Hikari Senju is a computer science graduate from Harvard University with a passion for technology and design.
He founded his first startup, QuickHelp, an on-demand tutoring app, which he later sold to Yup.com and became their Head of Growth.
During his time at Yup.com, Senju observed inefficiencies in content spending and the potential of AI in advertising, which inspired him to establish Omneky in 2018.
Senju’s background in computer science, entrepreneurship, and his experience with AI technologies led him to create Omneky, an AI platform that generates, analyzes, and optimizes personalized ad creatives at scale.
What you will learn
Discover how to leverage AI for generating and optimizing personalized ad creatives efficiently at scale.
Learn to identify inefficiencies in advertising spending and utilize AI to address these challenges effectively.
Understand the influence of Hikari’s background in technology and art on his entrepreneurial journey and innovative solutions.
Explore the strategies for building and leading a high-performance team with a focus on accountability and effective management.
Gain insight into how AI can enhance human creativity, fostering a symbiotic relationship between technology and artistic expression.
Transcript
Jeff Bullas
00:00:04 – 00:00:32
Hi, everyone and welcome to the Jeff Bullas Show! Today I have with me Hikari Senju. Before we have a chat, Hikari is a computer science graduate from Harvard University with a passion for technology and design. He founded his first startup, QuickHelp, an on-demand tutoring app, which he later sold to Yup.com and became their Head of Growth.
Jeff Bullas
00:00:33 – 00:01:03
During his time at Yup.com, Senju observed inefficiencies in content spending and the potential of AI in advertising, which inspired him to establish Omneky in 2018. Senju’s background in computer science, entrepreneurship, and his experience with AI technologies led him to create Omneky, an AI platform that generates, analyzes, and optimizes personalized ad creatives at scale.
Welcome to the show. It’s great to have you here dialing in from San Francisco.
Hikari Senju
00:01:04 – 00:01:07
Thank you, Jeff, excited to be here and excited for our conversation.
Jeff Bullas
00:01:08 – 00:01:29
So, Aari you ended up going, doing a computer science degree at Harvard University, which is one of the elite universities in the world. So, what led to that? In other words, not only the inspiration maybe to do computer science, but where did this all start this interest in science and, you know, in computing?
Hikari Senju
00:01:30 – 00:02:16
Sure. Absolutely. So, uh I, you know, I’m actually the first generation to go to an Ivy League for my family. Um My, uh you know, I grew up in Westchester in New York. Um, my dad’s a painter and my grandfather worked at IBM. And so, you know, I kind of grew up in, you know, but, but an immigrant, immigrant family, a first generation immigrant family. And so, um you know, I grew up in this intersection of, of kind of that immigrant mindset hustle but technology, um the, the, you know, like early access computers as a kid out of competing robotics competitions. Um And I, and also I had a deep uh I, I really like Payday because you spend the weekends at my dad’s art studio and being alongside him. And so, um I, I was, you know, a creative, I like creating things,
Hikari Senju
00:02:16 – 00:02:42
painting things like making things. And so I um was very lucky to get into Harvard, uh you know, and um you know, obviously studied very hard and did a lot of the, you know, one, you know, national awards and things like this and, and accomplished stuff. Uh that helped me, helped me get, you know, helped me, I think I was one of two kids from my high school to get to Harvard that year. And, and, um, uh, and, and, uh, and it, with a focus on computer science and it was,
Hikari Senju
00:02:43 – 00:03:09
uh, it was also, I’d also actually gone to MIT. Um, and so it was actually, I was cross registered at MIT, taking A I courses there when, um, when I saw the early generative A I models, I kind of a precursor to the generative adversarial networks. Um And that got me very excited about ja I, because I saw A I being generating paintings, uh you know, generative art, some of some of the very early precursors, generative art and um
Hikari Senju
00:03:10 – 00:03:47
about a decade ago and being the sort of artist being passionate about technology, uh and also kind of having an intuition of Moore’s law to some extent of like, because having screwing up around it, you kind of drop in intuition, the exponential rate in which technology improves felt that, you know, this technology was, was, was advancing very quickly and that it would probably surpass abilities within a couple of years. And so, uh within, within that decade, it was kind of my, my, my sense is I was kind of playing around with it, very excited. Um And that’s actually how I got hooked into general of A I was, was, was, was through gene of art and that part due to my, my up bring uh in New York,
Jeff Bullas
00:03:48 – 00:04:00
right? So, um so really, you had a creative bent initially rather than a scientific bent. You saw what? But you saw the intersection of A I and creativity and art does that, that was the motivator.
Hikari Senju
00:04:01 – 00:04:34
Yeah, I mean, I think it was, it was, you know, I had always been viewed as something that was creative, that robots will never be creative. Uh But what this showed me was that I can actually be very creative. Um And, and um that was, you know, I think it was very counter, that was very, it was very con at that time. Um that, you know, the, the view was that we’d all become artists and then the A I will do the non artistic work and humans will just, you know, automate everything but being artists essentially. Uh but it seems like in some ways uh there were,
Hikari Senju
00:04:34 – 00:05:14
is happening in which case I am actually uh becoming the artist. Uh And so, uh and, and so that, that was, I think the very interesting kind of surprising thing uh that I realized in college. It got me very excited about space. And so around that time, I started playing around with open source models. I mean, you know, you have some big open source companies around that time or a couple of years later. Uh and um and so uh all this technology started becoming available to regular developers and people who are just passionate about technology. I was uh you can download these embeddings from online and, and, and, and, and train them yourself. And, and so um that was, yeah, it comes from that.
Jeff Bullas
00:05:15 – 00:05:40
So uh what year was that? Because essentially public consciousness of A I generated art is really unimpaired since Chat G BT in a, in a large sense. Um In other words, it got a, I got given a friendly face called Chat G BT, which is the chat bots. And um so what, what year was this that you went to Stanford was this interest in the intersection of A I and creativity and art?
Hikari Senju
00:05:41 – 00:06:03
Yeah. So I was at Harvard from 2011 to 2015. And so, and so the classes I was taking in machine learning, uh 2014, 2015 at Harvard and, and particularly at MIT um uh was where I saw some of the early generative generative artwork. And at that time, it was kind of hallucinogenic. It was, you know, very trippy. Um
Hikari Senju
00:06:04 – 00:06:44
uh It was kind of, you know, it was very, it was very pixelated like, you know, but, but, but I still, I saw the, you know, like the, the, the phd student was like the T A of this class literally entering a prompt being like, you know, I think the prompt was like elephant in the sky or elephant flying through the sky and it would generate like, you know, a smudge of gray, like a blue background with like what appears to be wings and, uh, and, but, but like that was enough, that was enough. I was like, wow, if you can do that in 10 years, you know, and I think it’s played out, it’s going to be able to do anything. And I think now, you know, that that’s, that’s kind of proven out to be true.
Jeff Bullas
00:06:45 – 00:07:02
So you, you’re at Harvard, you’re seeing a, I start to develop and unfold in front of you and you then went out and I don’t know. So with quick help, which is a non demand tutoring app. Uh did that, did you do that after you left Harvard?
Hikari Senju
00:07:03 – 00:07:08
I did that while I was a student at Harvard. Um And, and um
Jeff Bullas
00:07:08 – 00:07:09
What led to that? You
Hikari Senju
00:07:09 – 00:07:43
know, the, yeah, I mean, partly I think I’d always been and always been entrepreneurial. I liked starting projects and building things. As I said, I was growing up, I loved painting and I loved making things. I love making websites. Uh I mean, I think if I, you know, around when I was a teenager, I made a website to recommend clothing. They use like an early precursor, recommended uh recommender network of recommender systems uh where we would tag all the uh I just scraped all the websites from um from Ralph Lauren and I would tag it with a, I then recommend similar clothes based on you, you kind of selecting image. And so um
Hikari Senju
00:07:43 – 00:07:59
uh it is, I’ve always been passionate about making things. I, like, I’m a maker. I like making things. I’m a, I, I, you know, I grew up in a, you know, in a family of crafts people to some extent. And so, um so I always was making these projects and, uh, and one of the projects I made in college was uh was this uh
Hikari Senju
00:07:59 – 00:08:24
app to uh connect students with tutors. Uh And the reason the impetus behind that was, you know, one of my roommates was um doing something called dorm, which is when, you know, students to make, uh to make extra money, they, they kind of, you know, uh do, do cleaning uh for, for the dorms and um, very smart person, uh very, very and, and like, it was kind of, I felt like an missed opportunity because this person could probably make a lot more money tutoring.
Hikari Senju
00:08:24 – 00:08:54
And at the same time, I had some friends that are also phd students and, uh, and they’re, you know, making minimum wage. Um but the minimum wage sometimes even uh in, in their, in their departments. And so, um I really saw an efficiency in the market uh uh of the, of this resource fundamentally being underutilized and felt that an app that could be a social and local and mobile. So it was a mobile app to connect students to tutors nearby or other students who could be tutors uh in, in, in, in a,
Hikari Senju
00:08:55 – 00:09:30
you could see how many mutual friends you had with them. And so there was some element of trust or you could see that you went to the same school, what school they went to, et cetera uh to then meet it on demand to help with tutoring. And so that was the impetus starting, you know, it was kind of a side project. I um and uh and it became, you know, I think I built it in a hackathon maybe even at Berkeley. Um because I was also doing a lot of hackathons then. Um you know, and um uh yeah, and they kind of, it’s got a lot of users initially, I was kind of shocked to see that there were a lot of users out, you know, maybe this would be interesting and then got into some incubator
Hikari Senju
00:09:31 – 00:09:36
uh in the summer. Uh And so that was kind of my first company,
Jeff Bullas
00:09:36 – 00:09:49
right? And you sold that? Ok, onto uh y.com. So, um I had a quick search on that and I can’t seem to see yup.com still in existence. Are they still around?
Hikari Senju
00:09:50 – 00:10:13
Yeah. So yup.com was acquired by another tech company. Um It was found. Yeah. Yeah, it was founded by Naguib Z who’s a really great friend of mine. Uh Now and, and um, and yeah, I mean, he, he, he started yup.com essentially, uh you know, to help with math tutoring at Stanford. Uh and um I really liked him and I, I’m a big fan of his and,
Hikari Senju
00:10:13 – 00:10:40
uh, and he, you know, said, hey, we’re working on something similar. Do you want to work, you know, do you want to work together on this, uh and join forces? And we had a lot of tutors on our platform. And, and so, um and so that’s kind of um uh and that’s kind of how we work together uh for uh you know, after, after uh I sold my company to him and I learned a lot, I learned a lot from M um uh and it was very, you know, like at that time, I had no idea how to run a company and how to manage and how to
Hikari Senju
00:10:41 – 00:11:15
uh do many of the things I think an entrepreneur needs to be able to do and, and I learned a lot and uh it helped then set up uh for, for Niki and also particularly, you know, also seeing the particular problem that is uh when I was running marketing at yup.com, I saw how much, how many challenges there are in terms of running effective content that advertising is a essentially like a trillion dollar industry. But half that half the budget, advertising budget is misspent on suboptimal apps, you know, even as recently as a couple of, you know, even as recently as like a couple of years ago and, and I think even still today. And so, um
Hikari Senju
00:11:16 – 00:11:53
that’s a lot of money, half, half of the money being wasted uh is, is, is a lot of money being wasted. And so, um that combined with my passion and of art kind of saw, put the two together to be like, why don’t we just quickly rapidly generate A B tests with A I and then also use computer vision to quantify design because it was a really great tool for analytics of design um at that time. And uh and, and, and then, and kind of close the loop. So if A I can in real time get insights about what types of contents, delivering clicks and delivering cells and then generate new content uh based on that.
Jeff Bullas
00:11:54 – 00:12:10
Cool. So, right. So that led you, I suppose after doing that into uh how can you create content efficiently and quickly using A I and, and, and technology? So when did omni key start? Is that how we pronounce it too, by the way, is that correct? It’s
Hikari Senju
00:12:11 – 00:12:12
omni omni
Jeff Bullas
00:12:12 – 00:12:25
omni key? OK. I was saying correctly pretty well, very, very close. But anyway, so omni key. So you saw this problem that you went out to solve an omni key I think started in about 2018. Is that correct?
Hikari Senju
00:12:25 – 00:12:27
It was started in 2018. Yes.
Jeff Bullas
00:12:27 – 00:13:02
OK. So essentially what have you done with it then? Because you can create the content, then we’re talking about. So how does A I come into the picture across omni you want to create content for marketers to actually create the content and then basically put it out onto platforms to actually use as advertising. So you’re doing video, you’re doing images, you’re doing writing. How does, so tell us a little bit about how a user would use it? Omni K? Those words, right? Let’s take us, take us through the steps.
Hikari Senju
00:13:03 – 00:13:33
A customer signs up, they connect the brand assets and brand guidelines. Step number one. So I learn what is unique about that brand, what are they, what are the products? Um It fine tunes on the customer’s product photos on their language use um And on their visual aesthetic. Yep. Um two customers integrate their ad networks, that platform. So meta Google, linkedin, tiktok, snap, et cetera, et cetera, reddit. Um
Hikari Senju
00:13:33 – 00:14:00
And so we can get real time data about what types of uh content is driving clicks and ultimately driving sales. Um Then we will uh you know, customers can sign up, they can see a dashboard with all the creative insights, uh predictive insights. Um So, you know, we believe that, you know, an ad with this feature will drive 20% increase in, in Lyft or uh you know, 2 200% increase in click through rates or uh you know, um
Hikari Senju
00:14:01 – 00:14:29
uh AAA certain percentage and increase the return of ad spend. Um you know, the creative insights will be married because what an A I can tag now when it comes to creative is, is pretty broad that you can tag language, the use of the language radical structure. Uh the persuasion tactics, the tone of voice, the elements of the within the creative using computer visions, all every individual element, the background color of the font, uh you know, is it male or female as a narrator, the voice, what is the, what is the action of the voice, et cetera? And so
Hikari Senju
00:14:29 – 00:15:08
we, we turn every creative into its into like a very uh in depth features set of, of, of, of uh of elements. And we find correlation between those elements and say a particular uh objective metrics. So we get to click the rates or uh to an announcement that the actual product that actually was our first product we launched in 2018 because the generation at that time wasn’t that great, but computer vision had advanced quite significantly by that point. And that was actually also the impetus of why I started the company in 2018, was that computer vision was pretty good. Um And, and so you can get insights about what types of creative content is driving results. Yep.
Hikari Senju
00:15:09 – 00:15:45
So you can, so customers can see all these insights and then customers that we then generate creative briefs for the customer with uh you know, suggested target audiences, suggested platforms. Uh The advertising platforms would suggest they should run more ads on or, or run ads on um uh concepts for creative imagery um uh templates uh as well. We generally recommend templates. Uh And then customers can then approve uh create a brief uh generate ads based on the creative brief and then, and then ultimately launch the ads across various platforms.
Jeff Bullas
00:15:45 – 00:16:05
OK. So essentially you’re learning the customer, in other words, brand voice, uh brand images. Uh So how does the model work? Is it a subscription? Is it a project? Is it done for your service? A lot of questions there? But is it something that you get a monthly subscription to do themselves? Tell us a bit about that.
Hikari Senju
00:16:06 – 00:16:25
It’s usage based, it’s consumption based pricing. So the more you use the product, uh you know, it’s based on the number of creatives that you’re generating uh and the amount of products, photos that you’re finding the models on. Um uh and uh and so the more you use the product, the more you charge and it’s, and it’s dynamic and flexible. Um
Jeff Bullas
00:16:25 – 00:16:30
Right. So do they, does the client actually do it themselves?
Hikari Senju
00:16:30 – 00:16:47
Yeah, the client can use, you know, we have a free trial on our site today. You can sign up and start using the tool yourself today. Um There’s also if you want us to do, you know, additional services on top of that, there’s that option as well. Um But because it is a self serve tool that customers can set up and use today,
Jeff Bullas
00:16:47 – 00:16:57
right. And what’s your target market? Is it, uh, basically the small to medium business more than enterprise? Is that, what’s your target? Your customer?
Hikari Senju
00:16:58 – 00:17:24
Yeah, I, I think our ideal customer is that, um, you know, S and B medium size business has kind of been our bread and butter. Uh for basically the life of the company, we do have a couple of great enterprise customers um in, in, you know, definitely the forward thinking and forward leaning enterprise customers are, you know, trialing our product and using the pro our product themselves. Um Some of the, the, the, the, yeah,
Hikari Senju
00:17:24 – 00:17:47
like some of the major consumer brands are, are trialing our product. But uh you know, there’s a little bit more of a hesitancy when it comes to enterprise and the use of generative A I due to IP concerns or uh copyright concerns uh data, things like this. And so they’re a bit more exploratory. But yeah, most of the usage uh and revenue is coming from uh small and medium sized business today for us.
Jeff Bullas
00:17:47 – 00:18:05
All right, cool. So let’s go back to the role of A I in the process. So A I can do a range of things, help you with the creation of the content obviously. And then of course, then it’s trying to see the pattern in the noise of the analytics and what works and what doesn’t? Are they the two major areas you’re using A I for?
Hikari Senju
00:18:06 – 00:18:09
Yeah, those are the 20 insights and generations.
Jeff Bullas
00:18:09 – 00:18:37
Yeah. OK. And, are you seeing any trends in the marketplace in terms of platforms to use or types of media to use such as video? Now a major digital advertising piece of content as opposed to still images. Um What’s the general, are you seeing any trends emerging in terms of how people are advertising on digital platforms? And you’re talking a lot about social media platforms, aren’t you?
Hikari Senju
00:18:38 – 00:19:18
Yeah, so video has been a big focus for social media for, I would say, you know, for some time I would and so that continues to be a big uh medium for creating and launching content uh advertising content, it’s very engaging form of content on social media, but also video applies to a whole area outside of uh social as well, which is streaming. Um And you know, the rise of streaming platforms and ad supported streaming platforms. Um Netflix invented the subscription streaming model, but really now the space is becoming increasingly more competitive with ads driven streaming platform models like, you know, there’s Disney is ad supported streaming and, and uh you know, yeah, less goes on FB to B et cetera. And so um
Hikari Senju
00:19:20 – 00:19:26
uh and so that, that uh you know, so you need video ads and video content for that, those platforms as well. Um
Jeff Bullas
00:19:27 – 00:20:09
Yeah. So in terms of the, so, so basically A I to help you create the content, then A I to analyze the effectiveness of the advertising. So tell me how that works and, and, and I’ve got another question about that. It’s about how your platform must start to become a bit sticky because it actually starts to learn about the customer. The customer then starts to rely on it, but that’s the next question. So how is the A I analytics um work on the platform and how, and do you have an example as a great um I suppose results from providing an uplift in, you know, click throughs or basically generating sales leads, whatever, tell us about the analytic. Yep.
Hikari Senju
00:20:10 – 00:20:52
Yeah, absolutely. Absolutely. So we have, for example, a company called Oana, which is a vegan cruelty free makeup brand. And uh for them, we delivered two X increases year over year cells. Um you know, phrases like cruelty free. Uh you know, it was correlated with, with driving predictive um engagement. Um uh We have uh we have a customer called Thunder uh that we saw a 50% decrease in, in the cost per install. And this is a mobile app based out of Egypt uh that does, uh you know, gives Egyptian consumers access to the American markets. Um We’ve had uh you know, a company called Ya, which is a consumer
Hikari Senju
00:20:52 – 00:21:37
banking. Uh, sorry, it’s a consumer savings account platform. Um, it’s like kind of gamify savings um, uh through, through their kind of weekly, um, kind of, uh, like random lottery e uh uh feature and, um, you know, we, we, we deliver them, uh uh uh uh a 50% decrease in cost per leads there through television ads as well. And so, um, you know, those are some examples of customers that have seen real results with izing our platform and tool. Um And again, um it, it’s really based on um you know, really analyzing the individual characteristics of each creative and, and what’s driving left and you can read more about it on our website. Uh And, and we have a large list of case studies uh on our site.
Jeff Bullas
00:21:37 – 00:22:06
So the customer signs up the sort of machine that learns about them, they’re advertising their brand voice and so on and creating content. So it sounds to me like it could be quite sticky. In other words, once you sort of go in there and teach it because A I is helping you create the type of content that works for their brand. So, and then of course, there’s the building of assets and so on. So do you see uh so does that reduce a lot of churn like, do you have much churn on the platform in terms of people coming and going?
Hikari Senju
00:22:08 – 00:22:42
Yeah, I mean, I think um you know, there is, there is like a virtuous cycle in the tool where you uh learn based on the performance, but also the customers and putting their own information about the brand. And so, you know, core part and like if this is really the third pillow of, of the of the product that isn’t really a feature but kind of underpaid everything is brand management. And I think that, you know, in the future brand management, it’s gonna become more important as competition for attention becomes fiercer, you know, the brands that are trusted are gonna elevate. Uh and uh and, and, and just be able to compete in the noise because they um um they just exude more trust. And so,
Hikari Senju
00:22:43 – 00:23:25
um uh and so as customers approve the ads and they edit the ads and they launch the ads, um then, uh then, then our system learns what makes our brand unique and that is, you know, you know, work and investment information that they’re investing into our tool on top of all the real time data we’re getting from the various ad platforms. Um And so, yeah, regarding, you know, the, the, the, the, the, you know, loyalty of our customers, our customers, I think are pretty on par with retention of businesses that sell to S and BS. Um you know, and you can take a look at, for example, Hubspot’s retention numbers or certain cohorts of customers and I think it’s very much in line with that. Um, and, uh, but, you know, we’re always working on making it better.
Jeff Bullas
00:23:25 – 00:23:45
Right. Right. Cool. So, in terms of, uh, the other thing that a lot of people would be interested in is OK, how did you start it? Was it self funded? Did you go and do a, you know, approach, you know, VC companies and say we’ve got a great idea, fund us. How did, how did, how did you actually fund Niki?
Hikari Senju
00:23:45 – 00:24:14
Yeah. So it’s really interesting. Right? I mean, I’m, I’m a solo founder. Uh, you know, I raised 13 million in seeds, you know, as, uh, and, uh, and, uh, it’s, you know, but, you know, it was, you know, first two years of bootstraps. I mean, nobody, yeah, I mean, nobody would. Uh, Adtech has never really been particularly sexy, uh, uh, especially recently and then A I back, gosh, you say, you know, you talk about a I in 2018, 2019, 2020 really? Anytime before chat he gets launched and,
Hikari Senju
00:24:14 – 00:24:39
You know, everyone thought it was vaporware. So, um, yeah, I mean, I think, I think, uh, you know, it was, yeah, I mean, I tried raising money, uh, for the first couple of years we were very lucky. We had some really supportive angel investors early on, not enough to really, um, be able to hire people that much. But, but, you know, Village Global Richard Soldier, the chief scientist that sells force. Um um
Hikari Senju
00:24:40 – 00:25:19
and others were some of the earliest believers in the company. Uh and, and uh and uh we’re very lucky to be able to work with them in the very early stages of the company. Uh just to keep the company alive uh to then, you know, be in position. Uh you know, when, when chanGE BUT gets launched, when COVID, you know, COVID was also another big impetus for our platform. Um GP T three gets launched. So I think 2020 in particular was a kind of a seminal moment for us because it combines several tailwinds, particularly COVID plus GP T three gets launched. Uh GP T three makes a lot of what we were doing scalable, more scalable because we were like individually fine tuning weights on a customer basis. So it was not as scalable. Um
Hikari Senju
00:25:19 – 00:25:59
And uh and, and then COVID suddenly created a humongous demand for digital advertising or, or reimagining digital advertising, you know, many brick and mortar businesses suddenly realized they had to be aligned. A lot of big advertisers realized that they had to kind of rethink their a their, their digital strategy. Uh And so uh that created humongous, um you know, they say chaos is a ladder and, and it created an opportunity for us to kind of start growing during that period. But if it wasn’t for, you know, some of the early truly angel angel investors, I mean, in the essence of being an angel. Uh uh uh in, you know, the first couple of years of writing lean times. But also I think it developed my job as an entrepreneur to be super capital, efficient and resource efficient and,
Hikari Senju
00:25:59 – 00:26:07
You know, stretch every dollar as far as you can take it. And um and uh and so it was also good training for you as well.
Jeff Bullas
00:26:07 – 00:26:30
Yep. So, really, you basically had a good idea that maybe the timing wasn’t perfect and was hard because I wasn’t in public consciousness really until Chat G BT launched at the end of 2022. So, so you, you must, so you saw a big surge in interest in your service when in 2023 or shortly after Chat G BT launched.
Hikari Senju
00:26:30 – 00:27:05
Well, I, we saw a single surge in interest in 2020 when G BT three, which was kind of the API version of charge G BT was launched. Um Open launches GPT three in 2020. Um And, and so, yeah, that was really uh that, that was really what made uh you know, started, started the growth and yeah, definitely. Uh since, since then, yeah, every year has been a further acceleration of uh of, you know, awareness of what we do. And so um uh yeah, it’s, it’s, you know, like the strategy from day one though, right when I started the company in 2018 was,
Hikari Senju
00:27:05 – 00:27:27
you know, what is the winning strategy for writing for, for, for writing the A I wave? Right? And, and, and the belief I had and this is for me playing around with open source models from, you know, 2014, 2015, when I was a student at Harvard was a lot of technology could need to be open source because um uh this is a technology that driven largely by research institutions um
Hikari Senju
00:27:27 – 00:27:56
and open source communities as well as private companies doing research uh in this kind of commons of, of, of, of, of research regarding content generation. Um because it’s just interesting to so many types of players and there’s just something inherently interesting as a human, I think um regarding the process of content generation. And so I believe that content generation would keep getting better every year regardless of whether we invested in it or not. Um And so then the question was OK, assuming that gen A I is gonna keep getting better. This is, you know, the strategy from 2018.
Hikari Senju
00:27:56 – 00:28:39
Well, where, what, what is, what is, what is necessary right? To get this to be a product that scales. It’s the last mile, it’s building the ad integrations, it’s building the great analytics engines. Uh It’s building great enterprise workflows and workflows for advertisers. Um so they can harness this technology. And so from day one, what we did was we built a workflow tool for advertisers that we can uh slot in the latest gen A I models as the technology improves. Um And, and we can be a great application uh for gen A I uh for advertisers. And so that was a strategy from, from day one was build the best workflow for advertisers in ja I um let the open source community
Hikari Senju
00:28:40 – 00:28:58
continue to advance ja I in terms of the image or video or copy generation of their own because they’re going to keep doing that. And there’s way worse, more people working on this than, you know, our small team of rag tags for uh you know, uh team. And for us, we’re going to focus on the last mile. And that strategy has seemed to have worked out through, you know, the,
Hikari Senju
00:28:58 – 00:29:21
you know, uh since 2018, since we executed is that, you know, every time there’s an advancement, our product gets better. Uh And we’re the only really the ones that have built a complete workflow uh uh that integrates with the customers data stack that integrates with their creative team workflows and marketing team workflows that delivers them the insights that they’re seeking um and gives them the control that they need to really harness this technology,
Jeff Bullas
00:29:22 – 00:29:51
right? So um in terms of what tools, what you use to basically drive your engine, is it, is it chat GPT four and, and other tools, what tools are you using? Cos you are developing a niche product, which is more vertical. OK. As opposed to general Chat, G BT four and B and the other different chatbots. Um So what other technology you’re using to actually drive will support your technology and feed your machine?
Hikari Senju
00:29:52 – 00:30:21
Yeah, so we’re big fans of, of open A I and, and uh you know, um um we, we’ve been partnered with Open Eye from G BT, the, the era, the, the days when we were downloading GP T two embeddings from hugging face and fine tuning them on the customer brand information uh back in 2019. So, so we, we, we’re, we’re close partners, you know, and we’re close partners with, with opening eyes since, you know, back when they were still making, you know, video games uh uh uh uh simulations. And so um
Hikari Senju
00:30:21 – 00:30:55
uh we’re, we’re, we’re big fans of their, their platform uh and, and, and uh the, the platform that they’ve developed, but also we, you know, I have to deliver the state of the art experience for customers. And so if that means, you know, you know, couple, you know, uh uh you know, a couple of years ago, you know, starting working with stable diffusion and utilizing those models or, but today, uh you know, utilizing models like flux um and finding fine tuning those um uh and, and, and other models, you know, Lama comes to mind when it comes to um you know, uh the open source on prem um go to market um version. And so
Hikari Senju
00:30:56 – 00:31:09
um it’s, it’s uh you know, we, we deliver, we, we’re a mix of experts. We deliver um the best results utilizing, you know, the state of art technology wherever we can find it and delivering the state of the art experience to our customers
Jeff Bullas
00:31:10 – 00:31:27
using basically the best technology. That’s the foundation for you. OK. So have you been surprised by the pace of change in the evolution of A I technology in the last two or three years since Chai bib hit the streets?
Hikari Senju
00:31:29 – 00:31:46
Um No, no, no, I mean, you, you, you, it’s interesting, you can literally look at like my business plan or like this deck I wrote uh back in, in, in, in uh in 2020. And I said, you know, within, within um within a year A I is gonna generate copy really well and within it,
Hikari Senju
00:31:47 – 00:32:16
you know, a year or two, it’s gonna generate images really well and it’s played out exactly as I wrote in this, you know, and you can see my solid chicken scratch and I’ve shared it with my management team and, you know, they’ve, you know, it’s like, it, it was pretty clear, at least for me sitting from the perspective of again, you know, from 2018, 2019, going to the open air office and having friends who work there, you know, being at the research uh where the research is happening at mit following the open source technology, uh you know, having investors like Richard Soldier, who is one of the um
Hikari Senju
00:32:17 – 00:32:56
highest in H index uh uh uh researchers out there uh and just being really just steeped in the JA I or the A I machine learning culture in Silicon Valley and in, in San Francisco that uh that this was happening. Um uh I think if you weren’t as, you know, embedded in, in, in and steeped in this kind of community of machine learning researchers, um you know, maybe it was a bit more of a surprise but, but I think, um you know, for the people who are actually building technology and doing the research, uh you know, I think it was pretty clear. Um uh And so, um that for me, I, I think, you know, um
Hikari Senju
00:32:57 – 00:33:26
uh if San Francisco is a, is a city that lives in the future, right? We know, uh you know, just this weekend, I was literally flying cars, uh you know, uh uh over, over over the San Francisco Bay area, you know, those vertical takeoff machines where somebody can, you know, fly, you know, uh fly off and fly around. I mean, you know, there are more surprises, self-driving cars and uh ride shares right now in San Francisco. So it is a city that lives in the future and uh and, and that applies to things like flying cars and self-driving cars as it applies to the RNA. I,
Jeff Bullas
00:33:26 – 00:33:42
yeah, well, I think Elon Musk tonight is actually announcing a breakthrough. Apparently, I think, uh, for, uh, Tesla in terms of self-driving cars, I haven’t, it’s, uh, uh, been teased out as an announcement tonight, so I don’t know what the announcement is yet. Um,
Hikari Senju
00:33:42 – 00:33:56
you hear folks, you gotta hear folks October 10th. Uh, yeah, we’ll see how, you know, uh, you know, it’s, yeah, I mean, I, you know, I’ve no doubt, I have no doubt of Elon’s ability to execute.
Jeff Bullas
00:33:57 – 00:34:26
Yeah. Well, the thing is, it’s, it’s not a matter of, if it’s a matter of when, isn’t it really? So, that’s, that’s the reality. Um, and, uh, and that’s the thing about any good business businesses, like your overnight success. But you maybe saw the vision or the future and basically it’s taken maybe five years or four years or 10 years. So, um, sometimes it’s just been around long enough to actually win the game, isn’t it? Rather than just, um, and also having enough money in the bank sometimes to win the game.
Hikari Senju
00:34:27 – 00:35:02
Yeah, I mean, I think we live, you know, what are the capital markets by the jungle? And, you know, you gotta survive uh in order to thrive and uh, and in order to have a chance survive and sometimes it’s, it’s, it’s, and it’s fierce for famine but if it’s famine, you still gotta, you know, you gotta be one of those um uh those that, you know, uh like uh like a tartar grave, you know, they, they just don’t die in, in any environment. You, you like, you could, you could put a tar grave in space and it will somehow just, you know, uh and they survive and then you can bring it back to earth, it’ll come back to life. And, you know, I think as an entrepreneur, you, you gotta be like a tard grave, like just, you know, I, I,
Hikari Senju
00:35:02 – 00:35:39
I love the tartar grave and this is an animal I learned like in college and I whatever, whatever, you know, it kind of really does seep into how I think of starting up like a Tarar grave literally like takes the DNA of all the tard graves and incorporates them into its own DNA. It survives in any kind of, you know, extreme environment. Uh And I think there is an element where, you know, a AAA founder does need to be like that uh just able to withstand any kind of environment and both feasts because you can die in feast as well. You can, you can um raise a lot of money, waste a lot of money, hire the wrong people, not mismanage, mismanage your resources. So uh you know, there are many ways for a star to die, both, both in feast and in famine.
Jeff Bullas
00:35:40 – 00:35:47
Yeah. Yeah. So what was it, what were you saying that there was a base that actually we need to, what was the name of it? You’re calling it? It’s called
Hikari Senju
00:35:47 – 00:35:49
I believe it’s called a Tardy Grave.
Jeff Bullas
00:35:49 – 00:35:51
How’s that spelled? I’m done.
Hikari Senju
00:35:51 – 00:36:10
Yeah. Yeah. So it’s Tardigrade, I think is how you spell it again, don’t, I mean, uh, maybe you can probably search it on Google and find some approximation, I mean, I might be wrong a couple letters here and there. But, but it, it, I do, I do. It is my uh my spirit animal. Maybe you can say so.
Jeff Bullas
00:36:10 – 00:36:13
Is that, is that an actual, is that not a mythical, is that a real creature? No,
Hikari Senju
00:36:13 – 00:36:19
that is like a real creature that is a real creature that exists. It looks like a weird, like warm,
Jeff Bullas
00:36:19 – 00:36:23
right? OK. And it exists in spices. Is that what you’re saying?
Hikari Senju
00:36:24 – 00:36:41
Yeah, it exists like in the bottom of the ocean, it can exist like you like you can put it out and like dehydrates and somehow survives in space. Um It um it’s like really hard to kill and uh uh uh and it’s, it’s, it’s uh yeah, it’s a very, you know, it’s an animal with a lot of grit.
Jeff Bullas
00:36:42 – 00:37:10
OK. Fantastic. All right. Just so just to maybe sum a few things up. Number one. what have you learned as an entrepreneur and basically, right from the start, including, you know, your quick help and then on to omni Key, what are some of the key learnings that you uh have learned along the way as an, as a digital entrepreneurs and doing start ups, what’s important,
Hikari Senju
00:37:11 – 00:37:12
right? Um
Hikari Senju
00:37:16 – 00:37:42
I think, uh you know, what I’ve learned recently is how to effectively manage. Um I do think that there and how to effectively hire and how to effectively, uh you know, lead, lead, lead, a high performance team. Um And I think this is so important for a start because as you said, it’s all about execution, right? It’s not about the idea, everyone has ideas and everyone has good ideas, but very few people can execute against the world and, and, and the ability to execute against an idea and win, you know, in, in an idea,
Hikari Senju
00:37:42 – 00:38:21
uh you know, in terms of the opportunity of that idea uh is all about execution and execution is done by teams. And so, you know, how to manage, how to not to, you know, like, for example, right, like there’s this thing related to founder mode, right? And, and it’s a big theme of like, you know, an entrepreneur uh you know, should micromanage and should be really in, in everyone’s grill and really putting the pressure on. Uh and, and I believe that I believe that I, and I do operate like that to, to, to, to, to, you know, maybe, you know, for most of my time as an entrepreneur, I have been operating that and, and, and that is my uh instinct, but at the same time, um you know, uh, you also have to hold your managers accountable.
Hikari Senju
00:38:22 – 00:38:57
Uh, and, um, uh, and you have to hold your people accountable and set clear expectations and set clear accountability. And if someone is not, uh performing in their role, you have to also be willing to switch them out and switch them and, uh, and make tough decisions like this. And I actually think it’s hard to fire people to hire people. I think firing well, is like the understated tech is like the understated skill of uh then that, you know, everyone loves hiring, right? Everyone’s happy. Everyone’s, it’s like having a baby. It’s like everything, anything. There’s anything that’s possible though, but then firing is I made a mistake, there’s a mistake,
Hikari Senju
00:38:57 – 00:39:35
not easy to admit that. And um and so, um you know, managing people properly um hiring people while uh you know, uh you know, knowing when you need to make that switch uh so that you continue executing and delivering against your goals, uh running a high performance organization by having clear accountability, but also keeping, you know, but, but, but also holding people accountable uh or holding people accountable also, you know, really like being very involved but, but not to not being involved in a way that is not, which doesn’t cause additional confusion or feel, create, create feelings of disempowerment uh where everyone feels empowered and that um
Hikari Senju
00:39:36 – 00:40:11
um and that uh uh and that it actually, you know, your involvement doesn’t create confusion. So it creates more clarity, which is what’s really important. And these are very nuanced skills uh as part of being a CEO and, and uh and especially in a growing, you know, in a growth space. And so um that is a skill that I’ve learned more recently through my, you know, process of being an entrepreneur. I think, you know, many people can build, have an idea, many people can build a prototype and many people can get a handful of users. Uh But then uh and maybe even get to a million or a couple of million in revenue, but then to
Hikari Senju
00:40:12 – 00:40:26
create an organization to create a company, uh that is uh that is something that is a different kind of skill as well. And, and uh and, and a skill that I think I’ve looked, you know, like just through, through blood, sweat and tears really learned that
Jeff Bullas
00:40:26 – 00:40:35
That’s some great insights and like the insight about, in other words, you need a fire as well. I think that’s actually very important. So, yeah,
Hikari Senju
00:40:35 – 00:41:02
it’s, it’s I, yeah, I think it’s harder. That is harder. Uh Yeah. Um And also at the same time again, if you’re micromanaging somebody, you’re not doing your job, right? Like you need to be at the helm, you know, doing your job exceptionally well, you know, you can’t be doing, you know, three other people’s jobs. But at the same time that no one’s, you know, steering the ship. So like it is, you know, you have to let you trust people, you have to empower people. You have to let people do their job if they’re not delivering, you know,
Hikari Senju
00:41:03 – 00:41:26
that has to be made clear and then you have to make the, you know, the tough decisions regarding that. Um, but, uh but uh, but you also have to be supportive because, you know, it’s not always their fault that they’re not succeeding, right? Uh It could be, you know, it’s, it’s, it’s, there’s a lot that happens at, at a startup and it happens in the, in a global uh cap, in a competitive market. And so if someone’s not,
Hikari Senju
00:41:27 – 00:41:57
you know, executing doesn’t mean it’s necessarily their fault. But, but, you know, you do need to quickly figure out why something isn’t working and, and, you know, figure out and, and quickly kind of solve it and move on to that, to that next problem. So that is, you know, to do that with empathy and to do that with, with, with sympathy and do that with emotional intelligence. Um uh is uh you know, where people feel excited to come to work every day. It’s um it’s not, yeah, it’s a skill. It’s very, it’s very, yeah, it’s a very important skill.
Jeff Bullas
00:41:57 – 00:42:28
Yep. OK. So 11 last questions before I ask you the last last question. Um And that last question is what brings you joy. In other words, if you had all the money in the world, what would you do every day without getting paid? And that’s the last question. But before you answer that, I’m curious about, OK, you’ve got the technology, you launch it. How do you generate leads and sales? What was your strategy there? And do you eat your own dog food along the way? In other words, use your own technology to generate leads. Tell us about how you actually built the business. I’m curious about that.
Hikari Senju
00:42:28 – 00:43:08
Yeah. So we’ve always been eating our own dog. We’ve always been using our own tools. And that was also another reason why we started, we started in ads for many reasons because they’re the most data because I believe that data and distribution is the edge in general of A I. That’s where most of the data and distribution is. Um And, but uh and that we could do with that product. And so our go to market motion is we generate leads through ads, through targeted ads. We use our own tool for uh for, for creating and, and, and, and orchestrating our advert our omni channel advertisements. Um as our product gets better, we become more efficient at getting good quality leads. And that has a virtuous cycle of, of improving uh improving our growth and our product.
Hikari Senju
00:43:08 – 00:43:30
Um But, you know, the very day one, you know, when I, you know, when it was, um, a lot of interns, like, I think we, we really harnessed a lot of interns. Um, Cold calling leads. Uh, and, uh, and the good old fashioned way. Yeah. Right. Just, just, uh, I think that’s,
Hikari Senju
00:43:30 – 00:43:59
yeah, cold emailing. Cold calling. Uh, especially during COVID. There was a lot of students that were looking for work, an internship while they were kind of working, studying remotely, um, you know, and, uh, and so that there was something a huge pool of, of, of young energetic people who had time on their hands, they were, they were hungry to learn. And so I was very lucky to be able to utilize that. Uh and, and use that, utilize that to kind of propel the, he’s uh you know, growth during that period.
Jeff Bullas
00:43:59 – 00:44:10
Right? Cool. That’s OK. The last question to wrap it up. Um What if you had all the money in the world, what would you do every day that would bring you joy or does it bring you joy today?
Hikari Senju
00:44:12 – 00:45:07
II, I mean, iii, I think just doing what I’m doing at a larger scale, you know, I, I like, I, I genuinely am a person that lives to work and I love my job. I don’t, you know, it’s, it’s truly uh so fun, like to be able to spend every day with technologists, imagining and scheming the future and building the future um with, with, um you know, people that you’ve you know, that you were working with. Um and uh and so that, yeah, and a technology and a mission they are passionate about. II, I think that is like, what I’m doing today is what I would love to be doing forever until, until the day that I’m dead. I would love to be doing this uh hopefully at a much larger scale. Um uh but, uh but yeah, I mean,
Hikari Senju
00:45:08 – 00:45:38
you know, uh you know, if I had, you know, if you had all the money world, I’d just be doing the same thing because this is really, hopefully now that that’s pressure for me though because in order to do that, I need to continue to level myself up to be, to be the best person to be running this company. So I, I do need to continue to, you know, like level up my skills and my abilities to be able to, uh you know, do that at a larger scale that’s on me. Um But, um but yeah, I mean, if I, if I had, if I could do anything, I’d be doing this,
Jeff Bullas
00:45:39 – 00:46:23
right? That’s awesome. In other words, you’re on purpose and you’re following a bliss as Joseph Campbell said, which is uh the, you know, author of uh Hero’s Journey. So thank you very much Akari for actually, you know, sharing your story. It’s inspiring and I look forward to seeing Niki and I. I look forward to catching up maybe in the future to talk more about A I and creativity. Um, because that’s fascinating and I have a certain take on it and, um, just wrote an article on it actually on my blog A I, um, is it, uh, basic creative genius or is it just a ghost in the machine? So, anyway, we’ll talk about creativity in a, I, I think next time we catch up I
Hikari Senju
00:46:23 – 00:46:27
I really would, I’ll, I’ll read that article and yeah, it’s a really interesting topic. It is
Jeff Bullas
00:46:27 – 00:46:56
a very interesting topic. Yeah, I, I’m fascinated by it because as humans, we thought that we had, we were what made us human was creativity. Um But the reality is that uh A I is actually asking big questions about humans. In other words, the machine is asking questions of us as humans. What does it mean to be human? If the machine can do most of the stuff we used to do such as writing and creating art, there’s an answer in there, but we’ll talk about that another time.
Hikari Senju
00:46:56 – 00:47:17
You know, there’s a, I, I was reading kind of some of the notes of, of questions you ask and one of the questions you know, you, you ask, uh you was it, the notes was like book suggestions. And, and uh, I really like this book called A Celer Ondo. And it talks about like, you know, several millennia into the future of humanity. And, and it’s about the story of this um essentially, it’s like this virtualized human
Hikari Senju
00:47:17 – 00:47:51
that like uh spans because they keep coming back to life as like a virtualized person and every like millennia essentially and like, and like the things they have to deal with at every situation. And, also the funny thing is, like the precursor of every um chapter is the total amount of compute in terms of both human brain power and A I brain power that exists in that universe. Uh And um uh yeah, I think it is, you know, it’s just kind of a symbiotic thing, right? Both human brain power and, and A I brainpower kind of computationally having a certain, you know, my and you know, me instructions per second, you know, like um like
Hikari Senju
00:47:52 – 00:48:09
evolving together to solve bigger problems, bigger challenges, overcoming bigger challenges, building Dyson spheres around the sun. Uh It’s, it’s uh yeah, I, I don’t, I view it as a competitive thing with A I, I would view it as like a symbiotic thing as humanity, you know, improving our collective intelligence. So
Jeff Bullas
00:48:09 – 00:48:14
I totally agree. I think it’s basically an amplifier of our creativity.
Hikari Senju
00:48:15 – 00:48:16
Absolutely. I agree with that. Yeah.
Jeff Bullas
00:48:17 – 00:48:46
So I’ve just um it’s something that continues to fascinate me in terms of a lot of people seeing it as a threat. I just see it as an amplifier of myself and my ideas. But what I love about A I, it actually breaks up the pattern of pattern and fixed thinking we quite often have about our own experiences that it is hard to break out of those molds. I think I actually challenge that by recombining human global intelligence in a spot where we actually can help us break our own bounds. So I
Jeff Bullas
00:48:46 – 00:49:21
um I totally agree with you. I think it’s a hybrid model. It’s symbiotic and it’s um it’s very exciting. So I’ve just read another book called The Creativity Code by, um, which is awesome as well. So, look, I think we need to have another conversation soon about A I and creativity because it’s for me, I’m absolutely, um, curious about the intersection of A I and humans because, um, I think can help us make us more human, but there’s a lot of elements and in the mix and so next conversation Akari.
Hikari Senju
00:49:22 – 00:49:25
Thank you, Jeff. Really? It was a pleasure talking with you and looking forward to our next conversation.
Jeff Bullas
00:49:26 – 00:49:36
Awesome, mate. Well, I’m looking forward to it, I’ll be key. I’ll check out a little bit closer and, um, we need to look, we’re launching a new product so we might, uh, go and do a trial and might try you out. So it
Hikari Senju
00:49:36 – 00:49:43
sounds good if you feel free to try, you know, sign up on our website, it’s a free trial. And, uh, and yeah, we would love to. Yeah. Um
Jeff Bullas
00:49:43 – 00:49:54
Awesome. All right. Thank you very much, Carrie. It’s been an absolute pleasure. Um, and, um, we’ll catch up soon about A I and human creativity.
Hikari Senju
00:49:55 – 00:49:56
Sounds great. Thank you, Jeff.
Jeff Bullas
00:49:56 – 00:49:57
Ok, thank you.
{Categories} _Category: Applications{/Categories}
{URL}https://www.jeffbullas.com/podcasts/generative-ai-advertising/{/URL}
{Author}unknown{/Author}
{Image}https://www.jeffbullas.com/wp-content/uploads/2024/10/Episode-Banners-39.png{/Image}
{Keywords}{/Keywords}
{Source}Applications{/Source}
{Thumb}{/Thumb}