Dhaval:
With me, we have Coco and John. They are the founders of Open Art, and we’ll start off with the question on what does your product do, what problem does it solve and who does it serve?
Coco:
Thank you for having us here. So in one sentence, open Art is a cutting edge platform where you can discover and generate AI art. Our website, openart.ai has been rapidly growing since our launch in August, 2022. And our long-term goal is to build an AI-powered workflow for creativity. If you think about creative tools today like Adobe Photoshop, Figma, or even Canada, they help you design or create something once you already have a rough idea, but they don’t really help you come up with new ideas. You’d still go to sites like Pinterest or be hands check out other people’s work to get inspired. However, generative AI can really come up with new and exciting ideas, for you and or with you like no other previous technology could before. So on our platform, we empower users to unlock their creativity with AI by one, generate images based on prompts. And second, train their own personalized AI models. Just as an example, so a fashion designer uploaded her past fashion sketches and trained her own fashion design AI model. And now she can actually immediately get hundreds of new designs waiting minutes. because these designs are based on her pastwork, but entirely new, created by ai and this magical process, really 100x her creativity. So we’re really excited to let more people 100 x their creativity with the workflow we’re building.
Dhaval:
Very interesting. So you said you launched in August, and this is a workflow product to empower creators with, creative graphic prompts. Did I get that right? Image prompt. Is that artistic prompts? Who do you serve? Like, is it, you mentioned fashion designers. Is there a specific segment within. Creative industry that you have nailed your product for.
Coco:
Yes, great question. To be completely honest, we’re still in the process of nailing down very specific target audience. But I would say right now, we are, our audience are creative workers like artists, designers, including, like the fashion designer talk about, but creative workers who have like ideation phase.
Dhaval:
Very cool. Yeah. There is a common theme among the AI product founders, which is if you empower the creators, then your product will go a long way instead of if you trying to replace them. Where do you draw that line? Where do you draw the line of empowering versus actually taking over a little bit of their creativity? Where does that line stand for you?
Coco:
I think it depends on how people use your product because I think the way I think about it is that for for creative workers like artists or designers, like.At least our product, they can really use it as a workflow tool that really 100x their creativity. And we’re constantly talking to this artist and designers how our workflow could actually help them monetize more like help them get more clients. So the way we work with them is very, very collaborative. Whereas I do think in terms of replacing artists I can see like in some use cases, for example, if you don’t have artistic skills, let’s say you are a writer and like you need some illustration, so perhaps before you don’t have good tools to get any illustration. But now you can use our tool to make some basic illustration. However, I do think if you need high quality illustration, you still need to go to artists. So I think the AI could replace some basic work of artists, but then like they want really replace artists.
Dhaval:
Very interesting
John:
also think about the human beings, right? So the, the most important or valuable thing is our creativity. The AI just unlock the whole potential of our creativity. So instead of wisdom that maybe 5 years, 10 years, to master like your skills to draw a picture. To paint on the wall. So the AI can help you do that. The most valuable thing I think is, of course your skills is much more valuable. But the creativity is the core part of the whole thing.
Dhaval:
Yeah, I completely concur with you on ai supporting the creativity and creativity being the most important thing for human beings, right? One thing you said is the customer experience, where in the customer journey. Is AI prompting tell us a little bit about your customer journey and where does AI get infused, like specifically in, in the user experience? Yeah, if you can share that a little bit. And how did you go about making that decision?
John:
I can give you an example from one with our user on the platform. That, that guy he has a, like a blue character called Coco, , he has like a maybe 20 to 30 image of the, that character he designed and he upload to our website and use our photo booth feature to generate a model.Then one of the special feature we had on our website is like a presets we have like a presets team in our community. With Very good at writing prompt, writing prompt for the Generative AI is really hard, right? But they’re super great at that. So we have a lot of presets and that guy will just by bunch of presets and apply those presets on his characters.And you can see the result are super creative. the blue character is on different off page is in different contexts in the background. And I think he really like that because he’s trying to upscale and enhance the image for lots of the results. That is really impressive thing.
Coco:
Yeah. And I want to add one thing is that after, for example, they see so many amazing results. They usually pick the ones they like, and then they can use it as a reference sheet to rejoy it. Or they can actually, if the quality is, they’re happy with the quality, they just need to do some touch up on the, on the on the image. And then they can actually, that can be their final like image.
Dhaval: Wow.
John:
I think the guy also came back later and he also purchased lots of model and know, keep changing that factor.
Dhaval:
Wow. This is very interesting. The three keyword I picked up here are preset prompts and outputs, right. Is the sequence presets, and then from the presets you create prompts and then the prompt from the prompts you create final outputs. Multiple outputs. Did I get that right? The sequence.
Coco:
almost, I think so the step are like one you need to train your model and training process is super easy. You just upload some photos of the thing that you want the AI to learn. It could be, for example, an image of yourself or image of a character or image of even a consistent style. So you upload it and after 30 minutes. The model is trained. And and then you can purchase those presets that John mentioned, which are essentially highly curated prompts. And then those prompts will be applied on your model. And then the third step is that after those prompts are applied on your model, you can see like hundreds of amazing results immediately.
Dhaval:
Very cool. Wow. I fully visualize it now. Thank you so much for clarifying. Beautiful product. One quick follow up question is, you said you launched in August, 2022. Tell us a little bit about your business results. How many users you have, how many users are using your product, what is your revenue, if you’re willing to share that et cetera, et cetera.
Coco:
Yeah I think last month we have around 800,000 visits on our platform. And yeah, and our SEO is really picking up, if you search an AI art generator, I think we’re the top three results. And in terms of revenue we wouldn’t be able to share the specific numbers, but then it is in the range of like tens of thousands a month.
Dhaval:
Wow. That’s a really good MRR for someone who’s, like just launched in August. Right. So six months in. Oh, thank you. Yeah. Tell us a little bit about whether you bootstrap this, whether you funded yourself, whether you raised some seed round. Where are you in that journey?
John:
Yeah I can tell a little bit story about this. It’s kind of like a also like a related topic of how to find a good program marketplace. I think that overall, the timing is really important because the before open art, Coco and I actually quit from Google quit Google together, last year. And we had idea to build a marketplace for social and medium influencers and e-commerce brand. But that didn’t play really well eventually. We pivot and we are very proactively looking for different ideas. I think it’s Back in, August open AI DALLE-2 started getting lots of attention, right? Coco and I have a daily idea of brainstorming session, and we do it every day. We analyze different ideas, directions and see if there is something we really want to work on. So I remember it’s a Friday afternoon and we were discussing about the DALL-E 2 because both of us already have the early access and we have been using it for a couple of days. And then we found a pain point that people are struggling to come up with a good prompt to generate like a high quality image. So we think we had a solution and I’ll tell you a solution a little bit later. And then we immediately decided to jump on this and we build our prototype, like first of version, In a week and on the day, I think it’s August the 26th. So Coco made a post on hacker news. And the title is Open Art, Pinterest for DALL-E 2 Images and prompter. Boom. There were like over 10,000 people visiting our website within 24 hours. That’s how, we started at the beginning.
Dhaval:
Wow. did you say you started in week?
John:
And then tell us about what happened after that. Like what has been your journey? Like, have you been bootstrapping since? Has the customer funding you, are you raising capital? Are you in middle of it? Yeah. Where is that journey?
Coco: actually,uh, we actually did a pre seed found race last year. Earlier. Last year March or April. But it was for another idea and just like many startups we pivoted but we are thinking of raising our seed round this year.
Dhaval:
Very cool. Very cool. So you said you launched a product on Hacker News. Within 24 hours you got over 10,000 users and it wasPinterest for DALL-E 2 image and prompts
John:
Oh, the, the goal is mainly for help people to discover different AI generator images and also know every image has the prompt so they can use the prompt or get some inspirations from a prompt to generate their own images.
Dhaval:
Wow. Okay. And is this, is your audience mainly creative individuals or is enterprise tools like Figma or Canva who want to be able to support their users with creative workflows. Where, where do you focus on right now and
Coco:
Yeah, I think. I think we’re constantly involving. At first when we launched, when John said, I think all we wanted to solve was a small painpoint that people have, they struggle to write prompts. So we want to have a search engine and a, like a Pinterest form to help people write better prompts. That’s how we started. We were targeting AI enthusiast, but as we were building, we realized how powerful it is for creative workers like Artist and Designer to actually use it to help them with their existing workflow, to help them like come up with new idea in the ideation phase. So that’s what I think we’re focusing, a workflow for creativity.
Dhaval:
Thank you. John, you mentioned about product market lessons. Do you have any other lessons to share with us since you launched? You have a lot of users, 800,000 people visiting your site every month, tens of thousands of revenue. Tell us about the product market fit lessons you have learned since your launch in August. Any big lessons that you wanna share with us, especially focus on ai. How do you find that balance for other AI product creators who wanna build an AI product and be as successful as you?
John:
Yeah I think Coco can add more, but I, I can write one more thing before I mention timing is very important, right? So we were really lucky to to write on the , tailwind. So I think another one lesson learned is in this industry specifically for Generative ai, so. The change is so fast. Every few weeks there, there’s something new and come to the industry and you have to act really quickly For example, at that time, August last year. The most appealing thing is the text two image, right? So people are creating, uh, different images. But the, the November, there’s a new technology, like a called Dreambooth you can train like a personalized model within an hour. We quickly adapt that technology and build our own product. So I think that’s the biggest lesson, I kind of learned so far. It’s like that you really have to act quickly and, keep listening. The community, the industry.
Dhaval:
You said there was dreambooth. That’s the product you use to customize your model. Based on that you quickly created an MVP. Any other quick product creation techniques or tips that you have for other aspiring AI product creators in this space?
Coco: I think I do have, one piece of advice for founders, especially actually non-technical founders because AI can sound very intimidating for non-technical founders. However, the thing is that a lot of, like a foundational models like stable diffusion or DALL-E 2 or GPT3 on these foundational models, they’re becoming commodities. So there are so many opportunities in the application layer. Like for a non-technical founder, it’s a really good thing because you don’t really need to be the expert who build all of this. Like technologies from ground up. Instead, you can use the APIs, you can use like these open source projects. And the most important thing is to actually have some unique insights in specific industry. I would say that the best AI products are probably the one, that is a combination of your unique insights in the specific industry and ai. So as a technical let’s say if you are a non-technical founder but you know, like specific industry so well, and you can just think about how these foundational models can be applied to solve some problems in your industry.
Dhaval:
Yeah, you had so many good gems in that last 30 seconds. I’m just gonna unpack that a little bit. So, for non-technical founders, what they can use is commoditize foundational models that has been abstracted out. So they don’t have to build ML models. They can use the existing models, but what they need is they need specific experience in a given industry to be able to pull insights from that so that they can customize these models. Using other tools like Dreambooth to make it specific for that insight that they have acquired from having worked in that industry. So use the commoditized model, use your ton of experience in a given industry to build on top of that, and now you have a product that can serve other people who are in that industry. And you don’t even have to be coding that you can just put it together using bunch of APIs and, existing models.
Coco:
Yeah, exactly. APIs for example, I think we are going to provide , our APIs very soon because a lot of developers are actually asking for it. So I think we will also make our service very soon as APIs so other people can even build on top of us.
Dhaval:
Very interesting. I was curious, when you, when you first release your APIs, are you going to focus on a specific segment of market or are you still discovering that?
Coco:
I think the most, the, asked for API is the. Like related to Dreambooth, it’s our photo boost feature which allows you to train your own personalized models. So the API will probably be the one that allows you to train your personalized models, AI models easily.
Dhaval: Got it.
John:
Yeah. We actually have like a group of people like, who really wants the API from us. And basically, like Coco mentioned, they want API to fine tune the model because we have the best experience. And also they’re asking for the crystal API to call to create images and search API because we have like over 10 million images on our website, we previously collected, and we also have the community to keep generating the new images. So those developers want to, gather a API to query the whole data set so they can use the images to somewhere else. The image itself is definitely for available for any, u usage like a commercial or non-commercial.
Dhaval:
John, you use a very important insight here, which is community. I wanna unpack that a little bit. You can build AI products using foundational models that are already pre-trained products like GPT3 et cetera, et cetera. On top of that, you can use products like Dreambooth to fine tune those models to make it specific to your industry. Now, with this two products, your technology stack is taken care of. What remains is your specific industry experience that you can bring to product creation and your customer experience, your user experience, and your community. Would you say, now that there is so much commoditization of ai, would you say that. The three differentiating factors for AI product creators is gonna be user experience, the community, and the industry insights that they have gained from having worked in an industry for long. Would you say those are the three differentiating factors for AI product creators, or is there still a need for in-depth technical, fine tuning to be able to create meaningful, valuable, net positive product?
John:
I think you summarized it really well. I think that three elements are so far as we can see right now is like a very core part for a AI company to success, at least at the beginning. And in terms of that more deeper technologies. I personally think it, it’s true you eventually you have to create the technical bar. To really to make something from good, great, and excellent. Right? Excellent. It’s like a copy AI or maybe a Japer ai. What they did before, they just leverage the existing tool or maybe API to build the first momentum. And regardless what resources they have at that time, but eventually they will have their core technologies to build a mode to build a the true differentiator for their company.
Dhaval:
Thank you. Thank you, John. Yeah, that was key insight for this podcast. Thank you so much for sharing that. Yeah, I think we are gonna change the gears a little bit and wrap this up. Now, do you wanna share. The future vision of your product, of yourself, and maybe tell us a little bit about where do you see you going? Where do you see yourself? Like where, what is the future for you? And then what is the future for your product? Tell us a little bit about where you see the future vision of yourself and your product. Yeah.
Coco:
Yeah, so I think the future for our product open art, I think the possibility is really unlimited. Even though I think I can give like a five year vision, but I think just based on how quickly the technology advances or like how, how much insights we’re gaining every day. It could be different from where we envision today, but if you ask me today, I would say we wanna build, a workflow that can really 100 x everyone’s creativity but I think in terms of details, it really depends on I think the technical breakthrough will really affect our roadmap because a few months ago we didn’t know about Dreambooth and like personalizing or AI models was like, we didn’t think about that. But then the technology came out and we really see there’s like great value in using this for creative workers. So I think in the next few months and years, we will have like amazing technical breakthrough that will enable a lot of things we cannot imagine today. But I do think. Our goal of like 100x creativity will still be the same. And in terms of ourselves, I think we are really enjoying this journey. We really hope to raise our seed round this year, and become profitable. I think we’re on a very good trajectory on being profitable. And yeah, nailing down our product market fit.
Dhaval:
Thank you. Thank you for sharing that. It was a pleasure to have you on this show. Thank you so much for being here. Any last thoughts before we wrap it? Let me ask you this question. Which CEO inspires you the most?
Coco:
I think for me, this will be a controversial one because I would say Elon Musk he is controversial, you know, but he, the thing I really appreciate about him is that he’s so crazy and extreme. Like I think he’s the extreme version of me that I wanna be, that I don’t think I’ll ever be.
Dhaval: How about you, John
John:
for me it’s Steve Jobs. I think he’s a really good person and he made like a huge impact on the entire society, right? So that, the one thing I really want to do because I started programming since I was really early nine years old. I build tons of shit from then. so I remember one person ask me a question, okay, you are keep building stuff and when do you stop? So this is how I tell him, okay, the time when I stop is when I work on the street. I see people are using my product and enjoy it and it, the product change their life and make their life better and make them happier. So I think that’s the point when I can try to stop. Okay. So I, I made my impact the world.
Dhaval:
Thank you. Thank you so much for sharing that.
