now ai

19:08in the in the in the mass media there's

19:12sort of this misconception or perception

19:14that

19:17AI is just this like all-powerful thing

19:20that's going to replace humans

19:24um but I think a more interesting way to

19:26think of it is that they are they are

19:29tools and each one is a specific tool

19:32for a specific purpose just like I would

19:34have like a calculator or a compass or

19:37something else

19:39so and you can combine and like what

19:41where you get interesting things is

19:42where you combine the tools right so for

19:44example

19:46um you can generate an image in Dali and

19:50then you can use cap cut which is a

19:53program which is like an app on your

19:54phone that will literally turn any image

19:56into like a 3D image then there's

19:59another problem which is uh like Dolly

20:01and mid-journey that and all I think all

20:03of these image synthesizers have which

20:05is that they make faces they don't do

20:07human faces very well there will be like

20:09artifacts on the eyes they'll just look

20:11kind of Droopy or a little messed up

20:14well guess how you can fix them you can

20:16run them through facetune the app that

20:18all influencers use and that'll make

20:21that'll make that'll fix it and so I

20:24actually facetune can fix it to that

20:26degree because I've seen some of the

20:27outputs from Dolly as an example and

20:29they're really kind of it'll fix them to

20:32be it'll fix them to be Flawless wow

20:34okay I mean it depends on the image but

20:36like I have seen I have seen it do it

20:39flawlessly okay

20:42it's like you know you could you could

20:43be the AI researcher who's like

20:45working really really hard you know it

20:48it gets harder and harder to reach

20:50Perfection right so you could be the AI

20:52researcher trying like really really

20:53hard to get that last one percent or

20:55like a Creator could just be like I'm

20:56just gonna run this through facetune

20:58yeah yeah

21:00I also wonder whether some of these

21:02tools become more specialized because

21:03these text to image generators are quite

21:05broad right they're being trained on

21:07like all of the internet and they're

21:08being used for use cases of like

21:11literally again all of the internet and

21:13because there's so there's going to be

21:14so many of them I do wonder whether they

21:16become specialized like maybe some of

21:18them become specialized for just like

21:20hyper realistic human photography and

21:24other ones become really specialized for

21:25really like generative

21:28um illustrations or something like that

21:29do you think that might happen or do you

21:31think they'll continue to stay really

21:33broad you know different companies will

21:36have different strengths and then

21:38they'll continue to be able to

21:39differentiate by sort of specializing in

21:41that but I do imagine that some of these

21:43companies are gonna buy other companies

21:44and then so they'll like you know kind

21:47of have more verticals kind of like how

21:49you know meta bought like WhatsApp and

21:51Instagram I see that happening with AI

21:54tools too right yeah so like if we use

21:57meta as an example they basically try to

21:59own messaging and so you might see like

22:01a similar phenomena with AI tools let's

22:05jump into a really cool example of how

22:08you deployed AI which is something I

22:10mentioned before you got commissioned by

22:12Cosmo magazine to create their first AI

22:15generated magazine cover we'll throw it

22:17up on the screen it looks awesome and

22:20you've also done as you talked about a

22:22behind the scenes explainer but for

22:23those who are not familiar with this

22:26project of yours can you share a little

22:27bit more about how it went down uh yes

22:30so um this was kind of shortly after

22:32dolly was released the people open AI

22:35gave me a call and they were like hey we

22:36have this really cool potential

22:38opportunity where Cosmo wants to make

22:40the first ever AI generated magazine

22:43cover and they were like and we want you

22:45to generate it

22:47that's crazy

22:49um and

22:53so I mean my first reaction was like

22:56uh worried I guess like I hope I can do

22:59a good job like

23:02um

23:04because keep in mind like it's not

23:07I can't emphasize enough how much like

23:09the human is needed in this process you

23:12know like yeah they could have just

23:15you can't actually have a fully AI

23:16generated magazine cover because you

23:18have to give a prompt

23:20and I guess you could theoretically ask

23:23like an AI to generate the prompt but

23:25still the human needs like a human needs

23:27to give some input right and so for for

23:30our process actually we went through so

23:33many different rounds of of iterations

23:36and

23:38um eventually we land on the creative

23:40direction of a woman astronaut and this

23:42was actually a I thought a really clever

23:43work around by openai to not show a

23:46human face because at that time human

23:48faces were not allowed in dolly in the

23:50early days human faces were not allowed

23:51for Dolly and so we and we also were

23:54like well what race is this one going to

23:56be too like we didn't want to make a

23:58race statement and so by making it a

24:00woman astronaut it could be any

24:01ethnicity as well

24:04um and so so they had that introduction

24:06and then they showed me a reference

24:08image that that was generated in Dolly

24:10and I was like

24:11like this is so good

24:13I don't know that I can even top this

24:15image and I was just like

24:18I I don't know that I can make a better

24:20image than this so I spent hours trying

24:23to make a better image than that and I

24:25could not and I was like like they hired

24:28me to do this

24:30I obviously have to do a better job than

24:32like what they quickly threw together

24:34so I just kept going for like a few more

24:36hours and then eventually I like

24:40I was like okay I I kept on

24:42troubleshooting and I was like okay I

24:43want this woman to like walk with

24:45swagger and it took me I kept on saying

24:47like walking confidently walking proudly

24:50like walking energetically and none of

24:52those problems working out and then

24:52eventually I was like walk with swagger

24:54and then boom

24:56then I got like the then I got the hip

24:58like thrust so like it was like

25:02you know it was like I I kept on asking

25:03for walking confidently and that didn't

25:05work but then when I asked for a walk

25:06with swagger it indeed gave me Swagger

25:08and then the other thing I was thinking

25:10of like was

25:11um there's like an influencer tip that's

25:13popular on social media for taking

25:15photos and it's like the hack is that

25:16you you bring the camera to the ground

25:18and shoot from the ground up to make it

25:20look really dramatic and I was like oh

25:21let me try that so then I was like how

25:23about wide angle shot from below and

25:26then boom it made it way more dramatic

25:28and then the image that I got actually

25:30its face was sort of messed up so then I

25:32did in painting and stuff to like fix

25:34the face and and then I did in painting

25:37to like extend the boots so I did I did

25:39quite a bit of like uh fixing up of the

25:41original image it was if you look at the

25:43original image it was it was quite a bit

25:44of a diamond in the rough

25:46um so if you look at my behind the

25:47scenes video there was a a significant

25:50amount of human work to make it happen

25:52yeah and we'll share that when you say

25:54in painting just for people who aren't

25:56familiar you're literally you're still

25:57using AI for that right but you're

25:59basically selecting a portion of the out

26:01put an image you're saying okay I like a

26:04lot of this but I don't like this area

26:05so like can you can you in paint can you

26:08replace that specific region in painting

26:10is like magic Photoshop like if

26:12Photoshop worked magically that's what

26:13it would be so imagine there's like a

26:15photo and let's say you don't like what

26:18you're wearing or you don't like the

26:19person's background or you don't like

26:20anything about it you erase that you

26:22type in what you want and then boom you

26:25get three options

26:26which is so incredible and that's what

26:28you also use partially for your fashion

26:31show so again we'll talk about that

26:33later but I want to get a sense of you

26:35talked about the fact that humans need

26:37to be involved and that it took several

26:39iterations like how many iterations are

26:41we talking in this in this spectrum or

26:44this this route to the final image it

26:46will depend a lot from artists to

26:48artists you know I think some people

26:49they'll share their first prompt and

26:52that's um that's interesting some people

26:54will spend you know maybe five to ten

26:57they'll just type 5-10 prompts and then

26:59for some people where they're really

27:01trying to achieve a very specific result

27:03for a specific project

27:05I mean we'll be in there for like hours

27:07and hours

27:09um running like dozens or hundreds of

27:11prompts and in the early days of Dolly

27:13it was unlimited and free so it was a

27:16little easier to do back then yeah and I

27:18think it's I don't know if they I think

27:20they may have changed their business

27:21model recently or their pricing but I

27:23saw an article

27:24um that this this woman was basically

27:26creating like a llama dunking a

27:28basketball

27:29um so we'll pull that up too but I think

27:31that one took a hundred iterations and I

27:33think it was like around 13

27:36um with that pricing so it gives you an

27:38idea and as you said it'll depend but I

27:41think the the maybe more interesting

27:42question is how fungible was Karen in

27:46that process and I don't want to yes how

27:48replaceable am I but if you gave Average

27:51Joe the same project and the same tool

27:55do you think they would come up with

27:57anything comparable

27:59I think a lot of people would have been

28:02able to make a really good Cosmo cover

28:05um I think those people would have been

28:09the people who would have done it best

28:10are people who have a good artistic eye

28:13have the kind of the patience and the

28:16motivation to keep going and keep

28:18refining on the prompt and who can like

28:22describe what they see in their head

28:25and so a lot of these people are not

28:27necessarily traditional artists you know

28:28I think a lot of people could have made

28:29a really really good result I there's

28:32actually an example that I have that we

28:35can show up that we can show on screen I

28:38a while ago wanted to expand The Girl

28:41With a Pearl Earring so you may have

28:43seen one thing that you can do that was

28:45actually just announced is you can

28:47expand paintings in Dali

28:51um and so you can basically like take a

28:53famous painting and then imagine what

28:54was all around it so you can take like

28:56the Mona Lisa and imagine like where was

28:57she you can take the girl the Pearl

28:59Earring imagine like everything that was

29:01around her

29:02and so I made a video a while back where

29:04I imagined her like in a like library

29:07and she was holding a book and I was

29:09like I want her to be like an educated

29:11woman you know

29:12and so I made that

29:14um and then I actually

29:16um am working with a project uh with

29:18openai actually right now it's one that

29:20they commissioned uh it will it's not

29:22out yet but it will probably be out by

29:24the time this podcast releases so if you

29:27go on my Instagram uh on Karen next Chen

29:29you'll be able to go and see this filter

29:31we made this Instagram filter where we

29:34expanded these famous paintings

29:37and

29:38um you can actually go inside through

29:40them and see these famous paintings

29:42and I looked at my girl the Pearl

29:44Earring and I was like

29:46this can be done better and so I hired

29:49August Camp who is

29:52um the person who taught me about this

29:53method and she is she has spent way more

29:57time in Dolly than I have she's so

30:00artistic and talented and then when she

30:03showed me I hired her to do this Pearl

30:05Earring and when she showed me hers

30:08I was just like

30:10moved to tears almost like

30:13jaw drop I was like

30:15that is what you made and so I felt like

30:19she was irreplaceable in this I mean

30:22I'm sure I could have hired 10 different

30:24artists to do this Pearl Earring and

30:26gotten 10 very different results and

30:29what I like about this example is that

30:31everyone is starting with the same

30:32Source image and so it almost

30:34establishes like a little bit of like a

30:36control you know for it and so

30:40um I think this is the perfect case

30:42study for like hey different humans get

30:44very different results with AI yeah it's

30:47it's a tool as you said but the reason I

30:49asked if you were fungible is because I

30:51just I really wonder how this progresses

30:53in terms of we know it's going to be

30:55somewhat of a democratizing force

30:57because now a bunch of people who like

30:58couldn't paint or couldn't do Photoshop

31:00in certain ways like can now do it and

31:03create all of this art or or or these

31:05outputs but I do Wonder then does that

31:09make a lot of people really successful

31:12or a lot of people really capable or

31:14does it still surface the very very best

31:17to the top right where where there's a

31:20different filter and now the filter is

31:21prompt engineering versus painting or

31:24drawing or singing or whatever it might

31:25have been in the past so do you think

31:27that's still the case where you're still

31:28going to see these like outliers who are

31:32just so much better than the rest or do

31:33you think it'll be more of a Level

31:35Playing Field than what we're seeing

31:36before okay so I think that

31:39um

31:40what this is gonna do what AI art is

31:43gonna do is it's going to

31:45significantly lower the barrier to entry

31:47to become an artist

31:48so be to be an artist right now you have

31:50to have a lot of time

31:53um a lot of training

31:56um sometimes the monetary Financial need

31:58means to be able to do that or the

32:00willingness to be like a starving artist

32:02to do it or it's a hobby on the side but

32:05there are kind of some there's

32:07definitely a barrier to being an artist

32:08and and now it's like everyone can do it

32:12and it's almost like I almost liken

32:15these image synthesizers like dolly or

32:17mid journey to like a Peter Pan but

32:19instead of stealing from the rich and

32:21giving to the poor it takes the artistic

32:25skill of artists and it's like gives it

32:28to everyone like here you go you know

32:30and so I do think that a lot more people

32:32are going to be empowered to be artistic

32:34or to be artists because they didn't

32:38necessarily have the patience to go like

32:40learn oil painting but now

32:42but they actually do have the talent to

32:44be able to describe what's in their

32:45imagination and continue to refine until

32:47they get the result they want

32:49so a lot of people are going to be

32:50empowered to be artists there will still

32:52be the standouts and the extraordinary

32:54people because they're going to be the

32:55ones who are finding like different or

32:59creative or innovative ways to do it

33:00there's you know as much Innovation as

33:02there is there's always going to be the

33:04ability to combine things in new ways or

33:07you know find the latest so I I

33:10definitely think that um there's going

33:12to be a lot more artists

33:16um a lot of unhappy artists

33:18um a lot of happy artists and also still

33:22standout artists yeah yeah I mean it's

33:26it's almost just reshaping the like

33:27skill ecosystem and like some people who

33:30have invested in skills for many decades

33:32like if they were just like an A plus

33:35oil painting unfortunately for them

33:38that skill is somewhat being replaced by

33:40this technology

33:42um and it reminds me a little bit of you

33:44know if we use like uber as an example

33:45like some of those taxi drivers had

33:47invested in their taxi license for many

33:50many years and so there's kind of a

33:52parallel there doesn't mean it's bad or

33:54good but it is just a fact that these

33:57AIS are like as you said almost like

33:58cherry picking skills and like dropping

34:01them on the rest of the population and

34:03it's this like great experiment to see

34:04what comes out of that

34:06um I also kind of want to hear from you

34:08as you worked with these AIS was there

34:10any like aha moments or like did you

34:13feel like it was truly like a back and

34:15forth process or did you feel like it

34:17was unidirectional and let me explain

34:18what I mean by that if it was

34:20unidirectional you're basically like

34:22giving the AI prompt you're seeing what

34:23comes back and you're like no no that's

34:25not what I want you have a vision but if

34:26it's more bi-directional did you get a

34:28sense of like you gave it a prompt you

34:30got something back and you go whoa I I

34:32didn't even think of that but like I

34:35like that I like so so was it more

34:37unidirectional or was it more

34:38bi-directional it's 100 of the second

34:40one

34:41um it does feel like a collaboration and

34:43so you get you oftentimes get things you

34:46don't expect

34:48and you're like oh

34:49I didn't think of that let me go down

34:52that rabbit hole now more

34:54so I mean that that reminds me of

34:56something I want to discuss with you

34:58which we've actually talked about before

35:00in a call we had which is the perception

35:04of AI so there's almost like two

35:07framings and I know this is a really

35:08simplistic but on one side there's like

35:10human versus AI or like human versus

35:12machine it's like an age-old Trope and

35:15then there's another which is like human

35:16plus AI right which is more optimistic

35:19it's like whoa like what can we do with

35:21this partnership and you've told me that

35:23you feel almost like a responsibility to

35:29portray this technology these tools in a

35:32certain way maybe even you know I don't

35:33want to put words in your mouth but more

35:35of this optimistic versus pessimistic

35:37doomsday lens so can you just speak to

35:40like why you feel that responsibility or

35:41almost like how you think

35:44um AI should be portrayed or how you

35:46want to portray it I've had to unlearn a

35:49lot of my bad habits in terms of like

35:54I am a trained viral video creator and

35:58so in many ways I'm trained very similar

36:01to the way a reporter would be which

36:03means that I'm rewarded for making click

36:05bait headlines

36:06and so when all this AI stuff came out I

36:09knew that like my my first instinct was

36:12like oh let me make a bunch of like

36:14human versus AI like let's let's have

36:16the AI do something and a human do

36:17something and let's have like judge's

36:19critique and say which one's better and

36:20like that was my first instinct to do

36:23that

36:25and then I realized like well what is

36:27that just gonna do that's just gonna

36:31it's just gonna freak people out you

36:33know

36:35um and I do think you know people are

36:37there there are reasons to be worried

36:40for a about AI I think there are very

36:42many legitimate ones and it will

36:45negatively impact some people more than

36:47others

36:49um but AI is a technology it can be used

36:53for good it can be used for bad and I

36:56just I think it clicked that like

36:59there's gonna be

37:01every gravitational force for the media

37:04is going to push towards the bad because

37:07that is what's going to get more clicks

37:09and engagements and Views and what's

37:12what's not the natural gravitational

37:14force is to actually think of ways to

37:16use AI for good

37:18and so I was also sort of in this

37:21turning point in my career where I just

37:22spent like two years optimizing for

37:24growing followers and I finally just

37:25sort of let go of that and I was like

37:26what would what would it look like if I

37:28stopped trying to go my grow my

37:29followers what would it look like if I

37:30actively stopped trying and just made

37:33like what I wanted to make I was like

37:35well since I no longer care about

37:37growing my followers which was which was

37:38very freeing for me I was like well I

37:40actually don't have to do this clickbait

37:42thing I actually don't have to try to

37:43get the most views by doing this

37:45controversial AI stuff I actually can

37:47just make like

37:49wholesome positive find creative uses of

37:52it because if it is a technology that

37:54can be used for good or for bad like why

37:57not you know it's more rewarding for me

37:59to make it for good

38:01even if it's less viral oh well it

38:05doesn't matter have you found that

38:06things have been less viral when you go

38:08that more optimistic route

38:10um I think on a

38:13one-off basis each thing is less viral I

38:16I there are there are many pieces where

38:18I was like if I had written the headline

38:20in a more controversial way it would

38:22have gotten bigger it would have gone

38:23bigger

38:24but I definitely think long term

38:28it's a better strategy for me to

38:32consistently put out positive stuff so

38:34people can people associate that you

38:36know with the content I'm putting out

38:39yeah because I I was just curious I mean

38:40you mentioned news as an example

38:43pessimism cells right click bait cells

38:46it captures your attention fear is like

38:49a very natural human emotion

38:51um but I do wonder if to your point like

38:54more optimistic Creations endure right

38:58so like people follow you because

38:59they're like Karen is this inspiring

39:02person she's learned all these

39:03Technologies she's creative so they view

39:06you as truly like this

39:08this creator that they more so like want

39:11to be like instead of I think pessimism

39:14sometimes is quite surface level right

39:16it does capture your attention but

39:17you're not like inching to go back to

39:19like hear from that Creator right it's

39:21very like transient yes I think that's a

39:24great Insight it's like you can make

39:26something go viral but it's a one-off

39:27and

39:29people aren't necessarily going to be

39:30inspired to continue to hear from you

39:33or you can continue to put out positive

39:36content in each piece of positive

39:37content isn't going to get as many views

39:39but you will kind of form a more

39:42long-term relationship uh in people's

39:45minds and there's actually this really

39:46incredible quote that I saw a while ago

39:48and it actually like changed a lot of

39:51how I think about things as a content

39:52creator uh in doing social media and I

39:55don't remember who said it but the quote

39:57is

39:58seek respect not attention it lasts

40:02longer

40:03and I was like wow

40:05because I was on this hamster wheel of

40:08chasing likes and Views and as anyone

40:11who has made a viral video knows it it

40:14it's over so quick you can have like a

40:17million views one day and then like it's

40:19it completely dies down by the next and

40:22so people get on this hamster wheel of

40:24addiction of like continuing to churn

40:26out content to try to get viral videos

40:28and I was on that hamster wheel for

40:30years and just being like when does this

40:31end like why am I doing this why am I

40:33Contin you know and it's it doesn't feel

40:35good it's addicting and then I saw that

40:38quote and I was like oh

40:40if I actually just put

40:43more effort into making more memorable

40:45good content I don't have to put out

40:48content as often

40:49and yeah I would much rather have the

40:52respect of a few people who I respect

40:56rather than the attention of the shallow

40:59attention of millions who forget about

41:01it by the next day

41:03I mean even on that point

41:05if if we even

41:07remove the like feel good I feel

41:11respected Etc part of it it's also good

41:13business and and what I mean by that is

41:15like

41:16a lot of people view attention as like

41:19the end goal like I want a million

41:21followers I want this uh this number of

41:24people to pay attention to me but

41:26ultimately like

41:28you need to convert that attention if

41:30you want this to be like a long-term

41:31game whether it's like to find clients

41:33or to like sell a course or to like

41:36start a rolling fund right there's many

41:37like versions of turning attention into

41:40business uh or Commerce but that second

41:44part of conversion does not work if

41:46you're just doing these like viral

41:48things and I I operate more on Twitter

41:50but as an example of this like um

41:52something I've been pretty outspoken

41:53about is just like these tweet threads

41:55which are just like so overdone and like

41:57so templated so um they work but they're

42:01they're just really really surface level

42:03and

42:04um

42:04ultimately what I find is that I see a

42:06lot of creators like feed into that and

42:08they're like okay I'm gonna go all in on

42:10this thing that I don't believe in I

42:11don't feel good about Etc but it's gonna

42:14bring me hundreds of thousands of

42:15followers but then again coming back to

42:17the point it's like well try converting

42:20those followers into something

42:23you know that that generates you money

42:24or that that is long lasting and and

42:26it's very hard right because those

42:28followers don't actually love your

42:30content they don't buy into you they

42:31just happen to be like oh okay another

42:33thread like it's almost like a habit for

42:36them to engage with that kind of work I

42:38am so glad you said that because I I

42:41feel the same way and ultimately this is

42:44my job

42:45the millions of followers are useless to

42:48me if I can't make a penny off of it

42:50like I'm not gonna put in all this

42:51effort it's my job right I need to make

42:53money out of it and I remember that I

42:56used to feel really bitter when I would

43:00see like these Tick Tock videos blow up

43:03and get millions and millions of views

43:04and they would be the most inane things

43:07like for example I'll give you an

43:08example of a tick tock video that can go

43:10really viral

43:11um it's a video where someone's like try

43:13to get the number of hearts and the

43:15number of comments to be the same

43:17right and so everyone's just like it

43:20gets like tens of millions of views

43:22because it's like a very smart

43:24engagement bait right it's a very clever

43:25engagement fate and I would be mad at

43:27that or I would or you know there's

43:29there's other examples that are less

43:31clever there's there's just straight up

43:33garbage on Tick Tock that is getting

43:35millions and millions of views and that

43:37is so demotivating

43:39for creators VFX creators artists who

43:42are putting like hundreds of hours into

43:45their craft to make something look

43:47really beautiful and then it completely

43:48flops and doesn't get very many views on

43:50social media and I would I used to be so

43:54bitter about that and then like someone

43:56told me and then I learned over time

43:58that this was actually very much true is

44:01that the the people who are putting out

44:02low effort content all the time getting

44:04millions of views they're not the ones

44:07getting hired by like major respected

44:11Brands you know they're not actually

44:14getting great opportunities off of their

44:16follower account

44:17um whereas the ones who are I know

44:21people who make incredible content they

44:24only have like 20 000 followers

44:26they're making bank

44:28you know yeah exactly no I mean I just

44:32have to say I resonate with what you

44:34said so much there as a Creator myself

44:36again I operate more with Twitter but

44:37same thing I'll see these threads and

44:39I'm like oh my God like what is going on

44:42why are people liking and retweeting

44:43this and why is this one account blown

44:45up it's it's really low quality compared

44:47to some other accounts that I follow

44:49where I'm like this is gold this is gold

44:52why is no one paying attention

44:54um but yes I think I think it's not

44:56productive to focus on that because

44:58ultimately the people who are putting

45:00the best stuff out there are the ones

45:03who if we take the whole value chain

45:05right like the whole process and

45:08consider that it's a business those are

45:10the successful quote-unquote businesses

45:12people put so much value on the number

45:16of likes and the number of views because

45:18it is the only number we can really see

45:20and judge ourselves by and so it's like

45:21we've been trained since like

45:25kindergarten to put value in like the

45:28number judgment you know when you get

45:30like a 97 on a test wow that's so good

45:32if you get like a 60 that's a d that's

45:35bad and so we are we are so trained to

45:37take that feedback of numbers as the

45:40truth and then so when it comes social

45:43media comes along and you get lots of

45:44views versus a little it's so easy to

45:46say well a video was great because I got

45:48tons of views and it's not because it

45:49got very little views and it took me

45:52years to learn that that is it is such a

45:54bad metric

45:56but it's the only like a objective one

46:00we have I know I know but it's hard

46:03sometimes as we talked about it's hard

46:05to just like to check out of that

46:07ecosystem to be like Oh I'm actually

46:10going to focus on the things that I want

46:11to put out there the things that I'm

46:12proud of instead of feeding into as you

46:15said what some people might argue is an

46:18objective metric of like this is what

46:19people want so you also have to accept

46:21that you have to go against so many

46:23natural human instincts

46:25yes

46:26um I want to talk about another topic

46:28which I've seen

46:30everywhere it's coming up in basically

46:31every conversation I have around

46:33creators around Ai and it's the idea of

46:36Ip right so we talked about like being

46:38hired for things but also like

46:41um owning your Creations is something

46:42that's mattered a lot for not just

46:45independent creators but if you think

46:46about like all the way up films

46:49entertainment like what matters is who

46:51owns the IP and AI introduces a very

46:55very interesting

46:57filter on this and and honestly I don't

47:00know who if anyone has the answers but

47:02but what I want to walk through is

47:04where is the value

47:08um along the AI creation process so like

47:10there's the there's the information that

47:14the AI is trained on which you could

47:16argue the output wouldn't exist without

47:18that so maybe that's where the IP is

47:20there's also the argument that the

47:22creators of the AI tools right open AI

47:26mid-journey Etc while the outputs also

47:28wouldn't exist if the AI tool didn't

47:30exist and then of course there's the

47:33prompt which you could also argue if

47:35someone doesn't enter a very very

47:37specific prompt the output also does not

47:40exist and I've even seen some a company

47:43yesterday that was selling prompts

47:45because it is valuable yeah to have the

47:48right I promise I've bought a prompt

47:50before there you go so there's there's

47:52probably going to be a whole economy

47:53around this so I want to hear from you

47:55obviously there's maybe no right or

47:57wrong answer but how do you think about

47:59that like where is the IP and how on

48:02Earth will we think about that as

48:04creators moving forward yeah I I I don't

48:07know

48:08um I actually I've thought about this

48:10but I don't I don't have the answers I'm

48:12not a lawyer I know that also like

48:15lawyers don't even have the answers like

48:17Congress doesn't have an answer to this

48:18yet it's also new

48:21um and I think it does

48:24yeah I don't know I think

48:27um well let me rephrase the question to

48:30something that I think you can comment

48:32on which is how creators get paid how

48:34creators charge so something that we

48:36talked about

48:37um before was that for your projects

48:40whether it was with Cosmo or we were

48:42actually talking about a project you

48:44said that

48:45um you feel or you felt like it was

48:48really important that those projects

48:50were paid for right and that might sound

48:52obvious but like as we talked about

48:54before AI

48:55in certain cases might not cost that

48:57much it might cost like two dollars to

48:59to run the props you're you're looking

49:01for

49:02um but some of those projects

49:04historically like I don't know how much

49:06Cosmo would typically pay for a magazine

49:08cover but it's not two dollars right so

49:10how do you think about the way that you

49:12charge as a Creator in this new

49:14ecosystem

49:16yeah I think it you know one of the

49:19things that when I work with clients

49:21I'll literally say them like hey like

49:24you know if even if you don't hire me

49:26um and sometimes it doesn't work out and

49:28so I'm like hey you know I'm you know my

49:29schedule doesn't work out for this

49:30project but whoever you hire like please

49:32pay them please allocate this budget

49:35because

49:37um

49:38we are at like a

49:42we're at like a moment in history that's

49:44very vulnerable where like certain

49:46standards can be established and I would

49:48like to see the standard established

49:50that if you hire a human to use an AI as

49:53a tool then you pay the human

49:57do you think that holds up though and I

49:59ask that because I think

50:01I think like many Industries the very

50:04best of the best If there really is like

50:06a one percent or like a 10x prompt

50:08engineer they will get paid very well

50:10we've seen that happen in many fields

50:12but then you also see when things get

50:14democratized like it's very hard for

50:16those price economics to not change

50:19right because because you imagine for

50:21example that at the beginning maybe yeah

50:24maybe certain creators are you know

50:26holding their ground they're saying we

50:28must be paid X but then you're gonna

50:29have insert Creator here that is like

50:32actually I'll take half of that or maybe

50:33I'll take a hundredth of that price I

50:35think that's a really good point I think

50:37there is

50:38um incredible downward pressure

50:41um and so what's probably going to end

50:42up happening unfortunately is that there

50:44will be like the more well-known people

50:46who can still charge higher rates but

50:49then

50:50for the vast majority of people like

50:52there's such insane downwards pressure

50:55on it and I do worry about that

50:59um for creators and I don't really know

51:01how that is all gonna play out yeah it's

51:04really hard to predict but

51:06yeah I wonder how some of these fields

51:09change I mean I want to hear from you

51:10because you've started to work with some

51:12of these companies

51:13at this point in 2022

51:16is AI or at least the tools that people

51:19are using are they seen as more gimmicks

51:21or or do you see companies like really

51:23investing in this and saying oh actually

51:25we're gonna like we are gonna replace

51:27our photographer with AI or you know you

51:30to use Cosmo as an example

51:32was working with you just something

51:33where they could like put their flag in

51:35the ground and say hey like we're

51:36Innovative we're we're working with AI

51:38or do you see them actually going and

51:40like doing every fifth cover with AI how

51:43are companies thinking about this in

51:452022 it is most certainly a gimmick and

51:48a headline Grabber it is like let's be

51:50Innovative let's be first let's be on

51:53top of this I would imagine in 2023 the

51:56answer to that will be probably be

51:58different

51:59um people will actually start using it

52:02as their workflow

52:04um because it makes sense and not as a

52:06headline grabber really interesting I

52:09saw that there was someone and a

52:11reporter who used a mid-journey

52:14art for his article I think about Alex

52:17Jones and then he went viral in a bad

52:21way in that he got like majorly attacked

52:24on Twitter for using mid-journey art

52:27for a magazine article

52:32um and he was attacked because he didn't

52:34hire like a human artist instead you

52:37know like um it was I think the Atlantic

52:39uh correct me if I'm wrong

52:41um but I think what what people didn't

52:45have context was was that the department

52:47he was in the image that he would have

52:50used instead would have been like a

52:51Getty Images or like some stock photo it

52:53wouldn't have been like a official

52:55commission right so that was what he was

52:58replacing but he still got attacked for

53:01it

53:03so I think you know the first few you

53:06know the leaders go the arrows attack

53:08the leaders right so I think the first

53:11few people who are going to start using

53:13AI are going to get backlash but it's

53:15going to get less and less and less and

53:17less until it's accepted I honestly

53:20discussing some of these downsides is

53:21very uncomfortable for me because I

53:23don't know

53:24what's going to happen and I do think

53:27that

53:29um

53:30it could be bad for a lot of people and

53:33I personally try to do try to stay

53:35optimistic and try to see the optimistic

53:36side but like there's no denying that

53:39like it is going to be bad for a lot of

53:41people and it is going to be good for a

53:43lot of people

53:44it is a game-changing technology I've

53:47heard some people say that it is like

53:49the technology of Our Generation in

53:51terms of how it shapes things I also

53:54heard um Sam Altman say which I thought

53:56was like an interesting frame

53:57um a lot of people will talk about like

53:58and we even did this at the beginning of

54:00our call they'll be like when when is

54:02this gonna like change things

54:04fundamentally

54:06um he didn't give at least in the

54:07interview I saw a definitive timeline

54:10because I think no one can pretty much

54:13no no no but we were just having fun but

54:16I think his framing was really

54:17interesting which is that he believes

54:19that uh again it's it's like the

54:22game-changing Innovation and if you look

54:25at humans and like the spectrum of time

54:27that we've been on this Earth if we're

54:30talking five years or we're talking 50

54:33years like that is still just like a

54:35drop in the Hat right like it's it's

54:37such a short period of time where it's

54:40like almost inevitable I think many

54:41people in this space will say that AI

54:43within 50 years is just like going to

54:46really really fundamentally shape or

54:48change our world and so again like

54:50coming back to his point it was like if

54:52it's in five months five years 50 years

54:55in the Spectrum or the span of human

54:57civilization

54:59it's nothing right and it doesn't seem

55:01like nothing within our lifetimes but

55:04um yeah I mean coming back to the idea

55:06of

55:07it helping people it hurting other

55:10people how do you think about like

55:13what creators or or people should be

55:17focused on should they be focused on

55:19like learning prompt engineering should

55:21they be focused on a field that they

55:23think won't be touched with it by AI if

55:25that exists like what are you doing as a

55:28Creator to say like you know what I'm I

55:30don't want to be left behind so I'm

55:32gonna be learning x y and z

55:34um well I don't think the answer is

55:36focus on prompt engineering because I

55:37wouldn't be surprised if

55:40um in a very short amount of time AI

55:42gets extremely good at prompt

55:43engineering and so those people are less

55:45you know relevant

55:47um in fact I was actually just talking

55:49to someone who was using gpt3 to write

55:52prompts to run through stable diffusion

55:53so

55:55um

55:57I would say that the best thing to do

56:01is not to learn a specific skill but to

56:03because that's going to change

56:06um probably by the time this podcast

56:07comes out you know

56:09um I think the best thing is to adopt a

56:12mindset that

56:15you have to always be learning and to

56:19accept

56:20that the model that humans had for much

56:23of humanity which is that you could

56:26kind of choose a career and then have

56:27that career for life

56:30that's

56:31gone

56:32we are at the like the last dying

56:35grasping breaths of those days

56:39I think the sooner you can accept and

56:42embrace the mindset that like the world

56:46is always changing

56:47the skills you have to learn always need

56:50to change

56:51the better off you'll be I don't know if

56:54we've lived in a time that's been so

56:56exponential where these Technologies are

56:58changing but humans have adapted so

57:01we'll see how we adapt I mean some some

57:03interesting examples you pointed me

57:05towards an um an article from Nathan

57:07Bashas where he talked about even like

57:10simple things like the introduction of

57:12unsplash completely changed the like

57:14design architecture that or almost like

57:16the design how companies stand out with

57:19their design changed because all of a

57:21sudden stock photography like hyper

57:23realistic pictures became widely

57:24available and so then you saw like what

57:27I think people call Memphis design come

57:29up and you saw all these like very

57:30similar looking illustrations and so I

57:32do wonder if there's some version of

57:35that right where like we have access to

57:37unlimited 2D art so maybe 3D art

57:40is important just as an example but then

57:43of course AI will eat that so I don't

57:45know I I think it's interesting to kind

57:47of think through how humans might evolve

57:50or do you think that this stuff is just

57:51moving so quickly that

57:53um I think humans are remarkably

57:55adaptable

57:57um and so we will adapt

58:00I do think that

58:03um

58:04you know already the world is moving

58:08much faster than human evolution can

58:10keep up with

58:11so for example like we weren't really

58:13evolved to sit in like office chairs all

58:16day we weren't evolved to

58:20hear the comments of thousands of people

58:22on social media we were evolved for like

58:2430 tribes right and so you see all of

58:26these like

58:27Mental Health

58:29and physical health problems that we

58:31have a lot of those are because human

58:33evolution hasn't caught up

58:36to how much culture has changed and so

58:39I I say two things one humans are

58:42remarkably adaptable but two there is a

58:45limit to how much humans can actually

58:48evolve from like you know a Charles

58:50darwinian evolutionary perspective

58:53uh to keep up with some of this stuff so

58:56we are entering a completely unknown

58:58time

58:59I'm just I'm trying to stay positive

59:02and try to highlight as much of the

59:05positivity as possible

59:07but um yeah there is a lot of

59:10uncertainty it is such a sea change and

59:14to your point like even as two people

59:17who believe fundamentally that

59:18technology is good or at least I don't

59:20want to put words in your mouth that's

59:21something I believe it is sometimes hard

59:23to just like fathom what the hell we do

59:25and who's going to be disrupted and how

59:28and and it's impossible to say that

59:31there aren't negative externalities yeah

59:34I would say I think technology is

59:36neither good nor bad I think I I like to

59:39view it as neutral it's like the neutral

59:42canvas and so I find that to be the most

59:45empowering because if it's neutral then

59:47we have the ability to influence it I

59:51like which direction I've heard other

59:53people use like it's like a mirror right

59:54it like it mirrors back to humans

59:58ultimately certain things that maybe are

01:00:00below the surface sometimes yeah that's

01:00:03good I like that but taking a little bit

01:00:06more of an optimistic lens I've seen AI

01:00:08used in a ton of different ways like I

01:00:11saw something recently which was like an

01:00:13AI that helped write Excel formulas

01:00:16which I thought was like quirky and neat

01:00:18um I also saw one that like takes legal

01:00:21jargon and simplifies it for the end

01:00:24user interesting um

01:00:27I tested it only a little bit it did at

01:00:29least the prompts I played around with

01:00:31did but I didn't play around for very

01:00:33long

01:00:34um curious to know from you do you have

01:00:35like a wish list are there things that

01:00:38AI has not touched yet where you're like

01:00:39man I would really love this and while

01:00:41you think about it I'll just quickly

01:00:43share one super simple but we're doing

01:00:45these podcast recordings with video and

01:00:48every so often I would just love to

01:00:51resay something right and things like

01:00:54the script allow you to resay something

01:00:56with audio very easily but I want that

01:00:59with video because if we're airing the

01:01:01video versions of these I want to just

01:01:03be able to be like remove my filler

01:01:04words and you know fix up

01:01:07my performance but obviously it'll look

01:01:10ridiculous if the video portion

01:01:13isn't equivalent so that's my little

01:01:14wish list for anyone at open AI or

01:01:16otherwise to to do I know they're

01:01:18probably coming out with video at some

01:01:20point but to sync the two what I have

01:01:22wanted for AI is video tools for

01:01:26creators that are AI powered that's why

01:01:28I've been doing so much hacking trying

01:01:30to get like these image tools to work

01:01:31for video because it's not quite there

01:01:33yet but I keep on like trying to make it

01:01:34work once it does though like

01:01:39I don't even know what's gonna happen

01:01:44and I was like be careful

01:01:46what you know

01:01:47yeah I mean it's coming right right it's

01:01:50it's for sure coming

01:01:52um

01:01:54it's hard for me to

01:01:57say a wish list

01:02:00I guess because I'm just trying to keep

01:02:02up with what's already coming out oh

01:02:04totally something we talked about before

01:02:06we jumped on which is like a neat little

01:02:08idea um just to get the wheels spinning

01:02:10for any listeners as we were talking

01:02:12about audio AI have you actually seen

01:02:15um any AI ASMR I wonder oh no I've never

01:02:19heard of that no no I haven't either but

01:02:21I was just you know what would be

01:02:22interesting with ASMR is like

01:02:24if they could like

01:02:26uh connect to your senses and measure

01:02:29like your response to it oh yeah and

01:02:31then like in real time like adjusting

01:02:34ASMR update it to like your sensory

01:02:37response

01:02:38oh I like that

01:02:42startup idea I mean people are already

01:02:44playing around with like audio Ai and so

01:02:47obviously I haven't seen the ASMR stuff

01:02:50but I feel like that that's coming soon

01:02:52and I've heard other people talk about

01:02:54this in the realm of video as an example

01:02:57where you're let's say in a movie and

01:02:59this is you know a little further away

01:03:01but based on your prompts it goes

01:03:04another direction right and so it's

01:03:06almost like personalized content to the

01:03:09consumer based on their needs or their

01:03:12prompts and so that's an interesting I

01:03:14have an answer for your question here's

01:03:15my wish list for AI my wish list is

01:03:18actually for all humans to take the

01:03:22ethics of it very seriously

01:03:26um and what I mean by that is like

01:03:30my wish list is

01:03:32that everyone working on AI be held to a

01:03:36standard to be using it for positive

01:03:38forces rather than negative

01:03:40and that we somehow develop an insane

01:03:43amount of cultural pressure

01:03:46that anyone who's not using for positive

01:03:48sources is like

01:03:52not ostracizes from society but but

01:03:54penalized in a way that it is severely

01:03:56incentivized

01:03:58how do you think about what is good and

01:04:02bad though

01:04:03in that context okay so I mean it's very

01:04:06hard to be the Arbiter of that but for

01:04:07example like deception is bad

01:04:10right so when you're altering things you

01:04:13need to disclose what you altered and

01:04:16how do you think that

01:04:18we will need to somehow label things as

01:04:23AI generated or human generated because

01:04:25you know a simple example you brought up

01:04:27deception but people do these deep fakes

01:04:29right of people without their consent

01:04:32they're not quite there yet but they're

01:04:34very close to being right well have you

01:04:36seen the Tom Cruise ones they are

01:04:38actually fully

01:04:40fakes I I can't tell I think most people

01:04:43can't tell okay so so we're already

01:04:45there it's not even a future we're

01:04:47already there on edge cases and we'll

01:04:49pretty soon be there for yeah Mass cases

01:04:51I think that will be developed I think I

01:04:54think um it will be necessary and it

01:04:55will happen and I think it'll be very

01:04:57similar to like nutrition facts on like

01:04:59all the food like it's a universal

01:05:01standard that we have nutrition facts

01:05:03and so there will need to be some sort

01:05:05of system that like

01:05:07shows you know this is a video this was

01:05:09produced in this way I've heard people

01:05:11talk about this and it sounds like I

01:05:13might be wrong because I've just you

01:05:15know just listening to another podcast

01:05:16but it likely would have to be at the

01:05:18hardware level as in run by like Apple

01:05:21or Google Etc because

01:05:24um if it's not at the device level

01:05:25that's like really easy to hack or to to

01:05:28abuse right and so it literally would

01:05:30have to be like okay this AI is running

01:05:32on this machine and this machine is

01:05:35plugged into stable diffusion or

01:05:37whatever insert AI here and the output

01:05:40in the metadata like and it would have

01:05:43to be

01:05:44like a case where you could not remove

01:05:47it

01:05:47has that information it would have to be

01:05:50the hardware level and I think it will

01:05:52be first adapted in like really high

01:05:54stakes scenarios where the veracity of

01:05:56something is extremely important so for

01:05:58example in like political messaging

01:05:59maybe they'll have like a very specific

01:06:02kind of camera or Hardware where it

01:06:06cannot be altered and then that is you

01:06:08know so for certain types of like really

01:06:11important messaging where it's really

01:06:13important that it not be faked whereas

01:06:15like you know like someone's social

01:06:17media influencer video like it's doesn't

01:06:20matter if something was like adjusted or

01:06:22tweaked you know so there's like

01:06:24standards or that's part of the

01:06:26regulation that in certain Fields you

01:06:29must use certain Hardware or at least

01:06:31you must use like a suite of Hardware

01:06:33that that plugs into these I mean

01:06:35regulation will catch up but I actually

01:06:36think it will be somewhat self-regulated

01:06:38in the sense that some people will need

01:06:41the design will have the need to show

01:06:43that what they do is true and not

01:06:46altered and then there will be companies

01:06:48that pop up that say okay this is the

01:06:51technology that does that and then there

01:06:52will be people who then buy that

01:06:54technology and use it of their own free

01:06:55will and Accord not because of

01:06:57regulatory requirements because it'll

01:06:59take a while for the regulation to catch

01:07:01up yeah but I do Wonder in that case

01:07:03like yes the good actors will use the

01:07:06technology that has that functionality

01:07:08but then that actors likely won't right

01:07:10unless there's regulation that's that's

01:07:13why I'm that's why I say that like

01:07:15somehow I want my wish list is for the

01:07:18culture to change

01:07:20because humans are self-enforcing

01:07:22Society you know we

01:07:25are very naturally tribal and we behave

01:07:29well largely because that's what's

01:07:31expected of us from other humans it's

01:07:33not like someone's watching us at every

01:07:35single moment but we still in general

01:07:37you know want to behave well

01:07:39and a lot of that is the cues that

01:07:42you're getting from culture and so I

01:07:43would love to see a culture developed

01:07:45where hey as as humans as part of being

01:07:48a good human

01:07:50um we use technology responsibly and

01:07:54ethically I don't know how that's going

01:07:55to happen and

01:07:57that's my that's why it's my wish list I

01:08:01love going down these rabbit holes

01:08:02because it does just get the wheel

01:08:03spinning I'm like man there is there's

01:08:04so much to figure out and when you're

01:08:06dealing with an exponential technology

01:08:09policy regulation IP a lot of the stuff

01:08:11it takes time right and so it like I

01:08:14also wonder whether we create

01:08:17regulation or policy let's say in 2023

01:08:21and then by 2025 like it's it's already

01:08:24just like you know it's completely

01:08:26different I actually think culture can

01:08:28change a lot faster than regulation much

01:08:30much much faster and I'll give you an

01:08:33example the fastest I've ever seen

01:08:35culture change was the handshake during

01:08:38covet you know how many like

01:08:42how many thousands of years or hundreds

01:08:44of years has the handshake been the

01:08:46golden standard

01:08:47and then all of a sudden like socially

01:08:49unacceptable so many things became

01:08:51socially unacceptable almost overnight

01:08:53because of covet and I was like Wow

01:08:56human culture can change like that and

01:08:59then now that the pandemic is better

01:09:00then like we're back to doing it again

01:09:02so

01:09:04um I do think that uh we're gonna need a

01:09:06lot of cultural forces

01:09:08um to kind of Sprint ahead of Regulation

01:09:10because regulation will take time yeah

01:09:12well

01:09:13I have no idea how this will progress

01:09:16It's been fascinating to hear how you

01:09:17think it might progress as someone who

01:09:19has played around with these tools then

01:09:21we'll be sharing a lot of your projects

01:09:23as we talk about them throughout this

01:09:24episode

01:09:25um but I want to end with something

01:09:27really fun which is that you've worked

01:09:29with tons of creators uh creators that

01:09:32people know Critters that people don't

01:09:34know

01:09:36um do you have a a wish list or is there

01:09:38one creator that you would just love to

01:09:40partner with I love throwing this stuff

01:09:41out into the ether because you never

01:09:44know you never know who might make this

01:09:46happen you know it's interesting because

01:09:47in the past I would have had like a

01:09:50dream client or a dream like something

01:09:52but I've actually learned over time that

01:09:55it's not about like which celebrity you

01:09:58can get or which brand you can get but

01:10:00it's for me I've learned it's actually

01:10:01about doing my best ideas my best work

01:10:03and so the way I do things is I try to

01:10:06match make like my ideas to a client who

01:10:09will accept them unaltered essentially

01:10:11or with the least alterations possible

01:10:13so I don't I actually

01:10:15don't won't answer that question or

01:10:17don't even have actually nothing

01:10:18actually comes to mind because it's it's

01:10:21mostly about like

01:10:23uh making the ideas that I want to make

01:10:25yeah yeah I love that it used to be or

01:10:28at least for me when I was younger it

01:10:30used to be like I want to work with

01:10:31these people or I want these companies

01:10:33to hire me and I think you're right

01:10:34there is like an evolution like we

01:10:35talked about before where you're almost

01:10:36like transcending attention and you've

01:10:39gotten to a place where actually your

01:10:40focus is how do I make something really

01:10:43yes it's almost like it's like almost

01:10:44like the mass was hierarchy of needs and

01:10:46it's like oh I'm actually I'm I'm

01:10:48actually just I'm actually focused on

01:10:49like doing the ideas that I want to do

01:10:52whereas in the past like I know 10 years

01:10:53ago I would have been like I want to

01:10:54dance in a Coca-Cola commercial like I'm

01:10:56not I'm good I don't need to do that now

01:10:58and then like even two years ago I was

01:10:59like I really really really really want

01:11:01to make a commercial for the Apple

01:11:03iPhone and now I'm like I'm good like I

01:11:05I don't actually need to do that anymore

01:11:06so

01:11:08um uh so yeah I I don't have I don't

01:11:11have one because it's actually about the

01:11:13idea and it actually is so cheesy it

01:11:15sounds so cheesy but it actually is

01:11:16about enjoying making the idea well

01:11:20that's a wonderful place to end things

01:11:21off so Karen thank you so much uh for

01:11:24talking with us today this was so fun to

01:11:27just think through you know the

01:11:28optimistic the pessimistic the dystopian

01:11:31utopian

01:11:32Futures that we might live in but at the

01:11:34end of the day I think one of the

01:11:36biggest takeaways for me is that AI is a

01:11:38tool and that tool is being democratized

01:11:40so that in itself I think is exciting

01:11:43for many people to be able to

01:11:44participate in that future

01:11:46yeah thank you so much thank you for

01:11:48having me thank you for these great

01:11:49questions

01:11:52thanks for listening to the a16z podcast

01:11:54if you like this episode don't forget to

01:11:56subscribe here on YouTube to get our

01:11:57exclusive video content we'll see you

01:12:00next time

01:12:02[Music]

🎥 Related Videos

What vaccinating vampire bats can teach us about pandemics | Daniel Streicker

a16z Podcast | Things Come Together -- Truths about Tech in Africa

2024 TSCRS Applications of anterior segments diagnostic instruments in cataract surgery

a16z Podcast | The Infrastructure of Total Health

The Robot Lawyer Resistance with Joshua Browder of DoNotPay

NES Controllers Explained

🔥 Recently Summarized Examples

Uncovering Corruption From Health "Experts" | Scott Carney

The Forgotten Geometry: A New Path to Unification

Bitcoin Holders...Watch Out for Sept

View original video