Tom Frazier is the Co-Founder and CEO of Redivider. He boasts an impressive 25-year career, driving transformational and disruptive architectural initiatives in future tech, B2B, and public sectors.
As a serial entrepreneur, Tom founded Redivider based on his extensive knowledge of data centers and a passion for revolutionizing the industry. Committed to prioritizing people, the planet, and profits, Tom is devoted to spearheading innovation in the digital economy.
It’s estimated that globally, data centers use two percent of the world’s electricity. The US averages 90 billion kilowatt-hours of electricity each year according to Energy Central.
In fact, they can generate the same amount of carbon emissions as the airline industry. Specialized computing demand is also growing significantly faster than energy capacity. Meanwhile, the data centers themselves have become huge facilities that do little or nothing for the local communities in which they reside.
Redivider aims to empower communities by bringing energy and investment that deliver prosperity (health, wealth, and education) to areas designated for economic development. This includes hiring and training new talent as well as offering them stakeholder opportunities in the company too.
Having served as a cloud and security strategic director for a Fortune 10 company, Tom has been instrumental in securing some of the world’s largest digital footprints.
He started his remarkable journey as the youngest faculty/staff hire at his university.
Tom holds a patent for a “System for Processing Customer Records” and obtained a Bachelor of Science degree in Information Technology from the Rochester Institute of Technology.
Outside of his professional life, Tom enjoys spending time at his hobby farm, traveling the world to teach dance, and cherishing his roles as a leader, teacher, husband, and father to his two young children.
What you will learn
- How Tom’s passion for technology began
- The inspiration behind Tom’s business, Redivider
- What is edge computing? Tom unpacks it for us
- Tom shares how he got his very first customer
- Discover the marketing sales process behind Redivider’s success
- Learn how Tom makes data centers sustainable
- Tom shares the impact AI has had on his business
- Discover how AI helps develop human creativity
- Find out more about Redivider’s unique selling proposition
- Learn why sustainability and social impact are incredibly important in business
- Plus loads more!
00:00:04 - 00:01:44
Hi everyone, wherever you are around the world, welcome to The Jeff Bullas Show. Today I have with me, Tom Frazier. Now, Tom is the co-founder and CEO of Redivider. He boasts an impressive 25 year career driving transformational and disruptive architectural initiatives in future tech B2B and public sectors. As a serial entrepreneur, Tom found a Redivider based on his extensive knowledge of data centers and data centers are exciting apparently but that's what we're gonna find out from Tom about what he's doing is a passion for revolution, the industry with a blank sheet of paper. He committed to prioritizing people, the planet and profits. Tom is devoted to spearheading innovation in the digital economy. It is estimated that data centers globally use 2% of the world's electricity. The US averages 90 billion kilowatt-hours of electricity every year according to Energy Central. In fact, they generate the same amount of carbon emissions as the airline industry, didn't know that by the way, specialized computing demand is also growing significantly faster than energy capacity. Meanwhile, the data centers will help become huge facilities do little or nothing for local communities in which they reside. Tom and Redivider aims to empower communities by bringing energy and investment that deliver property health, wealth and education to areas designated for economic development. This includes hiring and training new talent as well as offering them stakeholder opportunities in the company too. He served as a cloud and security strategic director for a Fortune 10 company. Tom has been instrumental in securing some of the world's largest digital footprints. Welcome to the show Tom, great to have you here.
00:01:45 - 00:01:48
Great to be here. That was quite an intro. Thank you.
00:01:48 - 00:01:52
Well, that's what your PR agency sent through to me. So I just had to read it, didn't I?
00:01:53 - 00:01:54
00:01:55 - 00:02:15
So, Tom, you've been in a variety of different, I suppose, activities and projects over the years. So you said you're a bit of a tech nerd. So where did the tech nerd-ness come from originally?
00:02:16 - 00:03:00
You know, I don't really know, I went to apply to school to go to college at the Rochester Institute of Technology and, you know, I thought computers were cool, you know, I remember buying my dad a one megabyte chip for his computer when I was in high school. I was so proud of that moment. And, you know, I found all these people at RIT and there's this group called Computer Science House, which is like a frat for nerds basically. And just seeing what these 18 to 20 year old kids were doing was mind blowing, you know, this is like mid 1900s and I just got hooked, I can't not do it, you know, it's a compulsion now to stay very close to the leading edge of technology.
00:03:01 - 00:03:30
So you did your technology degree effectively and what was your, so what was your first and you're learning along the way, obviously, I'm intrigued by tech, I can't program but maybe AI is going to be able to help me. So what was your first project job after you did your technology degree? What or what was your first entrepreneurial gig? What did it look like after you left, you know, UNI college?
00:03:30 - 00:05:31
Those are two very different phases. I have a very atypical path. I was actually hired as faculty staff in my university when I was 18. So after my first year in UNI, I found myself with an office right next door to all my teachers. And, you know that saying, you don't want to see how the sausages made. Well, in the mid 1900s, the teachers were learning the technology about an hour before they were coming to class to teach it because it was going so fast and, you know, it's funny you mention I feel like we're in that same exact moment of the mid 1900s. We're there now again with AI, but, so my first kind of career arc was, you know, working, doing IT stuff at the university, you know, running a lot of the unique systems and doing some of the networking stuff and I kind of got hooked on this one project around security and that sent me down this huge path for, you know, a big portion of my career doing IT security, information security, yeah, for some of the largest critical assets on the planet. And, you know, one day I found myself living in Sydney and, you know, I kind of got to kind of the top of the field and I looked around and I thought I remember this very, very specifically, I got to the top of the corporate ladder and I had this realization that I climbed the wrong ladder, you know, I didn't really enjoy working for a big company. Even though I learned a lot along the way and then I started my entrepreneurial journey, you know, where it's run it as cheap as you can, bootstrap it as long as you can and, you know, did another one and another one. And, you know, I haven't looked back since and really was taking all those lessons, insecurity about breaking down technology into smaller pieces. And I just applied that to businesses doing the opposite. Now, let's take these building blocks and use them to build up a company.
00:05:32 - 00:05:58
So Redivider, Where did that inspiration come from? What was the call to start this company that specializes in a sustainable specialized data center? Where did this idea come from? Was it over a beer in Seattle or Sydney or does it, you had too many drinks that day? I don't know. What, how did the idea emerge?
00:05:59 - 00:07:54
Well, I think it came from, you know, a career of experience in the field where you kind of know how the industry works and how it moves. And there were four of us that founded the company and we all got connected around this idea that, you know, good people create karma and when you surround yourself with people like you tend to attract more and repel the ones who aren't like you. And so a couple of us are surrounded by what was going on. And the light bulb moment was during COVID when COVID first happened and it, you know, in my brain, you can visualize this, close your eyes and think of a huge data center like a million square foot data center that is a really big internet pipe that connects to a big downtown building where everyone worked and COVID broke the remote work thing wide open. And now you have people working everywhere but the structure of computing and the structure of the internet wasn't really designed for that because it was designed for these big data centers, big pipes, big buildings. And that's kind of where the light bulb moment happened. It's like, you know, what edge computing has been talked about for 10 plus years and now is the moment that the industry is going to be forced into this direction. And so without having legacy infrastructure, without having a vested interest in owning, you know, millions of square feet of data centers, it allows us to dream the future that we want to deliver. And for us when Redivider that future is to uplift humanity through computing. And you know, how do we take something that is this baseload consumer of power, this generation center for data and how do we use it to make the world a better place and that's kind of what we've come up with.
00:07:55 - 00:08:07
Okay, cool. So just before we move on to what that means, you mentioned the term edge computing, a lot of people wouldn't understand what edge computing is. Can you explain what edge computing is?
00:08:08 - 00:09:00
Sure. I mean edge computing the way to think of it is a data center that's really small. It could be as small as the top of a light post at a traffic stop. It could be a closet in a hotel, it could be a purpose built building, that's the size of a shipping container or, you know, think of it like a spectrum. So it's not a large building that houses thousands and thousands of computers. The edge is trying to push the computing power closer to the source of the data, right?. So today, typically when you think of the data center, you're moving data, it's getting computed and then it's moving back, you know, like through a 5G connection or something. And edge just means you're pushing that computing closer to where people are, are using it.
00:09:00 - 00:09:16
So and making it as efficient as possible instead of having to have it like in the center of the USA powering everything at the at, you know, everywhere, all over the world or whatever. So in other words, you're put in computing where it's needed, right at the edge.
00:09:16 - 00:09:19
At the decentralized. That's right.
00:09:19 - 00:09:42
So distributed. Alright. Cool. Right. So you want to make it more sustainable. So at Redivider, you said you started with a blank sheet of paper, okay, and what are some of the core elements of what you do with your specialized data centers? What are some of the things that are really core to what you do?
00:09:43 - 00:11:21
Great question, you know, the core thesis is trying to insulate against risk. I think that's the number one thing when you look at the place we are now versus the place we started is how do we design a business that's going to grow for 20 plus years in a way that has the lowest possible risk. So we really focused on doing them as prefabricated facilities. So our data centers are made in a factory and doing them in a factory allows us to do a whole range of things so we can have a library of different modules. So if you're a movie studio who wants to rent a data center for a year, it might look one way with two modules or if you're a regional Telco with a 5G hub, it might look totally different, but there's still modules inside of the library. It also allows us to drive out a lot of unnecessary costs. So those two examples, you would have different design constraints and affordances that you would marry it to the used case. So we're really focused on that specialty computing need, making sure that our data centers are a lower cost, higher quality and by making them in a factory, we can produce them and deploy them at a scale that's extremely fast compared to a, you know, a two year project to build a huge building and do all this other, you know, project plan work that goes into making a large facility, you know, we can pump these out on a much faster schedule. So by the time we hit the same time period as a traditional facility, we've had a bunch of it operating along the way.
00:11:21 - 00:12:06
Right. So the thing that pops to mind for me is, okay, so you got chicken egg scenario here is like you got to go and find a customer who's gonna believe in you and your vision because you are not a trusted brand because you want people to trust your brand because they are relying on you for their data and computing, which is pretty much the the hub and core of any business these days. So how did you get your first customer? And how did that look like in terms of creating that trust, building it, did you build it first then chase customers or did you chase customers and then build it? How did you first customer look like?
00:12:07 - 00:13:11
Yeah, starting a data center company is not for the faint of heart, you know, these are not small projects and, you know, they're pretty monumental. So what we've spent, you know, we probably had the idea what two in a half, three years ago, we've been going for two years and all of that has been in preparation for, you know, where we've just come from, which is how do you build this thing that does what we promise, right? Super high scale, high sustainability, high social impact and, you know, that's largely setting up the manufacturing and the supply chain and all those things, you know, so in terms of getting the customer, they want to see all of that has been done, right? We're not at the idea stage, you know, we have manufacturing capacity, we have the carbon transparency platforms, we have all, you know, all the bells and whistles to do the things that we say we're doing. So that's really what it took getting all of that done in a way that can scale to the size of customer we're taking on board.
00:13:11 - 00:13:20
Right. So you went out and leased a building to do the prefabrication, basically you're doing. Is that what you did?
00:13:21 - 00:14:53
Well, that was our initial plan, you know, again, it's all about iterating for everyone's benefit. And I love to say that people should focus on the problem, not their solution and you end up with such a better, stronger business. And so my partner Eric, he's in manufacturing, he's a third generation manufacturer. And, you know, initially our plan was, yeah, we can just go set up a building and have them, you know, as assembly line process to manufacture these things and get all the parts and supplies and workforce and in the process, what we found was, you know, there are only a handful, but there are a few partners that exist in the world that can help us get to where we need to be faster. And so when we did the math, we're like, alright, if we do it on our own, we're gonna be another couple of years and we really shorten that cycle by at least 24 months by partnering with somebody who can do as good or better job than we could do in house. So that part, they're running our library, our design work. But when it comes to the manufacturing, you know, they already have lean Six Sigma processes in place, they already have lean manufacturing, they already have, you know, supply chain and, you know, made in America for the steel, for example, and how that can tie into our, you know, carbon footprint and have the transparency around that. So for that particular part of the business, we definitely chose to despite our own internal abilities to do that with the partner.
00:14:54 - 00:15:15
So because you're doing something quite specialized. So how do you find a partner to do what you wanted? I'm intrigued by that, like modular data centers aren't sort of just around every corner like a pharmacy or so, what sort of manufacturer did you find to do this?
00:15:16 - 00:15:58
Well, overall the, you know, when you're a mission driven company, you know, our mission is to uplift humanity, right? When you are mission driven, you surround yourself with A+ players in every domain, you know, that's the only way you can get a huge ball to move is like tons of experts. And, you know, so whenever we would look for partners, we would look for people who were the absolute leaders in domains. And we found one in the manufacturing space and frankly, they could do it better than we could do it out of the gate. And, you know, like I said, it's a small industry that I've been in for a very long time and, you know, we found the right partner to do that part.
00:15:59 - 00:16:15
Okay. So because you've been in the industry before you knew where the different state, where the different partners that you needed to come together. So, alright. So, he was your first customer? I'm intrigued by that as well. Can you reveal who that is?
00:16:16 - 00:17:23
Well, our first customer was ourselves. So, you know, we have other customers, but the very first one, which was, you know, over a year ago to dog food the business, you know, no one finds design flaws like when it's your business, you know, that and dog fooding is the best way to do that. So that was kind of our initial path and now we're kind of well beyond that and scaling for, like I said, it's larger companies. Our kind of position is to find the organizations that are very aligned with what are called SDGs or Sustainability Development Goals under the United Nations. And for companies that have made public statements around, we have these goals for sustainability or, you know, net zero or all these smart city initiatives. And that's really where we're getting, you know, the traction on the customer side. So, we're not kind of competing head to head against other data center solutions. Because that's not what we're trying to do. We're trying to have this other mission driven component that sits higher than the data center itself.
00:17:23 - 00:17:32
Okay. So you've got obviously a target market that's an SDG driven corporation or organization.
00:17:33 - 00:18:39
Not necessarily an ESG corporation, but somebody who wants to, large enterprises are multifaceted in their agendas and all of them have key risks in their business. So one of the key risks is access to power, access to data, manipulation of data, you know, how much does data drive their profitability. And so a lot of companies who have those as key risks also tend to align with some of the sustainability initiatives of the United Nations. Because like one of our core values is people, planet then profit. And you know, so many studies have shown that when you focus on sustainability, the net result of profitability is significantly higher. And so that's the kind of customer we're focused on. We don't want to take the political angle of ESG because that means a lot of things to a lot of people. We just think that by focusing on sustainability and social impact through data centers, we can create more profit for our customers.
00:18:40 - 00:19:18
So trying to find those customers because that's a challenge for any startup business. So what's your marketing sales process look like? So I'm intrigued by that because these are not like pick up the yellow pages and start at A and then call through to Z. So how do you find your customers? What, you know, what's your sales marketing process? I'd be intrigued in other words, from interest to sale. In other words, how do you get their interest? Is it a sales team? Is it telemarketing? What is it?
00:19:19 - 00:19:48
Yeah, it's, I mean, it's a multitude of things. So basically, it's a network effect methodology, right? I've been in the industry a long time. A lot of people that we have around us have been in the industry a long time. A lot of our advisers have been advising boards of major companies for a long time. And so we are, you know, we built the machine that we wanted to have and that has given us this network effect in the industry to go get those customers.
00:19:48 - 00:20:16
Right. Okay. In other words, basically the whole network you had before you started comes together. So in terms of your, the model for, you know, Redivider, is it a monthly subscription? Upfront capital deposit? How does it work in terms of a customer model in terms of revenue model? How does it work for you guys?
00:20:17 - 00:21:42
Well, we decided to not change that too much from how customers buy things today. Because if the proposition is, you know, there's a data center is, they're tiers of data centers and you have a workload that runs in them. Well, if you're gonna shift your workload to our data centers because, you know, higher transparency or it's closer to your facility or there's higher security, you know, all these benefits, we don't want to then change the buying process as well, you know, because that's easy for them to understand already because they're already a consumer of compute.
So that is largely derivative of what the industry does today, right? You're gonna buy space and power. In some cases, it's called bare metal where we can take, we'll buy all the computing power and you can lease it from us. But typically these are kind of longer term customer contracts, you know, we're not going after someone who wants to lease a server, right? We're going after someone who wants to lease, you know, a portion of an edge footprint across 200 locations, right? So think of content delivery or a cloud vendor that has an edge program, they want to take a portion of our entire edge. So those contracts are pretty standard in the industry.
00:21:42 - 00:21:56
Okay. Yeah, I'm not familiar with, you know, data center industry models really. It's but just so in other words, you, it's more the value proposition you're selling rather than changing the business model. Is that right?
00:21:57 - 00:22:38
Yeah, it's globally, not only does the data center industry take about 2% of all electricity, but the vacancy rate in data centers is almost zero. So the idea that we're bringing capacity online and we're doing so in a way that focuses on sustainability, social impact edge all that kind of other value proposition just means we're going to attract just like how, you know, we got together as a group, we're intending to attract to the people that see what we see and believe in what we believe in. But the buying process is pretty standard for the industry.
00:22:39 - 00:23:10
Okay. Let's dive a little bit more into the different, I suppose the value proposition you offer such as you talked about quite often, data centers don't bring jobs to a local community. You're talking about sustainability. Let's talk about some of the key elements of your value proposition that you're selling as part of your sales process. What are those like? Let's talk about sustainability. For example, how do you make your centers more sustainable than the next data center?
00:23:11 - 00:24:27
Great question. You know, the first and foremost, we're fabricating our facilities. So we have the ability to make the design choices that have the best sort of sustainable option for that facility, right? So a lot typically in the data center space, you're taking a building like a commercial building and you're making it a data center, right? So it might have been made with concrete or it might have been made with steel or might have been, you know, a wood stick building. But you didn't have, you didn't, you weren't given those design choices because you're just trying to find physical space. In our case, we're creating that physical space. So we can build that into the model that we have. And because we're building it in, we also can do accounting for what sort of carbon is embodied into our facility, how we're going to do sustainability on our energy production, our energy consumption, how do we meter all of that. So when you start with that blank sheet of paper and you think about these things, you can really differentiate against what everybody else has because they're largely integrating into somebody else's footprint.
00:24:28 - 00:24:53
So I sort of have this vision of you taking like a shipping container type look and putting in someone's, I don't know, where do you put these prefab data centers? Where, what does it look like? I'm trying to get a picture in my mind of I say not shoehorning a data center into an existing building. You are building, you're actually creating the space. Tell us a little bit more about that.
00:24:54 - 00:26:42
So think of it as like modules, right? So for example, our smallest footprint of a facility, you know, maybe 10ft by 10ft, right where it's still prefabricated. It's a purpose designed building like prefabricated unit that's really small that could go inside of another building. So you might be a manufacturing company that you're going crazy with IOT and sensors and telemetry data and computer vision and you don't want that data to go to the cloud for security purposes or competitive advantage. We can build something from our library that we can just shove inside of one of your existing manufacturing facilities. But the example you gave of a shipping container is a very apt example of like a small data center. It's, you know, roughly a shipping container size item, it's not a shipping container, but it's the same size. And then because they're modular, you can also stack more modules together. So instead of having something that fits on one truck, you might have it in seven pieces and then you ship them to the site and you join those seven pieces together. And then a year from now you want to double the size tack on a bunch more modules. And now you have a larger data center and because they can fit on a truck or a train, you know, we can send these facilities to anywhere they need to go. And because they're prefabricated, we can make them at massive scale. And so when you combine those two things of massive scale and go anywhere, we have the kind of unique ability to push computing closer to where it's consumed.
00:26:43 - 00:26:52
So that would typically be like a specialty or a manufacturer and you put it inside their existing facility, for example.
00:26:53 - 00:27:57
Yeah, a few more examples, right, a smart city initiative, a town is going to try to solve their traffic problems. And so they're gonna put cameras everywhere on top of the light poles and use computer vision to figure out how they can manipulate traffic such that it brings the quality of life up for a city. I mean, just one example, we can put a data center in that smart city so that, you know, for computer vision to work properly. It's, you know, a lot of data and so you don't want to shuttle that all to the cloud somewhere far away. You want that to be very close I mean, and another example would be training an AI model where you don't care geographically where it is. You want it to be the cheapest cost or you might care about the CO2 footprint of how that data gets moved around as a key initiative like for the airline industry or a company making an animated film who wants to rent a data center for a year and then they don't need it again so we can drop it off in their parking lot, pick it up a year later.
00:27:58 - 00:28:33
Right. Okay, cool. So you mentioned AI, so let's lean into what the implications for AI and we'll have a little chat about AI for business because I know it, we both find it very much of interest as a tech nerd. You are leaning into it and got a story about helping children learn. So how's AI impacting what you do? Like is this a gift from for you in terms of AI. Is AI a gift for you guys as in right time, right thing at the right time?
00:28:34 - 00:30:18
AI will, this is a very big question. So yes, it's good for us in the sense that it is driving the intensity of computing requirement through the roof. But the prior to, I mean, ChatGPT exploded this whole thing. But prior to that, the growth in data is already exponential. And when you think about data from a consumption a storage standpoint, there was already a massive demand for the idea of Redivider and our edge facilities. AI has changed that dramatically again because now everybody and their brother wants to use it as a competitive advantage and have their own custom AI models. So now everyone's training AI models and they all need this sort of computing power. So it is dramatically accelerating the computing industry for sure. How much of that is gonna go to edge versus the traditional facilities? Time will tell. But I would assume edge would get at least 50% of that market. Because you're gonna have two camps, you're gonna have the big guys like Microsoft and Google who do it all in their data centers for their customers as a service. And then you have all these people who do it more privately for their own benefit. And that group, they're not set up to have this style of data center. They're used to having a couple racks and a couple clauses with servers for email. And now you're talking about these, you know, crazy large infrastructure for crunching these models. So I think over 50% of the compute power would probably go to the edge.
00:30:18 - 00:30:32
Right. Okay. And the other thing that I'm intrigued by is that the challenge to an AI uses a different type of chip generally, which is a CGU rather than a CPU. Is that correct?
00:30:33 - 00:30:37
GPU instead of the CPU.
00:30:37 - 00:30:53
In other words, in other words, the chips that drive gaming industry, which is NVIDIA is the big players in the USA. How are you going getting supply of this? Because I've heard that it's a bit hard to get supply of these. Is that impacting you guys?
00:30:54 - 00:32:05
Not that doesn't directly impact us too much because that's really up to the customer to decide how that works for them. But I will say it is the most shocking thing I've seen in a long time that NVIDIA, you know, basically the sole source vendor for GPUs at scale, which is why you're seeing everyone come out with their own. I mean, Meta has stated they're coming out with their own AI chip, Apple, Microsoft, all these major players have realized we can't depend on a single supplier for all of this stuff. And I think that plays into the larger picture of kind of deglobalization that we're seeing to because a lot of these chips are made in politically. I wouldn't say misaligned but there is counter party risk at a geopolitical level. And I think that's where we're gonna start seeing a lot of things deglobalized into different countries, including the United States and you've seen the Chips Act and all the stuff that's happening to bring chips into America and our national security. And, I think that trend is gonna be the mega trend for the next generation.
00:32:05 - 00:32:28
And some of the investment in that is just huge and I've heard that a lot of companies are looking because of all the rebates and that are being offered by the American government. There's a lot of international companies saying, well, let's do this in America and I think it's like green initiatives and so on. So that must be quite.
00:32:28 - 00:33:29
But I will say, yeah, sorry to interrupt, I will say that in what I said before about all these companies that are having their own privately trained models and, you know, they're doing it for their own competitive advantage. There is a lot of advancement now to do that with CPUs instead of GPUs because a lot of these, what's called an LLM or Large Language Model, these are the do everything kind of models, right? And if you're building one specifically for FinTech, right? Bloomberg just announced they have their own language model, they don't need 90% of what you would get in ChatGPT to drive their competitive advantage. And so there is this amazing transition now back to CPUs for people who are doing these specialized models. And I think that's gonna ultimately be okay for the industry that we won't depend on a single source supplier like NVIDIA, even though they're great, nothing bad to say about NVIDIA.
00:33:30 - 00:33:48
I'm in the tech industry but I'm not, I suppose a programmer or anything of that sort of level. So I'm more just watching the business implications of this. So, tell us, what do you know about Bloomberg? Because they're basically a publisher, aren't they? Is that correct?
00:33:49 - 00:34:15
I only know what's been in the news that they've done their own language model and they have a unique data set, right? So the people who are going to win in the AI race are gonna be those that commit to computing resources and doing it in a competitive way with their own private data set. And I think that's exactly what Bloomberg is doing. And I think it's, you know, I think it's powerful time will tell if it yields a result. But I think that the strategies pretty spot on.
00:34:15 - 00:35:38
Yeah, it's really fascinating. So I came across a company that is basically cloning, doesn't know there's an avatar or clone of yourself. Does the voice using AI for voice as well as the image and then also it taps into learning from the content that you have. And for example, we've got a lot of content, we've got four or five million words just on jeffbullas.com. We've got, you know, so we've got a lot of content. So compared to the Bloomberg, this world, it's not a lot of content, but in essence, it's 10-12 years, 15 years of content creation. So AI and then you could become an avatar clone of yourself, becomes a coach that can answer questions all within that specific, you know, vertical that you're operating in. So, and we're just in the middle of this AI fire a storm, which is what it is, there's just so many different ideas popping up out of the ether, isn't it? And I'll be intrigued by what you think and what you're doing when you mentioned something about, you've got a sort of like side project that you're doing personally about children's learning and education. Tell us a bit about that.
00:35:39 - 00:38:37
Yeah. I mean, just to examine the potential for AI, you know, for me it's more of a hobby project to do something that's close to my heart, which is, you know, my children, I've got two kids that are eight and five and, you know, the idea of the world they're going into being completely reliant on technology, which means they're reliant on data centers, you know, just really intrigues me so to do a little test. I wanted to find a thoughtful problem. And if you look at the amount of education material for kids in the English language is plentiful and it's not as plentiful in, you know, Welsh or Icelandic or, you know, pick any other language, the collective body of works for children's material is insignificant, all of them combined compared to English. Yeah, so I did a little project to write some software and build an entire learning system for kids bracketed by age and in terms of domain. So like cognitive development or emotional development, physical development and really use AI to help create that framework and then use different AI to translate that into the proper language and how you would speak it in those languages as opposed to just, you know, translate. And it's technically right, but no one would ever say it that way.
You know, how you would use AI to do the corresponding imagery for that. And yeah, I've published a bunch of kids books to test that out. And my takeaway learning from that is, you know, if you're an entrepreneur in the 1900s, you needed $100 million to start a company. If you were an entrepreneur in the 2000s, you probably needed 50 million, 20 tens, you needed 10 million. And now with what you can do with AI, I think there's almost no need to raise any sort of funding to start a company. So the barriers to entry go to zero and that's going to create this whole new renaissance of inspiration of anything that people can dream up to try to solve. They're gonna be able to do that with virtually unlimited resources at their disposal. And, you know, as I reflect on that for Redivider, all of them require this massive computing. And it's critically important to me and my partners and our advisor and, you know, investors and everybody that we do it in a way that has the people of Earth in mind, it's got the protection of the planet in mind and we think those two things will lead to higher profits. And so, you know, hopefully I can do my little part to fuel the imagination of the next generation of entrepreneurs.
00:38:37 - 00:38:43
I love that because everyone said that AI is gonna remove your human creativity. I think it seems to be exploding it.
00:38:43 - 00:38:56
Oh, absolutely. It is like, you know, when you plant a seed in the ground and you see that first tiny bud, that's where we are right now. It is going to absolutely explode.
00:38:56 - 00:39:05
Yeah, this is something that's gonna be with us. It's basically gonna take over this century, isn't it? And beyond.
00:39:06 - 00:40:21
Yeah, I think it's the people who learn how to harness technology are going to position themselves significantly better than those who don't. And I mean, that sort of impact, well let me take a step back. I think that there's a major blocker to people to think that they can do this stuff. You think AI, it sounds super technical. It sounds really scary because it's artificial intelligence and then you start thinking about Terminator. But when you think about it from a social impact idea of, if you help communities learn how to harness these technologies, they're no longer geographically bound or constrained. So if you are in a third world country and you have wifi access, you could transform not only your own life, but you could transform the life of everyone in your village. If you live in a town in America, that used to be a, you know, a paper mill that was shut down and you kind of feel like the opportunities are limited to what's physically nearby. This idea of AI and computing and enabled by edge and all this type of stuff can really unlock multigenerational potential.
00:40:22 - 00:40:39
I totally agree with you. The future is very exciting. I think, you know, some of these stories and myths about, you know, the AI robot is gonna turn the world into one, just big paper clip and humanity is gonna disappear. I went to really.
00:40:40 - 00:41:28
That's a risk. I mean, you know, that is definitely a risk. The Air Force did a project very high governance, really great, you know, I think they executed extremely well and they were training this drone to, you know, take out something or other. And, you know, in the simulation it decided that the human that was evaluating the, yes, no, of should they take out this target or that target? It decided that that was the weakness from it, getting its points, right? The paper clip thought process, it was embodied in an experiment in the US defense department like, you know, so there are things we have to watch out for and we have to do this with respect and cons and constraint. But we shouldn't forget that we get to design our future. And, you know, this is our moment to do that.
00:41:29 - 00:41:40
So Sam Altman is touring the world basically at the moment talking to governments about, you know, the ethics and governance of AI, how do you see that?
00:41:41 - 00:43:09
Yeah, very, very complicated. You know, if you take it from a freedom perspective, you know, as an American, we have a deep history of freedoms built into our kind of soul and our life from birth. I think the idea of governments exerting control over things like AI is practical. But, you know, also a little bit scary to think of what that governance could turn into. On the flip side, I really respect thought leaders and technical architects that have spent way more time on this than me. And a lot of them raise a lot of concerns about doing this in a way that is going to be an enabler for mankind instead of something that could end mankind. So like, I think it's a complex topic and I don't think that there's a right answer but what I do know is that the governments of the world move very slowly and I think AI is moving at a pace that it's so fast. Someone who's in the industry can't even keep up. I don't care what level you are in the industry, you cannot keep up with how fast AI is going. And I don't know how that's gonna go. I don't know how the governments of the world are possibly going to create a framework that enables innovation at the pace that it's going.
00:43:09 - 00:43:14
Yeah, it's overwhelming, isn't it? Even people that are in the industry.
00:43:15 - 00:43:44
Yeah, every single day, you know, I learned about stuff and I'm like, how, when did that happen? You know, and it's like, I live this every day, you know, so it's very exciting. Again, this feels to me like the mid 1900s when everyone was literally pulling cable to connect the world for the first time and the excitement that was happening on the web and all, you know, the mid 1900s energy I feel is exactly what's happening in 2023.
00:43:45 - 00:44:54
And for me, I could even go back a bit further because my first four, I, on my original college degree was a teaching degree. So, I did that five or six years and I leaped into the technology industry and that was in the mid 1800s. That's when the PC industry took off. So, for me, it was a bit of the Wild West rush. Everyone's trying to create the software, the spreadsheets getting distribution, selling the PCs and there was the clones and there was, you know, like the NEC, Power Mate took on the IBM PC and then there was all other clones and you had all the whole networks just inside the building and then that network got connected to the internet as it emerged in the 1900s. So, yeah, I, for me, I watch it with absolute curiosity and the sense of play and fun because I saw what happened in the 1800s with the PC. And now you just talked about the 1900s with dragging cable to connect computers to the world. And then social media came along as well.
00:44:55 - 00:45:45
Yeah. And, you know, that old saying of history that doesn't repeat itself but it rhymes, you know, is very, it's what it feels like now too. It was IBM mainframe and then it swung over the PC and then it went to servers and then it went to this and then it went to cloud and, you know, there's this constant ebb and flow back and forth and you know, between the idea of edge computing, taking the idea of a data center and putting it in a bunch of places like the PC did all the way to these crazy application layer things like AI and how there's going to be millions and millions and millions of AI models that exist in the world now. You know, it definitely feels like we're probably halfway in that pendulum repeating or rhyming with the past.
00:45:45 - 00:45:56
And like you said before, we're just at the beginning of this, this is really in essence, it's the Wild West part of the start of the revolution, isn't it really?
00:45:57 - 00:47:04
Yeah. And you know, the filter that I apply in this space is a lot of the startup companies again that I think are the cost zero to make now they leverage if you take people process and technology is the three kind of pillars here. You only need one person or two people and process, you can forgo at the early stage. So it's all about the technology and that is really computing power. You know, if you look at how that's going to develop with these resources, as long as the filter is creating a business and not a feature, I think you can really divide the AI world into two. A lot of people are building features and then Google says AI on stage 87 times or whatever that joke was that happened a couple of weeks ago, you know, now all those companies are dead because they only built a feature that's now part of Gmail, right? But the ones who build something novel and unique and useful using their own proprietary data, those are going to generate meaningful change.
00:47:04 - 00:47:39
And the other part that I'm finding intriguing too is that you've got the incumbents and then you've got the new startups, right? So, some of these startups are basically built with AI, they're AI centric, whereas some of the existing aren't AI centric, they're bolting things on or enabling and amplifying what they do with AI. So there's a sort of are the incumbents gonna win or the AI startups that started with a blank sheet of paper gonna win or it's gonna be a bit of both.
00:47:40 - 00:48:49
I mean, from my perspective, it's probably a bit of both. But, you know, one thing I learned working for very large companies in the past and a bunch of, you know, startups, big companies do not invent, they scale. Small companies invent they don't scale, right? And that's because one has distribution, one doesn't. And so I think there's a good symbiotic relationship between the two groups of the big companies. Wait to see what sticks, then they buy it somewhat of a premium to what its scale is. But when they scale that to the size of their infrastructure, that's when it gets real adaption. And I don't think we've seen yet, I mean, OpenAI getting the 10 billion from Microsoft or whatever I think is the first example of big tech acquiring something to scale. And it'll be very interesting to see how frequently that happens in the future and how that happens in the future. But my money would be on uh big companies acquiring a significant amount of AI startups to try to get a little bit of scale and competitive advantage.
00:48:50 - 00:48:01
And a lot of the startups are plugging into the big computing power, for example. And models aren't they to power the small niche verticals?
00:49:02 - 00:50:02
Yeah, I mean, it comes down to the unique data sets and again, that's why we're pretty bullish on being the computing partner because whether whichever one of those groups wins, both of them are gonna require computing power and by doing it in an edge format, we just think that we're in the best position to win. But yeah, the smaller companies that are plugging into a generic model, unless they do something that no one else can do, big companies will just rip them off because it's cheaper than acquiring, you know. So, I think these young companies, you know, there are a few that are doing something novel but I think a lot of them are just, you know, nerds like me who are experimenting and testing and they're doing it under the parameters of a startup company and they may not be successful on their first one or their second one. But, they might, by the time they do their third or fourth because they figure it out, you know, they figure out that novelty is important.
00:50:02 - 00:50:05
Yeah. In other words, that unique selling proposition.
00:50:06 - 00:50:07
00:50:08 - 00:50:30
So, just to wrap it up here, a couple of things, number one is, we've sort of touched on it but can you sum it up in terms of what is, you know, Redivider’s unique selling propositions. And, you've mentioned some, can you distill that into a paragraph?
00:50:32 - 00:51:45
Yeah, I think it's, you know, time to power is one of the big things that we can deliver better than a traditional data center, you know, we can, our goal is to be able to go from a blank parcel of land to a operating data center in 30 days as opposed to 24 to 36 months. So I think that's a big one. The idea that we're aligning costs to the use case of the workload, I think is a big one that is going to enable significantly more companies to invest in experimenting in things like AI or computer vision or the enablement of smart cities. So, you know, I think that's the second one. And the third is the idea that unless we do this in a way that is sustainable, when you think about the environmental impact of data, the growth of data, the consumption of power by this industry. You know, there's a lot to think about there and we're really focused on that sustainability and social impact piece because we were able to build it in from the beginning. So I don't know if that was the summary you were looking for. But I really think those are the three key elements that allow Redivider to focus on our mission, which is to uplift humanity through computing.
00:51:45 - 00:52:21
Alright, that's great. That's what I wanted to distilt into three. It was perfect. Alright, great. So what have you learned along the way? Just to wrap it up, what do you learn along the way in your three years in you've done other startups before business projects, entrepreneurial projects. What have you learned in the last three years that you could share with our audience that are really important that you'd like to share?
00:52:22 - 00:54:17
I think biggest thing is if you're on your own journey to start something, make sure that you think big enough. I think a lot of people don't really think big enough and I'll just give a quick example if that's okay. And this is I don't know why this is my example, but it is, if you invent a bug spray that lasts two hours longer, a lot of people would just go, that's what I'm doing. But if you think a little bigger than that, it's actually you can go spend more time outside, right? You can spend two hours outside longer because my product does this. And if you think bigger than that because that's already a bigger thing. But if you think bigger than that, it's, I get to create memories with my family because I can stay outside longer because my product lasts two hours longer. And I think if you keep pushing this idea of thinking bigger, bigger, bigger, what you'll find is you actually become passionate about something that aligns to a problem instead of a solution. And that is the key thing that I've learned with Redivider in the past three years is, you know, one I'm doing with people I absolutely love and trust and the passion that I have and my partners have allowed us to bring together this group of amazing human beings that all believe the same thing. And instead of just starting a data center company, we've started one that is focusing on opportunity zones and how do we take the waste product of a data center and use that to change the relationship kids have with food or, you know, how do we help manage obesity or job training or, you know, education, like all these other things that we can think really broadly about and bake it in from the beginning. So I'd say think big do something you love with people you trust and should be right.
00:54:17 - 00:54:44
Fantastic, love it. So thanks Tom for sharing your vision and what you're doing and it's very, very cool and awesome. And I look forward to watching you guys continue to grow and I think where AI sitting in the middle of this as well. I think it's pretty exciting times, isn't it?
00:54:45 - 00:54:49
I've been a chatbot the entire time.
00:54:50 - 00:55:38
Well, I sort of have this idea of like an avatar that learns from my content that I've been writing for 10-15 years. A lot of content including, you know, the podcast which has got 100 and 70 guests and all the content that sits in that it's like and I just wrote my, you know, update in my will recently. I'm going well, maybe I need to allow to keep the power on or keep paying the bills for a web hosting company. Maybe someone like yourself so that my clone can actually be interrogated driven by AI that actually in 100 years time, Jeff is still alive, talking and answering questions driven by AI, in other words, to infinity and beyond, in other words, so maybe eternal life is real.
00:55:39 - 00:55:43
You never know. I guess we'll find out. I think it's within our generation to find out how.
00:55:43 - 00:55:53
Yeah, it's very, very, it's fun and interesting just to let the imagination soar, which is obviously what you've done when you started. So, congratulations to you and the team and look forward to watching you grow.
00:55:53 - 00:55:57
Thank you. Appreciate it. Thanks for having me
00:55:57 - 00:56:00
It's been great, Tom. Thank you very much.
Sign up to receive email updates
Enter your name and email address below and I'll send you periodic updates about the podcast.