I'm Yasmin Gagne. I'm Josh Christensen. And this is Most Innovative Companies. On today's episode, global design editor Mark Wilson. The CEO of Mischief, Gabe, like somebody screen grabbed him and was like, "He's playing this game?" And he got voted out within the first few days. Pinterest chief product officer Sabrina Ellis. AI actually helps with that. So you've pinned all these things and we're like, "Oh, you know what? You like Zen industrialist or you like Japondi."
And as always, keeping tabs. He doesn't believe the theory of gravity. He thinks it's the draft left by electricity. But first, here's the download. The news you need to know this week in the world of business and innovation.
A federal appeals court is allowing part of Biden's student loan repayment plan to resume. I think this is try number 72 on student loan repayment. The saving on a valuable education or save plan can now reduce payments depending on the borrower's income. But another part of the plan that would forgive loans entirely is still blocked by a federal judge in Missouri.
Just love how much power individual judges hold. It's so insane. Anyway...
So last week we talked about the European Commission coming after the Apple App Store. This week, the commission is going after Meta. Right now, European users must accept versions of Facebook and Instagram with personalized ads unless they pay up to $14 every month for versions without ads. The commission says this violates new laws regarding digital competition. Just get the ads off the fucking platforms. Yeah, just stop.
Stop it all. You know what? The European Commission is the hero we all need right now. They're doing great work with these antitrust and violation laws against big tech. Next news story, Boeing will buy Spirit Aero Systems. The two companies are recombining in the hopes of solving some of Boeing's airplane safety problems.
Spirit, which makes many parts for Boeing, including the fuselage of the 737 MAX, was sold to private equity in 2005. Boeing now admits that various supply chain safety problems, including the infamous door blowout, came from having two different companies work on the same planes. Why pay someone else to do a job when you can do it yourself? Am I right? Exactly. Or buy a company to do it for you. I want to buy a company to do my job.
Photos on Instagram will no longer be labeled made with AI, but instead they'll have an AI info dropdown with more details. The change comes after some photographers claimed their photos were given the AI label after using only simple editing software. Nice. Well, you know what I always want when I'm looking at photos? What? Reading. Yes.
The more I can get to read while I'm looking at photos and doom scrolling through, I'm never going to click on those AI info things on Instagram. No, are you kidding me? I mean, hopefully it makes the photographers feel better. That seems like it kind of sucks. And finally, Tractor Supply, which calls itself the, quote, largest rural lifestyle retailer, quote, listen, Bass Pro Shop probably has some words to say about that. Yeah.
They announced a rollback of their carbon emissions goal. Don't like that. They'll also eliminate all DEI positions. Also don't like that. And stop supporting pride parades or voting campaigns. Three for three right there. No bueno. That feels like a real, like, kick you when you're down moment. They're like, no voting campaigns. It's kind of interesting to see how thin...
companies support for marginalized communities are basically when it stops being more profitable to attract a more diverse market, they'll quickly abandon those causes. Yeah. And that's the news you need to know today. So Josh, what is your last Venmo log?
Oh, man. I do exclusively use Venmo, and I use it a lot. I think what my actual last Venmo log is something pretty boring, like paying my brother for my cell phone bill on our family plan. But other than that, which is a monthly expense, I think the last thing I paid for was my portion of dinner for me and my wife's friend Justin's 40th birthday, which was super fun. Happy birthday, Justin. And yeah, that was great. We went to this really great restaurant in Hell's Kitchen.
It was too expensive for what I wanted to spend at that time. That's an oxymoron, but okay. But it was very good and very fun night. What about you, Yaz? What's your last transaction on Venmo?
Brendan, our editor-in-chief, had to Venmo me because I buy snacks for the print team. That's true. You do have peanut butter pretzels over on your side of the office all the time. And you know what? He, I think for a while, thought that like the office paid for them or something because he kept talking about the snacks with other people. And I was like, Brendan, Brendan, buddy, buddy, you have to send me money. Like I'm spending my money on these snacks that you keep telling other people about, which then means more people come for them.
It was like a big problem. And also I needed to be respectful about it, which was hard. I do eat a lot of those pretzels.
They're so good. I know. I know. I should Venmo you five bucks for the pretzels. Yeah, yeah, yeah. Anyway, he was like, congrats on going from renting to owning because he like looked through my Venmo history and I was like, I got to make this shit private. I went private a while ago, but that's my favorite thing about Venmo is just stalking people. Lurking? Yeah, lurking. Stalking people. Well, the reason we're talking about Venmo is because now it's been turned into a game. Venmo.
It's called Mischief Plays Venmo. You sign up for $10 and every hour you're able to vote to eliminate someone from the game. With a current prize pool of over... This will surely be a blood money bloodbath. Is anybody else playing this game? Because if y'all are, we need to like...
Team up or something because we need to get this money. Brooklyn-based art collective Mischief, spelled M-S-C-H-F, like the worst bar you've ever been to, has created a lottery game on Venmo with a survivor-like voting system. And here to explain everything is Fast Company Global Design Editor, Mark Wilson. Hey, Mark. Thanks for coming back on the show. Thank you for having me. So, Mischief has been an MIC company before, and some of our listeners may be familiar with their projects.
But most people probably don't really know what this company is. So what is Mischief? Mischief is a Brooklyn-based art collective. But what that really means is about 30 people. They make a collection of like products you can buy for as like a little as $10 to really high art sort of like blue chip art.
art pieces that end up in galleries. They do everything in between. They like to poke at big brands, they like to poke at companies, they like to, I guess, instigate a little bit. What type of projects is Mischief known for? Mischief is best known for footwear, honestly.
Probably most famously, the big red boot. Each shoe was bigger than my head. So I put my whole leg in these things and tried them on with pants and shorts. And they really make you feel like a cartoon character. I remember the red shoes. I guess they're kind of like Astro Boy, but they were just like these big, fluffy, marshmallow EVA foam boots. And when they came out, they just went, you know, like hyperviral. They were like everywhere instantly. They made Mischief a lot of money. We don't exactly know how much.
But they've done a lot of other shoes, too. Like, you probably remember one of their first shops was the Jesus shoe, which was filled with, like, water from the Jordan River. Then they followed that up with the Sane shoe with Lil Nas X. So inside of this little...
There's actually real blood in here. How expensive were these shoes? That Jesus shoe was $1,425. However, that was like super limited edition. The Astro boots were a few hundred bucks. I think they debuted for $350, but the price may have dropped. It was right in there. Like you can get a few of them for $1,000.
150, 175. Some of them are just like on par with Nike's. They're not like super high design, unaffordable things. Do they work to engineer viral moments or do they really just put the product out in the world and see what happens?
I think engineering virality is an interesting question for anybody at Mischief. They see virality not as the ends, but as sort of the means to the ends. Really, they're about making art on the biggest, most sort of participatory scale as they can. So like naturally, social media is the way they do that, right? It's sort of the way they deliver and converse with the crowd.
Like these are artists at their core, right? Like these are people who actually like tried to make it work in marketing and they couldn't stand it. And so they started their own company and just learned how to make money making art, if that makes sense.
Right. But it's also marketing, kind of. It's also marketing. Right. I mean, all art is a little bit of marketing. They've had a lot of projects where they bring in the crowd and impart to like poke at big IP and brands. They had this thing called the Cease and Desist Grand Prix, where they printed Amazon and Subway and, you know, all of these Disney and a few other big brands like on racing gear. And the whole premise was that
whichever company calls and complains first and sends us a cease and desist order, they win. I love that. They knew none of these companies could sort of resist, you know, complaining in this legal manner. Ultimately, Subway did it. But I think Subway actually kind of, like, they tried to embrace it. You know, like, hey, we won. Yeah, yeah, yeah. They tried to be cool. It's like when actors show up to win the Razzie, you know? Yeah, exactly. Guys, I'm in on it. That's exactly what the situation was. Yeah.
So back to Mischief Plays Venmo, which is the game. How does it work? So basically, Mischief Plays Venmo is a website. You go to the website and it almost looks like a Game Boy, like a Grecian Game Boy, if you will. Essentially, what you're doing is you're playing a game that is built invisibly upon Venmo itself. You use your Venmo login to play a game that's on this website.
And Mischief was actually banned from Venmo years ago. They apparently tried to build a lot of their platform on top of Venmo back in the day and have since been banned from the company. So they set up apparently a shell company to set up this game. Nice. The way Mischief plays Venmo works is that anyone with a Venmo account could essentially sign up, put in $10 to a pot,
in a sort of winner-take-all lottery. And every hour, someone's voted off, much like the CBS game Survivor, but sort of like faster and ongoing. And a winner is crowned when someone gets 100% of the votes. So basically, people will keep getting voted off and off and off and off until everyone can decide on who wins. And that might be when there are two or three people left. Like, who knows? Have you been voted off yet? Not yet. I mean...
Soon. I'm going to vote you off. You can make it happen. It's actually really fascinating how it's working. So it's like it's based upon this ancient Athenian practice of ostracism, right, where people would write names on clay pots and whoever's name was like got the most votes would just be kicked out of Athens for a decade. That's so crazy. Honestly, bring that bring that back. Bring it back. Yeah, bring it back. There's a few people we should get out of here.
It's interesting because like they point out it was famous people. It was like politicians, you know, who would be kicked out of society. It wasn't like everyday people. Right. Which is a reason I'm still in the game because nobody's voting for me to win. Yet. Like I just like sort of like the non-existent level. The CEO of Mischief, Gabe, was already called out on Reddit. Like somebody screen grabbed him and was like, he's playing this game. And he got like voted out within the first few days.
So like you could tell the crowd is definitely pitchfork enthused. How big is the crowd? I think the pot is $109,000. So I guess that means about 10,000 players. Is Venmo like into this? Are they cool with all of this? Is Venmo getting anything out of this? I actually think Venmo is getting a lot out of it because we're talking about Venmo on here. And like Venmo is so weird. And I generally would not be talking about Venmo. More of a Zell man.
But Venmo's like non-official partner on this. They had banned Mischief off their own platform years ago. I think this is somewhat of a protest against Venmo that is also probably going to work out really well for Venmo because, again, like we're talking about. What is Mischief's kind of like central guiding principle? Or are they just kind of like letting artists be artists? Or do they have sort of like a central mission? I think artists be artists...
One of the interesting things about Mischief is you look at their projects and they're just like masters of virality, right? But they aren't really in the meme culture. They don't respond to moments that happen, you know, in mass culture. They actually all like brainstorm ideas. They put those ideas into a big pile. They measure like how feasible they are. They can actually like bring these products to fruition because a lot of them work with all sorts of different parts of the supply chain across the world.
And then they can sit on them for like months or years. They're sort of like not flipping these ideas really rapidly, as you might think. They're actually like letting them percolate and sometimes like just like dealing with logistical issues to make them happen. What do you think they're trying to say by getting people involved? I think like their art can only exist with the Internet. And they want to create something that's sort of of this moment of like brand, audience and like diversity.
maybe bullshit for lack of a better term, right? Like, you know, I've like written like they hold a fun house mirror up to culture. What's unique about Midship, I think, is that they sort of weaponize the brand and the platform sort of against like the people who hold all the power.
And Mischief's fans, you can buy collectibles and sort of there are all those aspects to Mischief. But I think Mischief's fans just like want to fund something that is like poking back at power systems in some way. So wait, when we're talking about funding, you mean just buying stuff, right? Or is Mischief a nonprofit? No, no, no. They are not a nonprofit. Mischief has tens of millions of dollars in funding from VCs. So definitely not a not-for-profit. They're VC-backed?
They are VC bats. Yeah, that's something like interesting and intrinsic to the core of mischief, right? Like they call themselves an art collective, but they are also like
backed by funding. How much have they raised? Publicly, I think about $24 million. Oh, my God. The thing about Mischief and like why they, I think they did need to raise money, though, I will say is like making products is really expensive. They are actually, on one hand, like a digital design firm. On another hand, an industrial design firm. They also build fashion, right? And so if you looked at this complete spectrum of products they do, there is like,
it does take a stupid amount of money to bring these things to fruition. And I also think what's interesting is like a lot of their stuff doesn't make money. So...
We talked a little bit about the fact that Mischief is sort of an organization of 30 artists, but tell us a little bit more about Mischief's founder and the company's history. Mischief kind of got a slow rolling start. You know, one of the founders, Gabe, essentially had his own independent digital art practice and he was doing sort of moonlighting and marketing. And then he joined up with these two other creatives, Lucas and Kevin, and they were doing a lot of the same sort of stuff he was doing. They were working on like
these viral projects that weren't built to make money whatsoever. Lucas and Kevin had this fake company where they promised to clone famous human beings into sausages.
So like these are the types of things they were doing, you know, but eventually they sort of coalesced and became mischief. It's crazy. These guys all sound like sort of the worst people I had crushes on in college. They'd be like, look at my Internet art. And I'd be like, fuck this. Normally, this is like not my thing, like this type of stuff. But like, I generally think that these guys are pretty clever and and this stuff is pretty cool. But you
You know, going back to the big red boot, it sounds like it's not just a stunt, right? There's some level of craftsmanship going on. They have a supply chain that they're working with. When you see a lot of these objects online, they look like they can just be like spun up. 3D printed. Yeah, exactly. But when you actually see them in person, the execution is really high end. The big red boot is like a well-made, well-articulated shoe. Like, I don't know a better way to put it. It
They're working with EVA foam in a very interesting way with that object. And seeing their stuff go viral online is almost a disservice sometimes because it's better done than it probably needs to be.
But then people see your art, right? It's like writing a really good article and having it go viral for the wrong reasons online. Like at least people are reading it. Yes. All press is good press is essentially the point we're ending on. All press is good. I think that's a good place to end this segment, Mark. Thank you so much. We'll be back with Mark for Keeping Tabs. But first, a quick break followed by my interview with Pinterest Chief Product Officer Sabrina Ellis. ♪
For the past 15 years, Pinterest, the visual search engine and mood board creator, has had a unique foothold in the creative industry. Now, as AI-generated images are taking over the internet, the company faces an interesting challenge figuring out how to classify them.
I spoke with the company's chief product officer, Sabrina Ellis, to learn how the platform is catering to diverse audiences and how it handles AI-generated images. I started by asking Sabrina how generative AI is affecting the platform right now. Yeah, it's kind of interesting. I feel like people come to Pinterest to find their inspiration. Now, a lot of that, though, I find most of it is still grounded in what they want their lives to be like.
In some ways, if they're finding things that are Gen AI created, it defeats the purpose a little bit because often you find that awesome dress, you find that awesome table. The fact is, in general, you want to buy that dress and you want to buy that table. And so with Gen AI, you may find it beautiful, but if it's not viable, it kind of defeats the purpose and people feel a little frustrated with that.
That's interesting. So you see it, I mean, at this point when people are using Pinterest, it is a vehicle for shopping a lot of the time. Yeah. I mean, I think the majority of the people and they say they're like, actually, I'm shopping for the thing. I could see situations where sometimes people also, though, will look at it for inspiration for some of their hobbies. It could be anything from I want a new idea for tattoo or I want a new idea for painting.
then I think Gen AI could be something that is a little bit more fitting for that because you're creating from it and you can bring it to reality without needing to be shoppable.
What were or are some of the biases you're seeing in the content initially? Where does that come from? If you look at the content, at the content that marketers, that advertisers are creating, I think it's inherently biased. I don't know that it's reflective of today's society, at least not the ones often that we see when we look in the mirror or when we look at our friends. It is biased.
still skewed, I would say, towards Caucasian, in women's fashion, towards Caucasian, thinner women. And I think many people are looking at that and being like, well, I don't know that that's me, whether it's in fashion, whether it's in beauty. And I think it was important for us to look at that and say that tends to be what marketers and advertisers are creating in their content. And when people are pinning and repinning and sharing, that's the material often that they're starting with. As Pinterest was trying to offer viewers maybe more
diverse options or options more reflective of society. You know, now when you search for, say, a dress or something, you may see it on a bunch of different body types or skin tones. Is that fair to say? I think you're exactly right. There's kind of two phases to it. There's one of actually identifying what this content even is. Yeah. And being able to classify it.
allowing people to actually choose, hey, I want to see something that looks more like me. And then the third and final step is that we actually make that diverse by default. So it's like if I searched, you know, I don't know, like A-line green dress, it would come up on a diverse group sort of automatically without me selecting that?
That is what we are trying to achieve. Now, there is still a skew that there isn't enough content, I think, to share. And by the way, ironically, I literally did a query this last weekend that was a line gown and it wasn't green. It was like aqua. So I don't know. I just want to note that. Yeah, there's something in there. Yeah, that's super weird.
I think something, as you were saying, that's interesting is like there's a preponderance on the site of content that features sort of white, thin women. You want it when you search automatically come up on a sort of range of body types. That's right. But how do you sort of make sure that more content gets created on diverse body types?
Well, look, we're not the ones who are creating that content. Yeah. So this is why, you know, I spoke, I will speak at different places. We actually try to really promote it. And we're trying to talk to those people who are the ones who are creating the content, the marketers, the advertisers, the even entertainment creators. It's like,
look, you all need to be in this because I think it's important for us to be reflecting what is in society. Some people did ask. They're like, look, in the age of Gen AI, why don't you just create it? So that was going to be my next question. What's the policy around that? Well, that's kind of horrible to me. Like it's just, everyone's doing all this real content with like skinny white women, but then we have to go to Gen AI to create any diverse. I just, to me inherently, that just seems wrong.
I think it's much more important for us to all really be trying to diversify who's in front of the camera and frankly, even who's behind the camera, because I think they're going to be the ones that are going to notice it more too. That makes a lot of sense. Something that, you
You know, you touched on earlier, which is that Pinterest is really like a shopping platform for a lot of people. As you've rolled out changes, you know, as you've rolled out stuff like now you're able to select what kind of sort of hair pattern you have, for example, when you're looking at hairstyles. Has that corresponded to an uptick in sales? We've seen it in terms of engagement.
Right? So that's the first step. So features are pretty new. What we're seeing, like, as far as hair pattern searches, that's up like 150% year-over-year. And as far as the engagement of people who filter on body type, their engagement's up 66%. And I think...
It makes sense. I was with my cousins this last weekend and we were actually at these kind of kitschy gift shops and everybody had all those name things. And we found these key chains that had our names on it. What happens when you see your name? You're like, oh my gosh. And let me tell you, Sabrina, I was like, I got to buy it now. It's rarely there. And I think it's that phenomena that you feel when people are looking at Pinterest and then they see themselves.
and they feel seen and they feel like this is for me, I want to engage more, I want to be here. We love that and that's what we want to bring to people. So yes, the first step is engagement from there, shopping comes from there. It's still early on some of those stats because the feature is just launched. What does tagging look like on the back end for that? Presumably, this involves categorizing millions of images. Tell me about how that works.
Actually, it's 3.5 billion images. 3.5 billion, wow. That's not all of Pinterest. That's actually just the ones that are in women's fashion that we're looking at categorizing. So very, very large number. And this is where AI does help us a lot in being able to process and categorize that large of a number. When you look at any given image, we actually have a whole slew of different information about that.
So that can be anything from this is a top, the color is red. We may or may not know the brand. One of the things we're looking at with AI is actually to start almost categorizing the type of brand, because then you can say, oh, I noticed that you like this type of brand. Here's another one that we know is related. So all of that information is in there. And one of the things that we needed to add with our processing was what is the body type
of the person in this women's fashion image. So a lot of technology went into that because you got to look at everything from, well, how do you categorize somebody who's in a flowy dress? How do you categorize someone who you only see from the waist up? All of those things are pretty tough. Our team did a lot of work technically first to actually say, how do we actually build it up? How do you actually use technology to categorize that?
Then from there, we've got experts that have built a lot of inclusive features where they would say, okay, knowing that now, how do we actually help understand? How do we actually help really the meaning of these? The system may have categorized these, but what do these segments look like? What do the clusters look like? And then from there, you know, then you got to decide how do you actually show this to the end user? How is it that they want to figure out like, what type am I? And so there's a lot of work there working with experts and getting a lot of feedback as well.
You previously worked on Android at Google, is that correct? I did. I was talking to a friend who actually works on that team, and he was saying that an interesting challenge with Android is that people in different continents and countries behave very differently and use the product really differently. I'm curious whether that's factored into any of your work at Pinterest, especially when it comes to these kinds of sort of services.
search results that you're working on? Pinterest is actually a global product. And actually, we're going fastest outside the United States. And so we're going fastest with Gen Z. That's growing like 20% year over year. And it's 42% of our market now. But we're also growing. Our growth is like 80% outside the United States.
So then you are looking at that content and how is it that you surface it, the right content for the right person at the right time. A couple of things that we are finding. One, actually, when you're looking at age groups, Gen Z content resonates the most with everyone, which is interesting. I think there's something where
I am much older, but I think we all want to be cool. And somehow Gen Z is producing the content that's resonating the most. The other that we find as far as the inspirational side, that actually crosses internationally quite well because it's visual.
I think in other forums where you're doing things like written content or video content that has language, there are more barriers. When you were talking about imagery, it often will actually still connect across different countries and international lines. The thing, the challenge that we have is actually much more what we were talking about earlier, making everything shoppable.
Now that is actually more of it because I may find something that I love, but if they only ship in France, then we have to find some equivalent and some match that I can actually purchase in my home country. So that's something that we're working on and that we need to think about.
As we look at some of these features, like some of the body type features, we are first launching in United States and Canada. We very soon will be launching in Europe. And then we'll actually need to take a look at different international markets and what resonates most with that. Why might those international markets be different from like a U.S. and a Europe? Yeah, it's both the content that might, you know, again, you want to be seen and you like to see yourself. But actually, somewhat, it's actually our method of how we've actually defined the types too.
One of the things I was saying, you kind of do the clusters, you work with the experts, you figure out what are the types. We worked with a lot of experts to figure out how many types should they even be? Should there be eight? Should there be six? How much is too much for people to not feel like they can actually even choose? We settled on four, and that was through a lot of expert feedback as well as user feedback. The other thing we figured was before we were trying to describe them with words, it was
We did a lot of research on that. It's very interesting. Things like the word stocky, some people just did not like. The word curvy, a lot of people actually were quite comfortable with. And by the way, that's the stuff when you're looking at AI. AI is never going to tell you that people don't like stocky and that people like curvy. That is all about really adding the human element to it. So we did all of that, but you know what we really came down to? It was better to use as few words as possible.
And so if you look at our types, we literally have four rows where it shows a variety of people in similar outfits and clothing so that people can identify and understand what those types are and figure out where they fit it. And we work really hard on those types.
Even the photography for those types, because like we have someone in a wheelchair, we have someone in a cane in one of them. And that's really important. It actually just adds to all of the inclusivity that we feel. Now, the reason I explain all that is because as we actually go through other countries, the thing that we find is we need to make sure that people feel seen in those categories, in those type definitions.
If it's too far where, you know, in the United States, it might be a whole set of different ethnicities in there.
We might need to do something quite different in Asia. That's interesting. Got it. Yeah. So I think those are some of the things that we are exploring on how we can roll out. So this is why I think it's easier in the United States. I think Europe and Australia would not be as different, but we want to test that out. We need to test out the audience of, do you feel seen here? Do you feel like you could identify yourself in one of these categories?
Is there a world where your line of business could say, we've seen X amount of people in this size and they're all searching for this one product? You know what I mean? Like a red swimsuit. And you would sell that insight back to company? You know what I mean? Like how far can you take this data?
It's kind of interesting. We have something about trends. Yeah, I see the trend report. Yeah. That one is fascinating. I've only been at Pinterest for less than a year now, but I think they've done that for several years. If you go back in history, they have been right over 80% of the time.
So we already actually look at what people are doing on Pinterest, how they're finding their inspiration. And we're finding it's very early in the funnel. I was talking with the CMO of a major beauty firm and she was saying that. She was saying, you know, on these other sites, I find people are maybe showing their inspiration. It's like, I think they're finding it.
on Pinterest. So when you talk about that, yes, all of that information has been very useful to us to optimize our product, but we also do help some of our partners understand what is happening in the industry. And we will tell them, a lot of them look to us for it. They're like, some of the beauty executives have said,
We will build whole and we will create whole new products based on some of the insights that you're sharing. What do you think is different about, say, the Gen Z TikTok user versus the Gen Z Pinterest user? Or what is different, even if it's in the same person, about the kind of behavior and discovery they're doing? Yeah, what we hear from...
our Pinterest users, it's actually pretty heartening. I mean, this I love because what they, the bottom line, if I look at it like a 15-year-old, like I look at my niece, it's a tough time right now. There's a lot of pressures going on for them, right? And so what she'll say to me is like, okay, look, honestly, Instagram, she feels a lot of pressure that she needs to look good there. And it's looking good for other people or others. They feel like, well, there I'm actually consuming information that others have told me
is important. I put that in air quotes, important for me to know. Where what I love is what they tell me, Pinterest is for me, right? Pinterest is for me, what I want, what I'm interested in exploring. I think the other ones are sort of what I would call lean back experiences and
Pinterest is I'm leaning forward. I'm exploring how I want to decorate my dorm room. I'm exploring what I want to wear for that, you know, for the prom, for that party. All of those things are actually engaging activities. Actually, just last week, I think Megan Thee Stallion was interviewed and she literally said, I'm not going to share my handle with you.
Because she's like, this isn't for all of you to learn about me on Pinterest. She's like, I love Pinterest, but I'm searching for all these things, everything from booty shorts to whatever. But she was very clear. She's like, this is for me. It's not for all of you. I love that because I think we kind of need a little bit more of that right now in today's society. Just because AI has been such a focus at Fast Company, what are some of the other big product initiatives that harness that technology and leverage
I guess the flip side of this is what are the sort of dangers or maybe negative aspects of generative AI proliferating on a platform like Pinterest? You know, we talked about the three and a half billion images in women's fashion.
We have tens of billions of images overall. So the ability to label those, to categorize those, to actually associate text with images, that's a huge benefit for us. That really unlocks a lot of power for us across the company. What that does, I'll tell you features that will show up in the end user experience then. Things like guided search. If you search for something like the A-line green dress, we're able to use AI to suggest ways that you might be able to filter it.
Do you want full length? Do you want MIDI? And AI can actually use that on any category to actually suggest filters. The other thing we find, a lot of people come to Pinterest and they love to learn about themselves. They always say, you know what? I know what I like when I see it, but I don't know how to describe that. And so AI actually helps with that. So you've pinned all these things and we're like, oh, you know what? You like everything.
Zen industrialist or you like Japondi, right? People wouldn't have come up with these names before, but people really like it because they feel like they've learned about themselves and then it helps them as they want to search more. And it's the, you know, what is it that I like about Zen industrialist? You know, do I like the more metallic thing? Do I like the colors? You know, those types of things. And in terms of the flip side, is there anything you want to control for or that you're sort of concerned about? Yeah, look,
It's incredibly powerful, but there are places where part of the reason it's powerful is because it's open-ended and you're not exactly sure what it can return. Any place where it's showing up on the UI, even on those suggested filters, we'll review them before they're coming out. The other is that we have decided that we're not going to use it on any categories that have a
people element to it. Right. So we just feel like that type of categorization and it could get risky. Yeah. So we don't tend to do that. Now, if you're categorizing furniture or, you know, style of furniture, it feels safer. And so we've actually started with those and we feel like we can learn from those. If it's more on the people side, we're not going to do that. My
My last question for you is what are your favorite Pinterest searches or do you have any like either hacks or folders or people you like to stalk? Tell me about that. Let's see. I think some of mine, you know, some of the most extreme and I think more creative tend to be on event planning. You know, people really go out with like summer parties or birthday party for five-year-olds. But
The creativity is just, I mean, look, I feel guilty. My kids are older now, but I'm like, clearly I was a failure as a mom because I did not do the things that these people are doing for their children. But I think that that just, on
unlock so much creativity because it's all the elements. It can be anything from how you're decorating, how you're doing food, the experiences that you're creating. And what I tend to like are the ones that have more emotion attached to them. Oh, you know, what am I going to wear to the Taylor Swift concert? That just has a lot of emotion and fun attached to it. I think those are the ones that actually I like the most.
I used to, it's been actually quite a long time since I found it, but I did find Sofia Coppola's Pinterest folders once. Was it really artsy? I mean, they were, she had a folder related to, she was working on an adaptation of an Edith Wharton novel, Custom of the Country.
She's one of my favorite novels. And she had a folder called Sea of Sea that I'm pretty sure was all costume inspiration and mood boardings. And did you see it? Did it come to life? I did. And then I think Appalach's The Project, I actually think she may have figured out that people could find it. She was like, shut that down, shut that down. Yeah, which is fair. And incredibly creepy of me to have figured this out too. I know.
I have to admit, I'm quite impressed by your soccer abilities now. So yeah, that was good. My journalism skills being put to work. Hey, you know what? You studied hard. You got the skills. But this was great. Thank you so much for taking us through all this. Absolutely. Well, thank you so much. Look, it's a ton of fun. I love listening to your podcast. It seems like you're always having fun and I'm drawn to that. So keep doing what you're doing.
Okay, we're back with Mark and it's time to wrap up the show with keeping tabs. Like most other people, I've got anywhere between 15 and 742 open tabs on my browser at any given time.
So this segment is where me, Josh, and our guest each share one of those tabs that we just can't close. Mark, since you're our guest, what are you keeping tabs on? So I've been keeping tabs on this thread by Hugh Bruyere. He's a technologist who works for DPT, which is, I think, like a creative agency. Like he basically dropped ink onto pans of milk. I know this is really exciting for this setup.
And it created these like very Rorschach looking like ink blots that are moving almost like a kaleidoscope. But then he filters that through stable diffusion. And so you see like flashes of like generated images that are very interesting. Like it looks like it could be a true detective sequence or something like that. So that's what I'm keeping tabs on.
What is stable diffusion? Stable diffusion is like a local generative AI model. So, you know, it generates images based upon prompts. And so somehow he like, he put those like
ink drop milk pans into Stable Diffusion. And then in real time, like you see these images appear. Is this now going to be the subject of the next True Detective? That's what I was going for. You can only hope. This is going to be the new subject of the next True Detective series. Josh, the Stable Genius of this pod, what are you keeping tabs on? So I'm keeping tabs. Speaking of Stable Genius, I'm
Has anyone seen how Terrence Howard has gotten really into math in the worst possible way?
Oh, isn't he one of those people who's like two plus two is five, can't tell me I'm crazy or something. So close to that. So Terrence Howard, so if everyone doesn't remember the actor's name, most famous for Empire, the Best Man series. He was in Iron Man, I think. And I recently served clips and now all of these tabs are off of him saying different things about math, where it started with him basically saying,
One times one equals two? Yes. And basically wrote a whole proof about it? This was a big thing. Yeah. Yeah.
When I was in college, this was covered on Gawker all the time for some reason. Terrence Howard specifically? Yes, Terrence Howard being weird about math. This is the thing. I'm just learning about this, but this is going back years. And he's been on like Rogan recently, dropped a new episode with him and some like former mathematician turned podcaster VC or something like that.
arguing about it but apparently like he wrote a whole proof to this he wrote about like re-examining and reframing the entire way we understand mathematics and sent it to neil to grass tyson who just ripped it apart and sent it back to him and another thing that just got thrown out there is he doesn't believe the theory of gravity he thinks it's the draft left by electricity
That's awesome. I love this man. Which is a whole nother thing. But the reason I'm obsessed with this is it's conspiracy theory, sort of like mentality around stuff. He's absolutely wrong in every provable way. But it seems like the most wholesome thing to be wrong about at this point because it's not actually affecting things. Now, hopefully it doesn't lead back to something where it's like,
something about like this math and one times one equals one because of some cabal or something like that, where it gets really dark again. But it's just really enjoyable. Like when conspiracy theories mostly consisted of aliens and Bigfoot, that this is where Terrence Howard is with mathematics. I'm imagining just like a bridge falling down though as you say this. Like, oh, it's just math. LAUGHTER
He's so confident about it, too. He's filed patents. Yaz, what are you keeping tabs on? Did you see that Wicked and Gladiator are coming out on the same day in the fall? Yeah. I thought that they moved the Wicked opening day so it didn't compete with Moana 2. Is it now competing with Gladiator? Yeah, it's now competing with Gladiator. Nice.
Two movies for the girls and gays. Yes, yes it is. Do we have an Elysian name like Barbenheimer? We need to come up with a good one. I think people have been trying to make it work, but it's a bit embarrassing. Gladiator? Gladi-kid? Glickid? I think Glickid is the right name. The photos of Gladiator look sick.
I haven't seen any of them yet. Oh, it's in the new Vanity Fair. I think Paul Mescal is probably kind of overrated. But then I watched After Sunny was good. And then I saw him on stage in London as Stanley Kowalski. He was incredible. Oh, that's cool. He's so hot in real life, too. Wait, so I want to go back to these movies coming out. So you're going to the movie on premiere date. Which one are you more likely to see, Gladiator or Wicked? I want both of you to answer this. Which one are you more likely to see or more likely to see first?
More likely to see if you can only choose one. Yeah, there's probably one in each eye. Like, I'm in between two theaters. I think I would see Gladiator first.
Because it's full of attractive men. I knew that was where the answer was coming to. Based on the Vanity Fair photos. I'm sold. Mark, what about you? I'm going to say Gladiator. I wasn't really big on the first one. I think Paul Muskell's fine, but you really sold him. No.
Just now, yes. So now I'm like going to reassess. I got to reassess my view. Yep. Let's end it here. Mark, thank you so much for coming on Most Innovative Companies. Thanks for having me. Our show is produced by Avery Miles and Blake Odom with editing by Julia Xu. Mix and sound design is by Nicholas Torres and our executive producer is Josh Christensen. Remember again to subscribe, rate, and review, and we'll see you next week.