What’s up, you guys? I’m trying something a little bit fun here, focusing on how to make money with generative AI. I’m calling this the Anti-Fragile Entrepreneur, and it’s all about sharing stories, lessons, ideas, and information that’s going to help you succeed in building a successful business from zero to one.

I am Adegbenga, a three-time startup founder and executive coach to entrepreneurs and creative leaders. This is a segment called “What’s the Deal?” where we talk about something that’s big news in the tech world, in the startup world, and unpack what it means for entrepreneurs, with a special emphasis on how to make money with generative AI.

Today, we’ll be diving into the topic of generative AI, exploring its potential and discussing strategies on how to make money with generative AI. As we delve deeper into this innovative technology, we’ll uncover the various ways it can be leveraged to create new opportunities and drive growth in your business ventures. So, stay tuned as we navigate the exciting world of generative AI and learn how to make money with generative AI.

We strongly recommend that you check out our guide on how to take advantage of AI in today’s passive income economy.

Understanding Generative AI

Generative AI, a key concept in understanding how to make money with generative AI, is the idea that artificial intelligence can be used to create things rather than just analyze things. This comes on the back of a couple of different developments that have been going on for a while now. OpenAI, a company founded by Sam Altman and funded by Elon Musk, with a lot of money from Microsoft, has been building a series of models to generate text and images.

They have GPT, and they just released ChatGPT, which really got the internet going. A million people were using it in five days, which is extremely fast growth compared to even the most viral social products. It primarily spread through the tech community but is now reaching a lot of other communities as well, opening up new avenues for how to make money with generative AI.

The Impact of Generative AI

OpenAI also runs DALL-E, an image generation tool that has opened up new possibilities for how to make money with generative AI. You can type in a prompt and generate a lot of images. If you’ve been following the news, you’ve seen these avatars and everyone styling their avatar based on another open-source model called Stable Diffusion.

Stable Diffusion allows anybody to train a system to learn like your face and then model your face in a bunch of different ways. All of these developments have been happening around the same time, getting the tech world talking. Let’s unpack what’s really going on here and why it matters, especially in the context of how to make money with generative AI.

For a long time, the joke was that AI is really just linear modeling or very basic algorithms. The joke was like whenever you hear a founder say, “Oh, we’re powered by AI,” the VC would be like, “Yeah, okay.” Unless you have a PhD from MIT or Stanford and you got some super secret sauce, you’re not really doing AI; you’re doing some basic stuff.

Maybe you’re even handwriting some stuff. There’s really a person in the loop. I remember one of the companies that one of my friends worked at back in 2010-2012, about ten years ago, was called Aardvark. It was supposed to do what Google couldn’t do, which is answer questions like, “What’s the best sushi restaurant for couples in San Francisco?” This was not super obvious and not super available at the time, but now, with advancements in generative AI, there are new opportunities to explore how to make money with generative AI in various domains.

The Evolution of AI

Aardvark said they had figured out a way to do that using human-based search. You would identify people in this network of products, and it would ask them, “Hey, someone wants to know about the best date ideas. Can we ask you? Or do you know someone you can ask?”

There was a system that was supposed to navigate through this tree of information that people had. And you guessed it, there was actually just a person on Google Chat running this. It was not necessarily real AI. Eventually, they used that to raise a Series A round of funding, which back then was like four to six million dollars. Now you could call that a seed round or whatever. Then they did build some interesting technologies and were bought by Google. In the end, it was a success, with a 50 million dollar exit, which at the time was not bad.

You might have had an AI-powered analytics tool, and you’d say, “Hey, we noticed that there’s a cluster of customers who have these traits and shop in these ways, and maybe you should do something about it.” So the best they could do was identify patterns and present those patterns to you, and then say, “Think about doing something.” Or maybe we hand-target them, do something more with them, right?

But now, with this generative AI, it really takes it a step further by saying, not just “Here’s a pattern,” but actually “Here’s actual information.” You see so many interesting applications here, like a doctor diagnosing their patient using some of the medical history, which is very impressive. You’ve got them taking the bar exam, generating drafts of scripts. I’ve seen some really interesting movie ideas that have come out of this.

And it has history chat, so you could ask it to modify what it gave you. If you wrote something, you could say, “Make it more scary,” or “Make it more professional,” or “Make it shorter,” or “Turn it into bullet points.” It could do all those things. Put it into a table. It’s also very good with code, which I think gets a lot of technical people very excited because it can write code in different languages. It’s not always right, but it’s often close. And it makes sense because you’ve got these programmers who are building this AI for themselves or for things that they understand, and programming certainly has a large corpus of code out there.

Sometimes it’s documented, sometimes there are descriptions about how the code works. And honestly, it feels very magical because programmers understand how hard it is to write code. I’ve written a small amount of code in my life, and I know how hard it is. So it feels truly like magic for even technical people, right? You think, “Oh, I’m super smart; my job won’t be affected.” And then you type in some commands and say, “Write an algorithm that can call some API and generate these things and manipulate them,” and boom, it does it.

Not perfectly, not every time, but a fair amount of time, and it’s only going to get better. That’s the thing that’s really scary for all the people who say, “Oh, boohoo, it’s not that good,” or “It’s wrong a lot of the time,” or “It’s not very creative.” Bro, this has only been around for a few years. You don’t think it’s going to get better? I watched this video of Sam Altman talking in 2019, and when he’s on stage, this woman’s asking him, “Are you planning to make money? What is your revenue strategy?” And he basically said, “Look, we don’t know. We have no idea. We’re just going to keep making this thing better.

And we told our investors that one day we will ask it to tell us how to make money with generative AI, and it will give us the answer.” And a ton of people in the audience laughed, and he acknowledged, “Look, it sounds ridiculous. It sounds stupid, but that’s really what we’re thinking.” Three years ago, he said that. Now, in 2022, the product is still free or very low cost, but it has captured the attention of everyone, and nobody is wondering now if that’s plausible. I think it’s much, much more plausible than it was three years ago. And a bunch of these companies, okay, let’s get into the applications.

Applications and Opportunities for Entrepreneurs

Founders, like, how should you think about this? What does this mean for you, and what lessons can we take away for you? So the first is that there are two paths now. You’ve got these companies that are developing these models. Some of them are developing them in-house, like Meta and Google.

They have their own proprietary internal ones of different levels. OpenAI is one big commercial one. You’ve got others that are like open source, and who’s maintaining them? I think there’s more being done with Stable Diffusion on the image side that people seem to be using. I haven’t seen too many applications of these open-source large language models. But you have these two ways of going about it.

You either are building these models or directly creating, manipulating these models, which is pretty challenging because, again, they all have teams of researchers who are working on this. So it’s very hard to advance the core model. And the core model gets better all the time.

And so there’s even this question of whether or not it even makes sense to try to train it for specific situations, like training it for medicine or finance, because arguably the advances that we’re seeing are so strong that even your tinkering around the edge is just going to get replaced by the next version. Okay. So then it becomes really about the interface and how easy it is for an individual to actually come in here and use this.

For a long time, OpenAI had this GPT playground, and you could make a free account, log in, type some things, and press submit, and it would generate things. But there was something about the chat interface that really worked because the chat has memory. But really, what the memory is just taking everything you said before and stuffing it into the original prompt, which you could have done as a regular person using the playground. But it’s awkward. It’s a strange way of interacting with technology, and so I don’t think people really understood to do that. And yes, they’ve updated the model.

We’re still at GPT-3. They’re saying that ChatGPT is not GPT-4, the sort of system that they’re running, maybe call it 3.5. And the point is that interfaces really matter. So you see some of the companies that have succeeded, like Jasper.ai, raised like 135 million off the back of an experience built on what I believe is GPT-3.

They’re not the ones running the models, but they’ve made it possible to generate ads, to generate social media copy. And you think about the next steps of publishing, right? Okay. Write a series of tweets. Already pretty cool, but then you still have to copy and paste those tweets into the thing. So if you could say, “Write a bunch of tweets for me,” press enter, and it’s been scheduled, that’s that layer of interface usability.

Lessons from the Mobile Smartphone Era

And so the second lesson is like, there’s a lot to be learned from the mobile smartphone era, especially when considering how to make money with generative AI. When smartphones started to emerge, and people started getting smartphones, what was exciting about that was that everybody has a camera, everybody has these real-time location sensors, internet in their pocket. And so most of the companies that were built in that mobile era weren’t building smartphones. Apple was building smartphones. Google was building smartphones. Microsoft was trying to do some stuff with smartphones and not successful.

And other companies tried to do stuff with smartphones. Not successful. And it ended up having these two major smartphone creators. All the innovation was in the software, in the interface. The interface to these technologies, the sensors, the cameras, the internet, right? And so I feel like it’s the same thing here. You’re probably not going to be generating the next GPT, whatever, or the next DALL-E, the next Midjourney. And there are companies that are doing that. And so if you are doing that, you know who you are. For everybody else, it’s going to be about the interface, right?

How can you make, tie this powerful, incredible technology and actually meet people’s needs where they are using software, using hardware. Maybe people are generating children’s books where they take pictures of your kid, and then they generate a cartoon image, and then they generate stories, and then they print the book. That’s the wrapper, right? All the wrapper. And the wrapper sounds, I don’t know, weak or minimal or whatever, but it’s a big deal, right? Wrappers are important.

Just because you have technology that can do something doesn’t mean people will be able to experience it unless you build the bridge of the interface. And I think there’s a lot of opportunity for great front-end designers and programmers to create these interfaces, people who think about experiences, right? That is going to be a big part of what makes companies that leverage generative AI successful, especially when exploring how to make money with generative AI.

The Importance of Detecting and Being Human

I do think that detecting when you’re looking at generative AI is going to be really important, especially when considering how to make money with generative AI. And being human is going to be really important because the problem with generative AI is that it

doesn’t really know what’s true. It just spits out a bunch of things and seems very confident in it. And sometimes it’s wrong. Now obviously, there are things you can do even in the prompt to say, “Be as truthful as possible.” And if you don’t know, say, “I don’t know.” So those things can work, and I’m sure the models will get better at providing confidence intervals, but the problem is not everybody’s going to use that when they encounter text online. Now you have to be a lot more suspicious than you’ve ever been because you don’t know if this is generated by a computer or a person.

Why does it matter? People can be wrong, right? But the difference is that people stand behind what they write typically. So if you are a real person, you don’t want to just be spewing out nonsense because you’ll lose credibility. You could make up a fake person, which means it’ll be important to know if profiles are real if a writer is in the name on the thing. Is that real?

But wanting to know what’s human-generated and what’s computer-generated is going to be important. Someone mentioned that in the GPT system, they do try to do some watermarking. And I think this has something to do with n-gram, which is like how frequently certain words appear in typical writing.

And so they skew it in certain directions to try to identify if a large body of text was generated by GPT. But not every model is going to do that. And then you could modify it and then scramble it a little bit, and then you lose that. So there is going to be a bit of an arms race where you look, trying to be ahead of the detectors, and then the detectors get better, like spam, hackers, and other types of security systems. So we now have this new arms race, which is, are the words you’re reading real? Which is computer-generated? This is particularly crucial when exploring how to make money with generative AI.

How to Make Money with Generative AI

As we explore the potential of generative AI, it’s evident that this technology offers numerous avenues for entrepreneurs to monetize their innovations. Here are some strategies on how to make money with generative AI:

1. Create AI-Powered Products or Services: Develop applications, tools, or platforms that leverage generative AI to solve specific problems or enhance user experiences. For example, you could create a content generation tool for marketers, an AI-driven design app for graphic artists, or a personalized recommendation engine for e-commerce.
2. Offer AI Consulting and Development Services: Many businesses are looking to integrate AI into their operations but lack the expertise to do so. By offering consulting and development services, you can help companies implement generative AI solutions tailored to their needs.
3. Build AI-Enhanced Content Creation Platforms: With the rise of digital media, there’s a growing demand for content creation tools that can produce high-quality, engaging material quickly. Develop platforms that use generative AI to assist in writing articles, creating images, or generating video content.
4. Develop AI-Powered Educational Tools: Educational technology is a rapidly growing sector. Create AI-driven tutoring systems, language learning apps, or interactive educational platforms that adapt to individual learning styles and needs.
5. Monetize AI-Generated Art and Media: Generative AI can create unique artworks, music, and other forms of media. You can monetize these creations by selling them directly, licensing them to others, or using them to attract traffic to ad-supported websites.
6. Provide AI-Driven Market Research and Analysis: Businesses are always looking for insights into their markets. Develop tools that use generative AI to analyze market trends, consumer behavior, and competitive landscapes, and offer these insights as a service.
7. Create AI-Enhanced E-commerce Platforms: Use generative AI to personalize shopping experiences, generate product descriptions, or recommend products based on user preferences. This can increase sales and customer satisfaction for e-commerce businesses.
8. Offer Subscription-Based Access to AI Tools: If you’ve developed a powerful AI tool, consider offering it as a subscription-based service. This can provide a steady stream of revenue and attract a wide range of users, from individual professionals to large corporations.

By tapping into these strategies, entrepreneurs can harness the power of generative AI to create innovative products and services that meet the evolving needs of consumers and businesses alike.

Embracing the Future with Generative AI

I think that’s going to be important, especially as we explore how to make money with generative AI. I believe that being very human is essential in this context. So none of this was written by an AI. I’m talking too fast to read off a script, frankly. People just aren’t, unless you are like a professional telecaster, able to read off text super fast. And I don’t know, I just wouldn’t do that. But there are companies that are trying to generate avatars and make videos seem more realistic. However, I think it’s going to be a lot harder to make a face look realistic. You’re not going to be able to generate that very easily.

So I’m going to have to start inserting ridiculous things like that to just make sure that you know that it’s really me. It’s I really recorded this in a real room, in a real location. This is information coming out of my brain. So we are entering a pretty strange world. But it’s exciting, right? This is the future. We’ve seen a lot of lack of groundbreaking innovation for the last couple of years.

I feel like crypto really tried to be that, wanted to be that, but 12 years in, there just hasn’t been anything significant that has actually changed the world. And I feel that this technology has the potential to do so, especially when considering how to make money with generative AI. So those are my three takeaways on generative AI, why it matters, and how to make money with generative AI.

We strongly recommend that you check out our guide on how to take advantage of AI in today’s passive income economy.