SOCIAL
Images and Inspiration With AI: Pinterest’s Jeremy King

Topics
Artificial Intelligence and Business Strategy
The Artificial Intelligence and Business Strategy initiative explores the growing use of artificial intelligence in the business landscape. The exploration looks specifically at how AI is affecting the development and execution of strategy in organizations.
More in this series
Jeremy King leads a team of 1,400 passionate engineers working on the continuous improvement of Pinterest’s image-driven platform. With a background that includes heading up a translation team at eBay and overseeing the technology behind Walmart’s U.S. retail stores and e-commerce business, Jeremy is now responsible for technology operations at Pinterest. To support the company’s mission to inspire people to “create a life that they love,” he and his team rely on advanced AI, machine learning, and a graph database to index and build a network of images so that users can find inspiration — particularly when they aren’t completely sure what they’re looking for.
On this episode of the Me, Myself, and AI podcast, Jeremy joins hosts Sam Ransbotham and Shervin Khodabandeh to talk about some recent advances Pinterest has made in the image-recognition space and shares his views on how generative AI will transform image-based content like Pinterest’s.
Subscribe to Me, Myself, and AI on Apple Podcasts, Spotify, or Google Podcasts.
Transcript
Shervin Khodabandeh: How does an AI-based platform use generative AI to engage users? Find out on today’s episode.
Jeremy King: I’m Jeremy King from Pinterest, and you’re listening to Me, Myself, and AI.
Sam Ransbotham: Welcome to Me, Myself, and AI, a podcast on artificial intelligence in business. Each episode, we introduce you to someone innovating with AI. I’m Sam Ransbotham, professor of analytics at Boston College. I’m also the AI and business strategy guest editor at MIT Sloan Management Review.
Shervin Khodabandeh: And I’m Shervin Khodabandeh, senior partner with BCG and one of the leaders of our AI business. Together, MIT SMR and BCG have been researching and publishing on AI since 2017, interviewing hundreds of practitioners and surveying thousands of companies on what it takes to build and to deploy and scale AI capabilities and really transform the way organizations operate.
Sam Ransbotham: Welcome. Our guest today is Jeremy King, head of engineering at Pinterest. Jeremy, thanks for taking the time to talk with us.
Jeremy King: It’s great to be here. Thank you very much.
Sam Ransbotham: It’s always good to start with an overview; maybe tell us a bit about what Pinterest does and what you do at Pinterest.
Jeremy King: Pinterest is, we like to say, the destination on the internet for inspiration. We have hundreds of millions of people that come to us every day to figure out what they want to wear, what they want to do this afternoon, or what they want to make for dinner, or how to decorate their kid’s cake or remodel their kitchen. And so we want to build a platform that allows anyone to create a life that they love.
It’s a great place to work with that kind of mission, and I have a lot of people who, when I tell them where I work, get pretty elated about their own Pinterest boards and what’s going on in the world.
Sam Ransbotham: Great. Tell us more about working at Pinterest.
Jeremy King: I’ve been at Pinterest now, the head of engineering, for just over four years. And just like any leader, I’m trying to figure out how to unblock my team and hire great people. I tell people all the time that when I came to Pinterest, I was just so impressed with the level of talent. It’s a relatively small team. We have 4,500 employees in the world and about 1,400 engineers. And as a result, competing against some of the biggest companies in the world, this team has to be extremely high caliber. We have some of the best graph database people in the world; we have the best computer vision people in the world. These people are really fun to work with.
Sam Ransbotham: None of that, so far, mentioned artificial intelligence, so, what role does artificial intelligence play with those 1,400 engineers?
Jeremy King: Yeah, that’s an excellent point. One of the things that is true at Pinterest and now at many other companies is Pinterest was actually born with machine learning and AI from the base. It doesn’t exist without it. So talking about AI and ML separately from Pinterest is almost like a misnomer; it really doesn’t happen without it. Nearly everything that we do touches our machine learning systems. We do millions of inferences every second. Every single request comes all the way back to the graph and publishes back out a very specific use case to every single person.
It’s not cached. … Every person has a specific result. I mentioned that we have about 1,400 engineers. We have about 350 machine learning engineers. As a percentage, when I tell other CTOs … they’re like, “That’s a really high percentage of machine learning people compared to my company.” That’s true also on the data platform side; we have a wonderful leader, Dave Burgess, who runs our data platform. And you can imagine our data platform is about as important as any other part of our capabilities. And so I mentioned the graph, but it’s also, you know, the sort of normal SQL databases and the real-time systems that keep this thing running.
Shervin Khodabandeh: Jeremy, you said “graphs” a couple of times. What’s a graph?
Jeremy King: Essentially what a graph is, is if you take an entity, an object, and you say, “What’s related to that object?” — so, in this case, we’re using a nearest neighbor graph, which we’ve written lots of documents on, which is saying, “Hey, I have an image, in this case, and I have another image that’s related to this image.” So that allows me to build this sort of network of images. And the reason the graph works at Pinterest is because every pin is essentially added to a board, which essentially makes it a node in a graph, and so it continues to increase the graph.
So it allows you to … unlike a SQL database or a relational database, where you have to sort of tag indexes and indices, the graph can be indexed. Essentially any image could be indexed to any other image. It allows you to do searches on things that are related incredibly efficiently. That’s the power of Pinterest.
Shervin Khodabandeh: Jeremy, you mentioned AI has been part and parcel of Pinterest’s journey and has been ingrained in Pinterest’s tech stack from its very early inception, so it must be that generative AI is all over the map for you guys as well, and I’m sure it’s on everybody’s mind these days. So maybe tell us a bit about what you are doing there.
Jeremy King: We have a state-of-the-art machine learning environment that interconnects these data sets across all the surfaces. We call it the “home feed,” the search system, the related things that drive personalization and recommendations and engagement across Pinterest. [And] we’re growing really fast from monthly active users. … Gen Z is growing even the fastest. We added 13 million monthly active users, so we’re really focused on how to increase engagement.
And not surprisingly, a lot of the improvements in machine learning, even in the last two years, have been driving that engagement. We have, of course, built new features [that] allow Pinterest to become more shoppable. But as we’re seeing the capabilities of the machine learning models getting effectively 10 or a hundred times bigger, which is relatively common today, we’re seeing that give an outsize increase in our results as the result gets more specific.
And frankly, Pinterest is also about looking at [vastly] different things. So it’s not [that] every single item is about “What do you want to cook for dinner tonight?” It’s also “I know you were looking at birthday cakes last week” or “New Year’s is coming up,” and we’re going to interplay some of that. So we’ve been taking advantage of some of the advanced AI capabilities that have only come to life over the last couple years. And that includes GPU [graphics processing unit] work and all the things that make it cost-effective to do these things.
So tying back to your question of generative AI, not surprisingly, given that we’re an image platform, generative AI is very interesting to us, and I typically break it up into three buckets. When we’re thinking about large language models — LLMs — we talk about, No. 1, how do I make my team more productive? We’ve got a couple of pilots running. We haven’t decided exactly which way to go yet, but it’s looking really promising, and, like lots of other CTOs, we’re really excited about that.
Shervin Khodabandeh: That’s interesting. Tell us how generative AI increases efficiency.
Jeremy King: I was talking to a number of CTOs, and one thing I thought was really interesting is, this one particular CTO was saying that, in general, it’s increased productivity by 10% to 15%, but there is a small set of users where it’s increased their productivity by 50%.
Shervin Khodabandeh: That’s fantastic. I was just going to say, it also feels like, if you go back a decade ago, there were lots of images and unstructured text and that kind of stuff that, for AI to train on that data and make sense of it, human intervention was needed, both to tag the content but also to make sure that the outputs made sense. And it feels like now, with gen AI, that a lot of those humanlike judgment calls are going to be made more and more with gen AI, particularly with lots of unstructured text and images and video and that kind of content. Is that right, or am I going too far?
Jeremy King: I think it’s right. I mean, Pinterest’s system is built on embedding, so you take an image or a piece of text and you essentially tag it — essentially, to your point, this is what Pinterest has really been great at. You take our computer vision technology and effectively build these embeddings to detect people or content or couches or birthday cakes.
We’ve been really good at that and have really led the industry for a long time on this. And you’re not wrong that what happens is, it just gets better; it gets more accurate, more specific. In the old days, I could say, “Hey, I know this is a lamp, and I know it’s got pendants, and it’s made of crystal, and it’s gold,” and that sort of thing. But now I can say, “I know exactly how many tines it has on it; I know what kind of light bulbs they are; I know probably who manufactured it,” and these sorts of things. I can get way more specific.
And at Pinterest, we really think about these as two different things. People come to Pinterest because they don’t know exactly what they want. And this is where Pinterest thrives, and I think why we have a long-term differentiation in the market. Because if you know what you want, you can always go to Amazon or Home Depot or Wayfair to go buy it, but if you don’t know what you want, you start with Pinterest or do a million different searches because [you] don’t know how to describe, you know, “classic barnyard kitchen” — [you] don’t know those words.
So even putting it in something like ChatGPT, I don’t know how to say that — you know what I mean? I’m sure that’ll get better over time, but a lot of it will be image-based, too, as well.
Shervin Khodabandeh: Are you finding that the foundational models that are available do it for you, or you’re building your own domain-specific models?
Jeremy King: Both, I guess, is the answer. What we’re finding — and it’s still relatively early — in sort of the generative imagery, what we’re finding is that the smaller models actually are much more specific, and so I think that’s what’s going to happen. And my CTO friends are saying the same thing, which is, each one of these models is going to be very specific to the use case, and that helps it not only be more accurate but also makes it much, much cheaper to implement as well.
Sam Ransbotham: I think it’s interesting that you mentioned cakes several times, because decorating cakes is, I think, practically my only use of Pinterest. And it ties to Shervin’s comment about generative, because I deeply suspect that those cakes that I see there, no human could create. I certainly can’t create them. So to what degree are … people putting [up] images that don’t really exist or products that don’t exist? What’s Pinterest’s perspective on that?
Jeremy King: Excellent point. I think, yes, and we see this a lot [in] one of our biggest categories we call “art.” Art includes a whole bunch of things, including things that you would traditionally put on walls, and paintings, and that sort of thing. But it also includes things like body art, and we have tattoos and these sorts of things. And again, lots of those are great cases where people need a little bit of inspiration, and it can be generative, right? It may not exist in the world, but you want to see what it looks like on a human.
You can see this coming, where you’ll be able to print out something that’s generative. And this is already happening, right?
Sam Ransbotham: Yeah. That really aligns with the idea of inspiration.
Jeremy King: Yeah, absolutely. And things like cakes and that sort of thing are very interesting. One of the harder parts is, I’ve seen lots of work for home and home improvement, but what it’s doing is it’s generating a whole bunch of ideas that can’t be realized, where I can say, “Hey, people have yet to really break out the camera and take a picture of their room.” People do that, but we call it a 1% feature. Like, how do you get people to actually engage the phone or the phone camera?
But if you can upload an image — I’ve seen some beautiful renderings of, like, “Here’s what my room is. Rearrange the furniture in my room, the furniture I already have,” and things like that become really interesting. Or you say, “I want to replace this couch with this couch from Wayfair. I want to replace this lamp [with] some other lamp,” and here’s 20 different combinations, and you can kind of click through it.
Those are like you’re taking a virtual experience, almost AR-like [augmented reality-like], but you’re putting real products in there versus them being generated products. But it allows you to build user experiences that are much more enhancing.
Last year, we built AR Try On, which allows you to try on makeup, and what we’re seeing, again, while it doesn’t get a ton of usage, the people that do use it are 60% more likely to buy something. It’s amazingly engaging when people actually get there.
Sam Ransbotham: What’s tough at Pinterest? You mentioned a lot of good use cases, where you’re doing lots of things with these technologies. Certainly, [it] can’t all be wonderful. What’s hard?
Jeremy King: We spend a lot of time on inclusive search and results. In 2021, we launched hair-pattern search. It was kind of this first of technology: When you’re searching for hairstyles, how do you identify hair that looks like yours? This is hugely important in order to refine what you’re looking for.
We had these same kinds of problems with skin tone, where originally we were trying to use the early models for skin tone detection. And what we found is, it was more on face detection. And so our team had done some great work doing skin detection — it could be a side shot or a back shot or ear shot or a hand shot. You have no idea what kind of skin. So how do I detect what skin is? And then you can detect what kind of tone you’re looking for.
Same thing had to be applied to hair. And we’ve got all kinds of hair: We have shaved/bald, we have straight, we have wavy, we have curly, we have coily, we have protective, and these kinds of things make the results dramatically better.
You can save your hair pattern and your skin tone, and then we’ll tailor your results to that. And as you can imagine, this is a complex problem. It’s not just a USA thing; every different country has different types of hair patterns and skin tones and different kinds of fashion, and so these kinds of things are really hard, but they’re the great projects to work on.
And we’ve had a wonderful team. The advanced technology group here, in particular — we’ve been working on inclusive technology and all these advanced models, and they’re definitely hard, but they’re wonderful when you get them right.
Shervin Khodabandeh: We have a segment here: five rapid questions. I’ll just ask you a bunch of rapid-fire questions. Just tell me the first thing that comes to your mind. What’s your proudest AI moment?
Jeremy King: The inclusive product feature, and actually, hair pattern was my proudest one, because not only was I involved [with] the product from the beginning, but it’s also first in the industry. And the amount of response that we got from our pinners was incredible. So, yeah, it was a really good one.
Shervin Khodabandeh: What worries you about AI, aside from bias and some of the ethical issues?
Jeremy King: It’s funny. It’s on every single forum. It says, “Hey, are you worried about bias?” Let’s see. What worries me about AI? Hmm. Yeah, I’m not worried. I don’t worry about too many things, but that’s a CTO …
Shervin Khodabandeh: That’s good. That’s a good thing though, right?
Jeremy King: That’s a longtime CTO’s thing. If you worry about too much, you can’t survive in this job.
Shervin Khodabandeh: What’s your favorite activity that does not involve technology?
Jeremy King: Mountain biking.
Shervin Khodabandeh: Very cool.
Jeremy King: I try and go at least two times a week. I go with a crew of Silicon Valley tech [colleagues] on Friday mornings and then usually on the weekends with my brother and a few other people.
Shervin Khodabandeh: The first career you wanted: What did you want to be when you grew up?
Jeremy King: I think my 5-year-old baby book says fireman, but both my grandfather and my father were engineers, so I actually thought I would build houses, because I loved working with my hands outside. So that’s where I thought I was going to go — architecture or house-building.
Shervin Khodabandeh: What do you wish AI could do that it currently can’t?
Jeremy King: Teleportation: Get me through the security line faster! Maybe that’s coming. I’ve always been super excited about translation. I took over the translation team at eBay way back when, and there were some wonderful people that were working on this. And I’ve always thought about a universal translator as something that would just be so amazing. And it’s getting so close, but it still seems like so far away before there’ll be a consumer version of this device.
Shervin Khodabandeh: But why do you think that is? Because we’ve got LLMs now, and we’ve had, for a long time, audio to text. What’s the limiting factor here?
Jeremy King: I think it’s form factor; like, how do I get out my phone, hit “translate from Spanish to English” — you know, that kind of thing. I don’t know; if I’m walking up to a vendor in the street in San Francisco and I don’t speak the same language, and I want something. … It’s got to be close, but it’s a form factor thing. But I think it’ll come shortly.
Sam Ransbotham: Seems solvable.
Jeremy King: Yeah, solvable.
Sam Ransbotham: Jeremy, we really appreciate you taking the time to talk with us. It’s been very interesting to learn, sort of behind the scenes, what’s happening between all those images that we see but maybe didn’t appreciate all that goes into making something that looks as easy as Pinterest work. And it’s been pretty fascinating to see how much artificial intelligence and machine learning is behind that. Thanks for joining us.
Jeremy King: Of course. It’s been wonderful.
Shervin Khodabandeh: Thanks for listening today. On our next episode, Sam and I meet with Damini Satija and Matt Mahmoudi from Amnesty International. Please join us.
Allison Ryder: Thanks for listening to Me, Myself, and AI. We believe, like you, that the conversation about AI implementation doesn’t start and stop with this podcast. That’s why we’ve created a group on LinkedIn specifically for listeners like you. It’s called AI for Leaders, and if you join us, you can chat with show creators and hosts, ask your own questions, share your insights, and gain access to valuable resources about AI implementation from MIT SMR and BCG. You can access it by visiting mitsmr.com/AIforLeaders. We’ll put that link in the show notes, and we hope to see you there.
SOCIAL
Twitch to shut down in SKorea over ‘seriously’ high fees

The platform said it had tried to lower its costs by reducing the maximum video quality but it was still losing money – Copyright AFP Chris DELMAS
US-based streaming platform Twitch said Wednesday it would stop its service in South Korea in February because of “seriously high” network costs, dealing a blow to millions of users in one of the heartlands of e-sports.
The Amazon-owned company said in a statement signed by CEO Dan Clancy that costs were 10 times higher than most other countries, making it impossible to continue operating.
South Korea allows internet service providers to charge data-heavy companies like Twitch extra fees, which has already led to a long dispute with Netflix.
Big telecom firms in Europe have pushed for a similar deal, which they call “fair share”, but an EU consultation concluded in October that the idea was not popular.
Twitch said it had tried to lower its costs by reducing the maximum video quality but it was still losing money and would pull out of the country on February 27.
“The cost of running Twitch in South Korea is currently seriously high,” said the statement.
– ‘Stellar player’ –
Twitch, acquired by Amazon in 2014 for close to $1 billion, gained significant traction among gamers in South Korea.
The firm does not publish user numbers but it was widely reported in 2021 to have six million users in South Korea, more than four percent of its global total.
The country is known for its passionate, competitive, and dedicated gaming community, as well as its megastar Faker — a gamer hailed as the Michael Jordan of e-sports.
“We would like to reiterate that this was a very difficult decision, and one that all of us at Twitch are deeply saddened by,” the company’s Wednesday statement said.
“South Korea has always been a stellar player in the global e-sports community and will continue to do so.”
Shares in South Korean video streaming service Afreeca TV, Twitch’s competitor, soared almost 30 percent in afternoon trading in Seoul.
Some of the country’s Twitch users were devastated by the news.
One streamer, yummy_2 said: “It feels like losing my job right now.”
– Biden vs Trump –
Netflix was the first major international firm to cry foul over South Korea’s rules on network fees, getting entangled in lawsuits with SK Broadband, one of South Korea’s biggest internet service providers.
However, the two firms announced in September they would drop the legal cases and would now instead “collaborate as partners for the future”.
While the usage fees are a boon to telecom companies, they are bitterly opposed by tech platforms around the world.
European lawmakers and digital rights activists also argue such an arrangement could break rules on net neutrality, whereby telecoms firms are barred from selling faster internet speeds to particular companies.
The issue has been at the heart of a years-long dispute in the United States with former President Donald Trump rolling back net neutrality rules and his successor Joe Biden struggling to restore them.
SOCIAL
Baltimore food influencers do it for the ‘gram. Do restaurants benefit?
The Cake Out Maryland bakery in Columbia was a labor of love for sisters Sade and Azia Castro.
Between traveling nurse gigs, Sade Castro would take orders over social media for the sweets otherwise found only in the Philippines, advertising flavors from ube flan to chiffon cake with a milky caramel glaze. But few outside their community knew of the shop.
Castro saw foodies on Instagram in videos that garnered thousands of likes and followers. More people had to be searching for “Asian tastes” in Maryland and Virginia, she thought. Why couldn’t her cakes be the next viral sensation?
So she reached out to a food influencer.
Over the last few years in Baltimore, the practice of connecting restaurants and burgeoning food businesses with social media personalities has become increasingly common, according to public relations executive Dave Seel, who has built an arm of his Blue Fork marketing firm for the task.
“There can be a dearth of coverage for certain subsections of the city,” he said. “Influencers have taken up that space and used it to build followership.”
Baltimore is a small city, especially in food media. There is no Eater, Infatuation or Michelin Guide. People are thirsting for creative, diverse angles, Seel said.
With the rise of food influencers in Baltimore comes an opportunity to provide platforms to communities, voices and cuisines that have been traditionally alienated. But this wave of restaurant marketing has also raised questions about the authenticity of social media tastemakers and where the quest for that viral video leaves small businesses, many of whom are fighting for survival following the pandemic.
Marketing is an extension of community building, Seel said, and to that end, some restaurants have modified their aesthetics to attract new customers over social media.
Seel cited BLK Swan in Harbor East as a prime example for its well-advertised community events and “selfie walls.” Customers cannot visit Gunther and Company on Toone Street without taking photos by its “Instagram-worthy living green wall,” he said. At times, he has recommended that restaurants invest in a “particularly ooey, gooey picture-worthy” dish.
It does not always go viral or attract the attention needed to generate business, but it’s an increasingly popular strategy.
“Has it eclipsed all other strategies? I don’t necessarily think so. … But do [influencers] have a seat at the table? Absolutely,” Seels said. “You can’t ignore it.”
‘It’s a marketing job’
Tim “Chyno” Chin always dreamed of hosting his own television show about food.
He grew up an army brat, born in Germany and shuttled between bases before landing in Sandtown-Winchester, a Baltimore food desert. It was not “lavish,” he said; food was utilitarian and purchased with food stamps. There was no one like him on TV: Black, Chinese and gay. But as Chin remembers, he had a “charisma” that allowed him to persevere.
Chin now considers himself part of a “freshman class” of influencers who rose to foodie fame before the local restaurant industry came to embrace the world of social media marketing. Until about six years ago, eateries looking for publicity were beholden to legacy media platforms. The big players trusted to show Baltimoreans where to eat were radio personalities like Downtown Diane and Dara Cooks, he said.
“We slowly started replacing that,” Chin said. “They didn’t understand [social media] was going to catch on the way it did.”
Chin had worked in kitchens and as a server, so he believed he could relay the importance of a social media presence to the old guard of small businesses. He started by running the social media of the former Joe Squared in Power Plant Live, and then shooting food pictures at the now-shuttered Pinch Dumplings in Mount Vernon Marketplace in exchange for free meals.
“I would post something and then a restaurant would sell out of it,” he said, calling it “the Chyno effect” — a byproduct of his time hosting a YouTube show. He’s now garnered followers as “The Baltimore Foodie” and “The Boy with the Blue Beard,” building a more-than-135,000-person Instagram following and appearing as a host for the “Fresh, Fried and Crispy” show on Netflix.
“I’ve got an Emmy waiting for me somewhere,” he told The Banner.
To show restaurants he was serious, Chin drew up a rate sheet for his services. “A lot of influencers have it,” he said. The sheet explains an influencer’s cost per post, Instagram reel, links, video and stories. “Everything has a monetized value.” Chyno did not say how much he charges, but as his audience across platforms rises, so does his value.
“People don’t understand this is hard,” he said. “You have to constantly evolve with technology, learn algorithms, follow these trends. … It’s a marketing job.”
TikTok celebrities like MMA fighter turned foodie Keith Lee, who recently made news for a video critiquing the service at an Atlanta restaurant, can change an eatery’s reputation with a single post.
Anybody can call themselves an influencer, Seel said, but “it doesn’t mean they have a core following or an engaged following that really creates the marketing effect that can get restaurateurs that return on investment.”
‘It’s such a difficult thing to decide what you’re worth’

The world of social media marketing is still largely uncharted. The Federal Trade Commission has codified guidelines on sponsorship transparency for influencers, going as far as to issue $50,000 penalties for failures to adequately disclose paid partnerships.
According to estimates from Insider Intelligence, more than $6.1 billion is expected to be spent nationwide this year on influencer marketing.
Local influencer Rachel Lipton learned about rate sheets herself in 2017 when 7-Eleven offered her $100 to post its iced tea on her “like the tea eats” Instagram page.
“My wife pulled me aside and said, ‘I think you should be charging these large businesses,’” said Lipton, who already had a full-time job. “It’s such a difficult thing to decide what you’re worth.”
Her pricing varies. Video content took far longer to edit, so she charged more. Her rates also went up depending on the size of the company inquiring about a post. She also is particular about who she will work with — or not. She said she will never post about Chick-fil-A due to their alleged culture of homophobia. And since news broke in 2020 of Ouzo Bay allegedly discriminating against a Black woman and her son, along with follow-up complaints against the owners, Atlas Restaurant Group, Lipton has promised not to promote dining at their restaurants.
Kimberly Kong, the creator of a series of food photography pages known as Nomtastic Baltimore and Nomtastic D.C., has amassed more than 100,000 followers, in part, for making a point of dining at Asian-inspired small businesses in Maryland and Virginia.
“I let [businesses] know that you’re only going to get featured if I genuinely like your food. And it’s going to be disclosed that I was invited and food was comped,” Kong said. Yet she cringes at the “influencer” title and the lack of authenticity it evokes. A large number of her posts were not paid for, she said, and were born out of an interest in wanting to try new spots.
Kong also does not charge small businesses for promotion, citing pandemic-era losses as a reason for many of them to be skeptical of investing in the world of social media marketing. Chin and Lipton also said they offered reduced rates to try and boost local spots.
“I understand the restaurants’ point of view with how slim the margins are and how tough it is right now,” Kong said.
‘Every time we posted something, it just got sold’
Sade Castro never met the Instagram celebrity that sparked an interest in her Maryland shop.
Neither did her sister and business partner, who repeatedly called Castro “crazy” for inviting someone with more than 100,000 followers to sample their cakes. For three years, the two-person bakery had sold the desserts almost exclusively to a group of Filipino moms over Facebook — and even then, they struggled to meet demand.
“I trust that you really believe in your food recs and that you’ve actually tried and loved every food post,” Castro wrote to Kong on Sept. 9. “With that, I would like you to try our Filipino-style cakes.”
Shortly after, Castro was leaving a sampler of nine cakes at Kong’s door.
On Sept. 21, Kong posted footage of her digging into a gooey can of chocolate cake and slowly slicing into the ube flan’s purple center.
“I was at work when my phone started to go off,” Castro said. Within a day, the video had gone viral. The number of people viewing the bakery’s Instagram page rose by over 900% in a matter of hours, and then again by another 2,000% by the end of the week. About 3,000 new people had followed their rarely updated Instagram by the end of September.
“Why would you do this?” Castro remembered her sister asking. “It’s just the two of us, we’re baking from home, and we have full-time jobs.”
The bakery that had provided roughly 120 cakes each year catering to their Filipino neighbors had received hundreds of orders in a matter of days. “We were messaging people saying we don’t have [the cakes],” she said.
Unable to meet demand, they started a lottery. By the end of October, the attention faded some, with viewers of their content down by 42%, according to Castro’s Instagram analytics. Still, the success of the post presented an opportunity for Castro’s self-proclaimed “side hustle.” “Every time we posted something, it just got sold,” she said.
But a restaurant has to be ready. Seel explained that influencers will often receive a tailored experience: sampler cakes, private dining and even custom sandwiches. The business has to be able to execute at the same level for the regular customers, too.
In October, Fells Point eatery Little Donna’s claimed to be “screwed” after a New York Times critic placed the business on the paper’s list of most exciting places to eat. The now-shuttered Local Oyster also faltered in the spotlight after an influencer-promoted sandwich spurred high demand and community backlash, forcing it to be 86ed from the menu.
“All of a sudden, there can be an onslaught of people and it’s hard to keep up,” Seel said.
Castro has no regrets about Kong’s effect on her business. As of November, Cake Out is searching for ways to increase output and serve the Filipino neighbors who had leaned on them for their traditional holiday treats. Plans to move to a larger kitchen are in the works, due to the support from new customers, Castro said.
“For now, we are grateful.”
SOCIAL
The Age of Virtual Influencers is Coming, Which Will Bring a Range of New Considerations for Brands

While the current spate of generative AI tools are interesting, and are already changing discovery behaviors and interactive processes, they’re really only scratching the surface of what’s possible, and are far from actual “intelligence” as the AI name suggests.
Indeed, most of these initial models are data matching tools, able to predict elements of text and images based on the most likely sequencing, by applying probability to huge datasets. And they’re becoming increasingly at doing just that, but they’re not “thinking” as such, these systems are not developing new concepts all on their own, and there’s no intent or reasoning behind those matches, other than mathematical likelihoods.
That’s the next level of AI, which many experts have expressed concern about, in that such systems will one day have the capacity to think independently, and potentially exceed our own mental capacity as a result. Though creating a digital “brain” as it were is still a long way from being a reality.
But even so, through probability alone, we’re also just touching on the expanded possibilities of generative systems, with the latest advances now pointing to a whole new phase of digital creation, which could cut many humans out of the process.
Last week, a Spanish ad agency made headlines after it revealed that it had created an AI character, which is now earning $US10,000 per month from brand contracts.
Aitana requires no payment, has no qualms about what she promotes, and is available 24/7. And she looks real, and no doubt many of her 200,000 Instagram followers were not aware that she doesn’t, in fact, exist.
You can see the appeal of virtual influencers in this respect, and Aitana is not the first to build a huge following, and certainly won’t be the last.
Even before the arrival of Dall-E and Midjourney, virtual models were already gaining traction, including characters like lilmiquela (2.7m IG followers), noonoouri (424k followers), and Shudu (241k).
More advanced creation tools are now making these virtual identities even more life-like, while the next phase of digital animation could take them to another level of realism, in replicating human trends.
This video, shared as part of Alibaba’s “Animate Anyone” project highlights how advanced image recognition and video sequencing can now replicate actual human movement, to an increasingly realistic degree.
It’s still not perfect, but again, we’re really only at the start of this process, and you can see how, as these systems continue to evolve, virtual influencers, in both still and video form, are set to become much bigger elements of online interaction.
Deepfake characters, where celebrity faces are superimposed over real actors, are another aspect, and another vector for security concerns, but fully virtual creations, animated from still images, would be cheaper to use, faster to customize, and easier for any brand to create, based on templated actions, animations, and movements.
And they are coming. Every platform is already rolling out AI labelling requirements to get ahead of this, but realistically, it’s going to become increasingly difficult to know whether you’re looking at a real person, or an AI creation, with the blurred details and distortions rapidly being ironed out by newer processes and systems.
Sure, right now, it’s easy to spot those AI-generated promotional photos showing up within your Facebook feed ads. But refinements are steadily bringing these tools closer to reality, and ironically, taking consumers further from it at the same time.
So what does that mean for your marketing efforts?
Well, if you’re camera shy, and have reservations about making video content, soon, you might not have to, with viable alternative options to create digitally animated content. You’ll have to disclose such, but realistically, it’s the concept that will resonate with viewers, not the composition, and if you can avoid the tell-tale markers of current generative AI imaginings, it could be an avenue for your future development.
Though it could also be bad news for human influencers, who are just now having their moment in the sun, as more brands come to realize their worth in reaching certain audiences.
The “Creator Economy” in this context could be set for a rapid recession, as even short-form videos become increasingly AI-simulated, sparking all new trends in promotions, with brands happily welcoming the cost savings.
I do think that human creativity will remain an essential, and that no matter how realistic your AI creations are, you’ll still need human-centered emotion at the core of any promotion.
And until machines can actually think like us, that will remain the key differentiator, though the actual process of expressing your message looks set to change significantly.
This will be a key trend to keep tabs on in the new year.
-
SOCIAL7 days ago
Musk regrets controversial post but won’t bow to advertiser ‘blackmail’
-
SOCIAL6 days ago
Is this X’s (formerly Twitter) final goodbye to big advertisers? It looks like it
-
SEO6 days ago
A Year Of AI Developments From OpenAI
-
SEO5 days ago
GPT Store Set To Launch In 2024 After ‘Unexpected’ Delays
-
SEARCHENGINES5 days ago
Google Core Update Done Followed By Intense Search Volatility, New Structured Data, Google Ads Head Steps Down & 20 Years Covering Search
-
PPC7 days ago
5 Quick & Easy Ways to Get More Referral Traffic (+Examples)
-
MARKETING7 days ago
Take back your ROI by owning your data
-
TECHNOLOGY6 days ago
Next-gen chips, Amazon Q, and speedy S3