What a brilliant marketing strategy that you send your people out or they leave to go to other companies and then they buy your products. That's so smart. So to me, I learned from that: how do you release people and keep them close to you? You're generous, you're kind about it, you don't perp walk them out.
All right guys, we got David Steinberg here. We are out here in Vegas for CES. It's going to be a fun week. Yes. You ready? You go to CES every year?
I am ready. I wouldn't say every year, but it's been a lot.
What are you looking at? What's your eye on this year? Any trends?
No, you know, listen, everything right now is about artificial intelligence, right? It sort of really exploded with the launch of ChatGPT last fall, and we're now seeing it really ripple through every technology, and you're going to hear a lot about it this year.
Yes, I'm excited because I love to learn, and ChatGPT has made learning like 10 times faster for me.
I agree. Well, it's certainly made getting answers 10 times faster, right? I'm summarizing books because books take a while to read, let's be honest, and like 80% of books are kind of fluff. There's really only like key messages within books in my opinion. I learned when I was a kid: read the first paragraph of every chapter and the first sentence of every paragraph. And I just started reading books like that when I was much younger, which sped it up for me. The funny thing is, I would bet you that if you went into the average, you know, ChatGPT query and said, 'Summarize this book,' it will literally give you a lot of that.
Wow, I need to try that. Yeah, well, I took the last week off of work and I was just doing personal development and I summarized 50 books.
Isn't everything you're doing here today personal development?
Show pretty much, yeah. I mean, we're developing together, yeah. But yeah, I summarized 50 books on ChatGPT that would have taken me a full year to read.
So I got all those key points and feel a lot smarter, I guess.
It's amazing. I love it. I'm using Gemini inside of my Google. I'm using ChatGPT on the regular. I mean, of course, our company develops, is one of the largest developers of artificial intelligence, and we started programming in AI seven years ago, not seven months ago.
How do you have that foresight seven years ago?
So I read a lot. I, you know, I read hundreds of different things a month, and I do what I call triangulation. If you can take three separate points, triangulate it to a central theme, I believe that theme is going to change things. And about eight, nine years ago, I started reading. We were already working heavily in automation at Zeta, and we were thinking about machine learning, deep learning, and we'd always been sort of a big data platform. So all of this data, no human could process it anymore, right? And as we started thinking about the future, we really started focusing on natural language processing, the ability to ingest trillions of data points and convert them into intent, you know, what do individuals intend to do next? And it was funny because when we took Zeta public in June of 2021, we had a big sign.
You put a sign on the side of the New York Stock Exchange, right? It's a thing to do.
Right. And our sign said, 'Data plus AI equals intent.' In June of 2021, put it in perspective, people were asking me, 'Who's Al and why is he in charge of your data strategy?' Right? It was that foreign to people even three years ago.
That's fascinating. Well, they were probably working on OpenAI for years, right, before it dropped.
Yeah, yeah. They'd been working on OpenAI for many years. But listen, AI has been around for a long time. The big leap was truly the launch of ChatGPT. And I look at it and I think we're going to look back 10 years from now and say this was the Renaissance moment for artificial intelligence. It's when AI went from science fiction to boardroom conversation.
Yeah, I'm very excited about it. I get companies offered to buy my raw podcast episodes because they're training their models to learn from podcasts. Think about Joe Rogan's episodes, if you could learn in like 20 minutes the key takeaways from it.
Oh yeah. And then by the way, it's not just that 20 minutes, it's the themes. So a true algorithm two, three years from now will not just say, 'Here's a summary of Joe Rogan's podcast,' it'll be, 'Over the last 10 years, here are the five podcasts you need to listen to to become an expert on this topic.' And it'll topicize it. It'll ingest all of the information from 10 years. By way of example...
Shout out to today's sponsor, Specialized Recruiting Group. Navigating the professional job search is hard. You know the perfect job is out there, but you're not sure how to find it. The good news is you don't have to go at it alone. You just need Specialized Recruiting Group. You can connect with them at srgpros.com. They're here to guide you and help you find a role that fits, all without costing a dime. They're offering a tailored approach to find your next role at srgpros.com and help you get on the right course. Your local Specialized Recruiting Group team knows which businesses are hiring and can offer you a path to contract and full-time roles. If you don't see the role you are looking for on their website, Specialized Recruiting Group also recruits for confidential roles, so give an office near you a call to learn more. Take the next step in your career by starting at srgpros.com.
By way of example, it'll sort through all of it and it'll build themes from a curation perspective that will allow the algorithm to not just curate one episode, it will ultimately be able to put together the same theme from large numbers of episodes with one output that you can ingest in 20 or 30 minutes.
Wow. So AI is going to revolutionize the education industry, though.
So where do you see that changing to over the years?
Well, I mean, listen, education is complex because people learn differently. When I was a kid, to date myself, you weren't allowed to bring an HP calculator to class. You don't even know what an HP calculator is. I say that not in a pejorative way, but because of your age versus mine. It was like the first sort of computer. It was a calculator that could build formulas. I come from a technical background, so sort of like that. When I was a kid, my kids, my older kids, I have three older kids, two younger, my older kids bring their laptops to class now. And it's just the evolution of education. You're not in five years as a professor going to be able to say, 'Write me an essay on this,' because the essay will all be automated. They're going to have to come up with more complex learning and more testing, which will be in person, so the person can prove they're actually taking the test versus using an AI.
Yeah, a lot of teachers are dealing with cheaters now, right? Well, there was this kid at Princeton University. Did you see this story?
No. So my theory on AI is you're going to have people using it for stuff they shouldn't use it for. My counter theory is we're going to have algorithms that protect us and algorithms that say whether something started its creation process with AI or with a human. It's all coming. This young person at Princeton, I shouldn't say kid, wrote an algorithm that allows a professor to figure out if a paper had its creation with OpenAI. He's now left Princeton, raised a bunch of money, and he's turning it into a business. And you know, somebody asked me the other day, how is AI going to affect encryption? Right? Because one day quantitative computing is going to match with AI. In fact, not to sound like a total name dropper, recently I was talking to Elon Musk about this. And Elon and I were sort of riffing on how is quantum and AI going to come together. And I believe that the creation of quantum computing coupled with artificial intelligence will get us 10 orders of magnitude greater than we are with AI today. But what's going to happen is as soon as that launches, every single piece of encryption, every passcode you have today is crackable in seconds because quantum computing, which will think on a nonlinear basis, will allow it to crack any code. So we're going to need to create algorithms that protect us, and it's going to be sort of a if-then, if this happens, then you're going to have to have that.
That's scary as someone in the crypto space. If your seed phrase gets leaked, that's your whole bank account. They could send it in a second.
Well, once again, you're going to have an algorithm that will protect your crypto.
Yeah, so platforms need to start looking into that, right?
They're already doing it. It's interesting because so often technology is adopted by people who want to do stuff with it that we might not like, right? More nefarious stuff. And often it takes a while for the tech to protect us to come out. I think we're going to see a bit of a paradigm shift here because I'm already seeing organizations building algorithms to say whether something started with AI or not, and I think you'll see a very similar thing with encryption.
Elon's very concerned AI is going haywire, right? Were you talking to him about that?
We got into it a little bit. The truth is we were talking more about, you know, getting to space and the election, and really focused on AI, which is what I'm really focused on and what we do with Zeta. But you know, he certainly has spent meaningful time talking about how he believes that if we don't have gates on it, we're going to have a challenge. Now, of course, he's got xAI, and then you've got OpenAI, and you've got, you know, sort of everybody out there.
He's got Grok too, right?
I haven't used Grok yet. Have you used Grok?
Not personally. We've played around with it. We do a lot with OpenAI. We do a lot with Gemini and with a number of the other algorithms. We at Zeta have our own algorithms which we built on top of Python, although I joke they're midsize language models.
I'll tell you what that means, Sean. So a lot of people talk about large language models, which is effectively the whole internet. Then you get into small language models, which a lot of people are talking about as the future, right? The ability to really hyper-build and allow the algorithm to explore. At Zeta, because 245 million people in the United States are in our data cloud, and because we're ingesting a trillion pages of content a day, I joke that's a midsize model. We're not ingesting the whole world every moment of every day because, quite frankly, for our enterprise clients, what we're really focused on is helping them more efficiently run their marketing. The way we do that is by making sure that we only show their ads to people that are interested in their products, but also will have a propensity to want to buy their products and or be able to be credit approved, right? Like, you know, companies show ads all the time to people who won't get credit approved for the products. It's a waste of money and it's going to piss off the customer. So we try to eliminate that prior to our software deploying their marketing for them. We do that by ingesting all of this information. We build what's called a CDP, or consumer data platform. We ingest all of their data and we match to the 245 million people in our data cloud. We then ingest between 5 and 7,000 incremental data elements in addition to pulling in a trillion pages of available content into the algorithm. And then there's no data exhaust. So by way of example, all of this stays proprietary to our enterprise clients. Enterprises are very worried about their information sharing back to other models, right? Our platform ensures that doesn't happen. And then by doing that, we're able to lower their marketing cost by about 50%.
Huge, big, especially for big companies. 40% of the Fortune 100 largest companies use our platform today.
Wow, that speaks to the product, right?
That's a great product, especially with all the strict regulation on marketing with Facebook and Apple now. You solved a really big problem.
We do, and we plug into both those guys. So we think they're both amazing companies, and our Zeta ID number matches to pretty much any available ID for walled gardens and the open web.
Wow, I need to start using your product.
Well, but we primarily focus on very large enterprises at this point.
That makes sense. Yeah, yeah, people that are spending millions a day, right?
You might have to hit up Gary Vee if you're not working with him yet.
I know Gary. Gary's a great guy. Gary runs a really great agency, and they're doing really interesting things.
Yeah, damn, that's super exciting. So Zeta is public?
Zeta is public, yeah. We trade on the New York Stock Exchange, and it's been an interesting journey.
Nice. And that's your second big company, right? You've had another?
That's my seventh company. I've sold four, taken two public, and then I'm chairman of another.
Sold four. Wow. Taken two public, sold four. Well, how old were you on the first one?
I started the first company when I was 21.
Wow. Yeah, I didn't know any better. Yeah, right out of college, right?
Yeah, right. I came out of school. I went to work on the Hill for about eight months. I read about this whole new thing called wireless back in 1990, and I started a series of wireless companies.
Wow. So you attribute a lot of your success to reading?
You know, I would tell you that reading is a very important component of how I've built the strategy for the businesses that I've wanted to start. And then, you know, the single biggest differentiator in what separates winning businesses from losing businesses, in my opinion, is execution. It's probably the most undervalued and most important component of any business. So, you know, the good news is for a dyslexic like me, I've been able to surround myself with some of the world's greatest people. And I bring in operators who can run their components of the business substantially better than I can. And I do something that's hard for a lot of entrepreneurs: I let them do their jobs.
You're not a micromanager.
You know, there's this whole founder mode thing, and there are times I go into founder mode. I'll go very deep if there's an issue on something. But you know, and I know the metrics of our business as well as anybody, but I try to focus on the things I'm really good at, and I try to hire the world's greatest people to do what they do best.
Yeah, when you study the top companies, it seems like they got a visionary like an Elon, a Bezos, someone like you, and then they have operators around them.
Yeah, and my partner Steve Gerber, who's our president and chief operating officer, is the best operator I've ever been with. My CFO, Chris Greiner, is the best CFO I've ever worked with. My general counsel, Steve Vine, my, you know, CTO, Chris Schaengold, our chief data officer, Nate Gore, these guys, each one of them could run their own companies at this point, but they choose to be a part of our collective.
I love it. Did you self-fund at first?
Yeah, I did. So for this company, I put up the first capital, which is one of the reasons I think I'm the control investor in the company today.
Wow. Yeah, because a lot of people are quick to raise these days.
