Feb 26, 2021
This is an interview with Chris Lu, Co-Founder of Copy.ai. For those interested in our content marketing and blog writing services, please visit:
Hi, this is Ryan and I'm your podcast host today. We'll be having Chris Lu here on from Copy.ai. Welcome to the podcast, Chris.
Thanks Ryan for having me. Of course.
Well, we've been following you for a while now and wanted to get you on a podcast, discuss where copy.ai is going and where you see it going. We've seen some growth within your different tweets online and just an upward movement, especially with COVID-19 hitting. There's been some significant growth with people doing stuff online being able to grow their businesses because that's where everybody's hanging out is online. So if you could just tell me real quick your background and how you got into orchestrating this, this great software.
Yeah, my co-founder and I move on to Paul Jacobian. He's the guy who tweets about our company. And I have worked together for five years now. And we were at an investment firm. However, what constantly fascinated us was entrepreneurship specifically, like how do you start it? How do you find the creativity for it? How do you really push it to the limits? And we realized that's a very strong creative exercise. And when we launched our own thing, we realized how much writing there was to be done. It's unreal, like everything from the mission statement and the landing pages to the initial emails, the launch posts there's just so much writing to be done. And that was one of the hardest parts. And then we realized that, you know, the world needs more entrepreneurship and COVID really helped celebrate that. And at the exact same time, a very powerful AI algorithm was released. And we realized this AI algorithm is the first foray into AI powered creativity. And so we decided, you know, what, this needs to exist. We need to distribute this AI to the entire world to really prove entrepreneurship and our goals to have a billion new entrepreneurs in the next decade.
That's great. So with the creating the software, was it an iterative approach? Like, did you create a little tool and then expand upon it or did you have the idea to have it do what it does today and where do you see the growth of that?
Yeah. when we started, we were very focused on highly repeatable writing processes. So in our eyes that was like writing ads. So our first version of that tool only had like seven or eight tools and it was around Facebook ads, Google ads, and Instagram ads and captions. And then what we realized is, well, that's actually a pretty small piece of the market. A lot of the small businesses, they don't actually even run it, but instead what they really need help on our product descriptions, blogs, blog, intros. And that's when we started expanding the scope today, we are aiming to build out tools for almost every single use case you can imagine. And the reason is when we build out a new tool, we really want to help empower people for a specific problem that they have. So for us, you know, for example, coming up with ideas to write about, to coming up with that first intro, that book I could really continue to keep so much continued reading. We think that those are really, really hard problems and it would be mentally tired. And so we're kind of build out as many tools as we expand to help facilitate that creativity.
That's great. So with, with this technology, I know if it's built off of open.ai and they use a technology called GPT three or GPT four, I think is GTP four is out and you're playing with it right now. What is the difference between the two?
Yeah, it's actually deeply two, three for now, and we are fascinated by what GBT four will be. So a little background on the technology it's open AI had this thesis that if they can train a margin of model and a large enough algorithm, it will become smart. And so they launched something called tributary to about a year and a half ago. Now, maybe even almost two years ago now. And it was pretty impressive. But maybe one in a hundred or really mind-blowing results. And then maybe like one in 20 was a passable and the rest are complete crap. So Paul actually spent a lot of time, but keep it to two, really trying to explore the capabilities. And that means wading through a lot of, you know, bad content to look and find the gems. GPD three is a model that's just a hundred times bigger than GPC two.
And it's probably one of the largest models that has been commercialized to date. And when open AI released it we saw immediate, it was like, wow, maybe like three and 10 where like really good results. Five out of 10 were probably passable. One in 10, maybe one in 20 were like, mind-blowingly good. And we knew immediately that was much more commercially viable because, you know, with one or two clicks, you get really, really good results. The way we see GPT three and give me a little bit of background. There is it's a model where they just try to predict the next word so where it's not technically a word, but we'll say it as the word here. So if you have like my name, like the next word has a very high probability to be is. And so that might try to predict that but there's also a little bit of randomness thrown in.
So there are other words that could go after. Hi, my name. And then, you know, it may choose one of those other ones on separate runs. But what this results in is everyone is trying to set up the context for it to finish running. And so it looks very much like Harry Potter stuff, the better you can set up the context, the better you can prepare the AI to complete what you need it to be the better the results. And so even though it's a lot of companies are being built on JPC three, the way you implement it actually has a very drastic differences in the quality changes radically as well.
Got it. So is that kind of your competitive edge, bringing a creativity model to it where you can take this open API that they're providing, but the way you're reverting it to create is a little bit different than what competitors might be doing or what's your competitive edge?
Yeah, that's probably one of the strong ones and I think our competitive edge comes down to our mission and vision. It's our North star metric. And while it's not very powerful today has the moat. I think you have to look very different five, 10 years down the line. Yeah, I think the way we approach the market matters a lot. It really tells you which tool to build first, which ones to prioritize, and it also builds out a brand in the process.
Got it. So are you limited right now to TP? I gotta get the acronym, right? The GPT three, or are you kinda like really anxious to get the GPT four in your hands? Or are you maxed out at what it can do right now? Or are you able to take a little bit further or where do you stand right now with the technology?
Yeah, GPT three is already really, really powerful and I think we've barely scratched the surface. When it first launched, people were showing up all kinds of cool demos, like having it write code having it create designs. Of course, you know, the w the way you implemented is very different than how we're implementing it, probably Revit words. But it is just as impressive those tasks. And as a result, I think that there is a whole world of possibility that we haven't even tapped into all of our tools are basically the most simple building blocks or today but down the line in the future, I think there's going to be a lot, but really, really cool things you can do with it that are much more advanced. That will be much more helpful to users as well for if, and when they launch it, I can only imagine how much better it will be. If it's anything like the jump between Japan two and three, it's going to be mind-blowing. But it's already pretty hard to see how much better it can get that. One thing that is on the plate though, and this is a little tangential is open AI, released a blog post about a Dolly, which is a image generation API that hasn't been released yet, but effectively you'll be able to describe an image and then the AI algorithm will cry.
Wow. That's great. So do you, do you, is there like a list that you've created internally of like all these ideas that you could create with it? Because really the sky's the limit in terms of creativity, I can only think of a, you know, five or six use case applications, and then I get inside your application and there's a 30. And do you have another hundred that you've put on a Google doc somewhere or secured in a safe I, I don't expect you to share those with me, but I'm just curious as to where the ideas STEM from, or are you doing some highly classified brainstorming somewhere? Or w what are your, what are your ideas behind that?
Yeah, our tools usually fall into two categories. One is business. So it's like anything that requires you to write something, a business, like it could be used for that. And so we talked through it, our product a lot. If we need to write a tweet, if we need to, you know, do anything, we will like test it out. Hey, the GPC three recently we were looking at job descriptions and it's like, Oh, we need to create a tool for that. So that got added to the list. It's not ready yet, but it added to the list. More recently the other side of tools that we have are more first-line more for personal utility where it's like, you know, stuff that's just been painful to her. Right. then it's less, you know, associated with marketing or SEO or anything along those lines.
So, for example, for students writing your cover letter, like I remember when I was a student, I had no idea where to start. And, but now we have a tool. So if I were a college student today, I would just type in the job that I want, or the company that I'm looking for, click a button that I have a great starting point for a really, really solid cover letter. One other fun tool that we'd launched recently was for Valentine's day. We had this thing called Valentine's day.ai, where we would help you write Valentine's day cards. We believe that people really want to express their love, but sometimes struggle to find the words for it. And when they struggle to find the West for, you know, it just comes out cheesy and it just doesn't feel right. With AI, you can actually help brainstorm the different words and you can input like inside jokes or, you know, more contextual stuff about your specific relationship and get back words that you could use for your own Valentine. The way we see a copy AI right now is much more like an assistant that helps you brainstorm you take bits and pieces from all the different generations that has, and then you put it together. And that final product is around and you may have, well, you wrote it. AI just helped you brains in different ways, but you're the one who edited it and figured out this is exactly what I'm trying to express.
Yup. I hear that. Exactly. It it reminds me of a experience I had in early and then late nineties when I was in an English class. And I literally that's when the internet and more content was coming online. And I literally was reading the great Gatsby and I copy and pasted a section, and I put it into my report and the teacher kept up bringing up plagiarism, and I didn't know what she was talking about. And then eventually I got a little slap on the wrist and my grade dropped a little bit because I had plagiarized not really thinking anything of it, but something like this, you put that paragraph into copy.ai, and it's going to regurgitate a different way of saying it, right. Not to say that we should go off and teach our kids this, but I think it allows them to get more creative in coming up with verbs and adverbs and nouns that would allow them to express what they're trying to say.
More effectively they still have to make completed sentences and crafted, you know, make it personal. And I don't think AI will ever get to a point in some industries that will like manufacturing and that sort of stuff. But with the creative realm, I don't think it's going to replace the copywriter. It's just going to empower them, just like these other tools, other manufacturing tools. So with that, are you, are you seeing a big organization here with copy.ai? I see that you're hiring. What are those positions and where do you see a copy.ai going in the, in the, you know, six months, year, two years.
Yeah. You touched on a really, really strong point there, which is we are deeply, deeply passionate about AI to empower people, not replace them. As a matter of fact, one of our founding mission statements. So we first left our jobs was there was a narrative out there where it says AI is going to take over, take our jobs. And if that is true, there is a second narrative where it will, AI will make it easier to start Coopers. So we are on a mission to help shift people from losing their jobs to, Oh, no. Now I am part of being that company that gets the benefit from the AI. And that it's a very subtle thing, but I think it's really, really important. And so everything we do has been around, how do we empower to do this? We don't do auto.
We do suggest we really, really, we think that ultimately is great at generation, but it's down to the human to churn and bring it back to exactly what you want to express. That being said our long-term vision is to create creative tools that help entrepreneurs start businesses. We believe that very deeply in the passionate economy. I don't know if you've heard of that, but it's basically, you will be able to make living focused on your passionate nature. Content creators are probably the early signs that are living in the future and it's because you can make content about anything. So they're, you know, they're already, it's a very meta but to already do it, you know, it's like, Oh, I am passionate about trains. And so I make a lot of videos about different trains. I get a lot of views. I can run ads on it and there you go. I have a living.
So we, we really do believe in
This. And one of the hardest parts though, of starting a business is getting all of the words, right, getting the messaging, right, connecting with the customer. And so our tooling wants, they want to create these creative tools that help them get there. One other point that you touched upon is we're not going to replace copywriters. As a matter of fact, copywriters will actually be empowered in a, in a much more powerful way than everyone else. And the reason for that is the point I mentioned before, where every single word you use to input into the AI matters and a good choppier, your writer will have better words to input, which means that AI will understand your intentions in a much better way. So if you use a word like, you know, I, I like, you know, she, or something like that and then you change it to like, I like this specific type of sheets. You know, it, it, it really changes the output. And the better you are with words, the better you are with telling and describing at this, when you were looking for the better, the results. So CA good copywriters of good storytellers will find this as a super tool. Yeah, that's what we see it as well
Internally. We are not just our copywriters are creating content for our clients, but we're actually paying them to create content for us internally. Right? So we practice what we preach in terms of, we were curating a whole bunch of content because when people type in those long tail keywords, we want them to find us in order to do that. We need to rank high on Google and that you bring up a good, interesting topic that we could probably talk an hour about, but essentially this, this, the passion I guess influence of today you know, people that like, I want to do what I'm passionate about. They have the ability to take content. That's already been created, put it into this tool, whether they have to do several copy and paste to generate it, but they can spend like a half an hour a day doing a blog post that was curated, you know, 10 years ago.
And they can repurpose it as their own because they changed it up and up where it's not plagiarism, where they can do that. And when they start doing that, if they're consistent with it and they put in the keywords that they're looking for to be ranked for, they can rise in the Google rankings with the ability to get to those people that are interested in that specific niche. So I think the empowerment that we're allowing people that even aren't copywriters, that they will have the ability to say, I'm passionate about XYZ. And this is a tool over here that I have to spend. What, $40 a month, $50 a month. I'm not, you might raise the price to a hundred dollars a month. I'm not sure, but essentially what you want them to do is say here, here's another tool in your tool belt that allows you to start a business.
Not just that you're not a copywriter take copy that's over on this site and bring it to this site. So are you seeing yourself using copy.ai as you expand the options? You know, the, the passionate person say there's a hundred different niches. Are they going to be able to go in there and say I want to do it for this specific niche, or it's already to that point, or, you know, when you get so many options, are you going after the small group over here, or are you going after a huge group over here, you see yourself branching out into different, you know, making an umbrella over coffee.ai, where do you see it going?
That's a great question. And we are still debating it internally. There's just so many ways to go. But what is our North star is helping empower more and more entrepreneurs. So we definitely will want to launch more tools for beginner entrepreneurs to really be able to ramp up quickly. And it's very much like Shopify is arm the rebels in a motto where it's like, we are trying to arm the rebels. You know, that AI taking the jobs thing. It's like, we do think that's going to potentially happen. But you know, we're going to arm the rebels. So not only are they going to, I thought they're not going to get replaced, no employees, we're going to empower employees to quit their jobs first, and then start these other things. The result's going to be the same, but the employee the everyday person is going to be far better off on your other point about the different niches and whatnot.
One of our favorite tools is actually changed tone, and we need to rename it. It's a horrible name for the tool, but basically you can type in like, you know, a product description or something from a different, you know, competitor website. And they choose the tone that you want, click the button and get the same thing and rewritten in a different tone. So for example, if you have a candle brand and you want it to have this more playful, witty sort of feel to it, you can literally take a competitor's, you know, candle description, you know, and then put it through and then click witty, pressed the button and get back with your channel descriptions that will have a lot more personality. And you can build an entire plan around that around being like a witty camel grant. And you know, you can already imagine what that would look like.
And you can do this across everything. And so as long as you can find something that you're passionate about and you're willing to spend the time to, you know, really make it happen, really define the brand, the voice, you know, what you really want from it. You can start a business today in less time than ever. As a matter of fact, you could probably start that business within a day, you know, get it up and running on Shopify, find the manufacturers that are going to drop ship these candles, for example, you know, and then you can probably launch it by the end of the day. Before it might've taken a lot more time, because it would take a lot of times, or right, the top of the, you sort of really define the brand, but these days it's a lot easier.
And that's one of the big motivating factors. It's Paul and I think about this a lot. If every single one of your business ideas has a 10% chance of success, how many times do you have to do it before overall? All of our projects has like a net. At least one of them has like a 95%. And the answer right now is 30. You have to do it 30 times and you'll have over a 95% chance of one of them succeed. What we want to do is not only decrease the number of times, but also we want it in garden and make it easier to do more projects, but we also want to improve the potential champions Fest. We can push that up. You know, it's a 20%, you know, the number of tries to find a successful business actually drops significantly. And we believe that this will happen across everyone, you know, talk to any entrepreneur. They've had multiple, multiple failures. Me, myself, personally. I made it take, talk about it. I have like 10 failures underneath my belt. And you know, this one just seems like an overnight success, but it was many, many years.
Yeah. And you make a good point. I think that there's various reasons why a business would fail execution, but this essentially is giving them a tool to say, Hey, we'll take care of you over here. Yeah. You got to do the marketing aspect. You got to pay for ads, you got to curate content, or you gotta take it to your audience. But this is just another way that we're empowering the marketplace to, to bring it to the percentages that you're talking about. If they have one thing fixed on their plate of to-dos, then there'll be able to, you know, take a breather and say, Oh, I can focus my energy or my time on this over here. But if, if they're not able to grasp writing or, or able to create creative writing that penetrates or excites their audience, then they're really, they're not going to get anywhere. Right. They can have a website with a nice picture, but if it's not explaining it properly, then they're not going anywhere. So it really, I think it levels the playing field, like you said before, I think copywriters, it's not going to replace them, but it's going to make copywriters out of non copywriters as well.
Exactly good copywriters, you know, and the experienced ones are probably the best storytellers on the planet. And that is a skill that just won't be easily replaceable. They'll have a very good sense of how do you tell this story? How do you make it exude through all the websites? So if I were starting a company today and we did I would use copy AI to potentially run through the first version, right. Get some traction and validate the problem. If we can validate it, then the next step is to probably bring on someone who's a really good storyteller, some more experienced copywriter to really help refine the story and make it feel and flow a lot better. Now, that being said, you know, these days there's so much content that needs to happen. You don't need a full-time copywriter on every single Instagram post. Yet you do want the Instagram posts to have high quality. So copying, I can fill in that gap. So there's just, I think there's going to be so much more that needs to be written. There's probably not enough copywriters out there,
Right. So I'm just curious in respect to the creative element and not any boundaries. So if I were to take some content from a website that is written like a sales page, and it has a story and has a hook, and I regurgitate that into copy.ai, is it going to change it enough that would allow the person? Or is it because that story or that it's written a specific way that you can't repurpose it with copy.ai and you don't want to deal with the legalities there, or I don't know, have you looked into that or what that looks like?
Not yet. We are focused very much on a shorter form, like maximum 400 characters or so. And the reason is when you start looking at full landing pages and the rewriting entire landing page, it does get pretty complicated. But what I can tell you though, is most managers do tell a story in a pretty similar way. It was like a lot of those, you know, very common copywriting, formulas, I pain agitate solution, ADA, whatnot. And we can help you brainstorm boats, which sort of thing if you saw a sales page and it's like, wow, they did an it solution really, really well for this product. Could I do that for my product? You can run it through, copy AI, get it back, and you could find other pain points or other ways to excite them. One of the early users we had was so excited because they'd been thinking about this topic for 20 years.
It wasn't about AI or anything, but just thinking about a topic for 20 years, he ran it through copier. They gave him four or five new angles that he hasn't even considered before. And he's like, wow, like I've been focusing so much on trying to understand this topic yet with one click of a button in 10 seconds, I was able to get new ideas and that just blew his mind. And that was what we think the power of this AI can be. It's a really, really good at lateral thinking. It's really, really good at pattern recognition is really, really good at trying to find similar things. And it also has a huge knowledge base. I should have explained earlier, but this model is trained on about 10% of the internet as of October, 2019. So it knows a lot of contexts. And as a matter of fact, one of my favorite tools is a startup ideas where you can go in there type in you completely unrelated topics.
And it will try to create business ideas for you out of these, out of the intersection, this topics. But there was one where I did like, you know, a dinosaur as you know, and then like amusement parks. And then like, you know, something I know is giving me incredible, like BR dinosaur experience, you know, apart, or like, you know, and it was mind blowing. And so we believe with these tools and if you put in your passion, you will find something that you back intersect all of your passion, that you would the right person to pursue this idea, just because this is who you are, what you believe in and what you're really, really tasked with.
Well, that's great, Chris. I appreciate your time today. I just, I think that's a good stopping point for us because I want the, our audience to understand your tool, its capabilities in a trajectory in which it's going in the future. So please what, what's a good way that our audience and your audience can reach out to you if they can reach out to you personally. Great. If you just want them to go to the site, but how, how can we get in contact with you?
Yeah, we live and breathe on Twitter. I highly suggest following my co-founder today, his Twitter is twitter.com/all your Coovian. And my Twitter is Twitter at optoms slash Chris underbar. Underbar leu, there's 200 bars in there. Paul actually made a joke the other day that our Twitter is our Slack channel. And it feels like that many times.
Well, it sounds like it's your revenue channel as well, especially with the posts and the, what the traction you've been able to get from it. So congratulations, you've grown significantly. How many users do you have right now that are paying, if you don't mind?
Yeah, we have about 1,200 now. It might be a little higher than that grows a little bit every day, and we're very, very fortunate. But we're really excited. Every single user, you know, read to us represents, you know, a entrepreneur or someone who started growing their own business. And these are the people that we want to support.
Wonderful. Well, thanks for your time. Once again, Chris, and I appreciate it and we'll, we'll see, continue to grow and we'll keep following you so
Perfect. Thank you very much. I appreciate it. And look forward to it.
All right. Take care. Thanks. Have a good one.
Connect with us:
SEO.co // DEV.co // PPC.co // Link.Build