This is Part 2 of our two-part “Investing in the AI Revolution” series with GigaOm CTO Howard Holton. To see Part 1, please click here.

The tech world moves fast, and no one wants to get left behind.

Emerging technologies like generative AI, large language models, and open-source platforms have the potential to completely transform individual businesses or even entire industries. Those who embrace them will profit, while those who don’t will become irrelevant.

Yet a “Hype Cycle” also tends to accommodate new technologies. Several new movements in the tech world that were believed to be the Next Big Thing turned out not to be. 3D printing and NFTs are recent examples.

How should forward-thinking and growth-minded investors separate out the game-changers from the flashes-in-the-pan? What new technologies are actually gaining momentum, and which will never live up to their expectations?

To answer those questions, we’ve brought in an expert. 7investing CEO Simon Erickson recently spoke with Howard Holton, the Chief Technology Officer of GigaOm. GigaOm brings the decision-making executives of progressive companies up-to-speed about emerging technologies and then helping to implement them across their organizations. (You can also  see last year’s conversation with GigaOm CEO Ben Book here.)

In Part 1 of their conversation, Simon and Howard first addressed the status quo of generative AI. AI is being used for ‘fun’ things today — like creating lifelike images through MidJourney — but even this requires significant computing power. Howard explains that innovative companies are already deploying AI at scale, but that they need appropriate data strategies and governance policies in order to maximize their success rate. This is similarly true for the flood of recent large language models; those that endure will require filters to curate the noisy flood of data from all across the internet is a way that is actually usable and trustable for businesses. One key advantage of AI over human beings is that it does not have the same biases as humans.

The two then turned their sights on hardware, specifically the custom silicon being designed by hyperscalers like Amazon, Meta Platforms, and Microsoft. Chipmakers like AMD and NVIDIA will still have an endless runway of future demand, though niche applications will also continue to be served by customizable chips like FPGAs.

Here in Part 2, the cloud’s Infrastructure-as-a-Service providers like Amazon Web Services, Google Cloud Platform, and Microsoft Azure are finding that cloud computing is becoming more commoditized. Each of the Cloud Titans is looking to create a platform for developers, who are comfortable with their capabilities and eager to deploy what they’re already familiar with at their organizations.

Howard then also spoke in detail about the Metaverse. While intriguing in theory, he also believes it will be very difficult to moderate or to control offensive content, and that monetizing the Metaverse for any corporations’ profit interests could be counter-productive to furthering the interests of its users. He and Simon do agree that digital advertising is a likely income stream that will result from the Metaverse; a next-evolution of the personalized advertising we’ve gotten used to in display ads on websites or video platforms.

In the final segment, Howard discusses the importance of trust in the future of AI. While he believes several AI projects are likely overhyped and will eventually go bust, some that are well-designed and execute well could be incredibly valuable and profitable. Companies should hire a “Chief Trust Officer” who can verify the biases purposely imposed on AI models.

Publicly-traded companies mentioned in this interview include Alphabet, Amazon, AMD, Confluent, CrowdStrike, Meta Platforms, Microsoft, MongoDB, and NVIDIA. 7investing’s advisors or its guests may have positions in the companies mentioned. 

Transcript

Simon Erickson  22:35

(Continuing from Part 1)

Let’s go back to the cloud. Let’s talk about the infrastructure providers or the cloud Titans out there.  I’ve followed lots of your coverage that has kind of said that computing is a commodity market now. Right? Whether you’re working with Azure, whether you work with AWS, whether you’re working with Google, it doesn’t matter. The computing costs themselves are becoming more commoditized. It seems like it’s more about the platform and what the uniqueness of Google can offer you versus what Amazon versus one of the other cloud providers can provide.  How are these companies, the big three, we’ve got the big three out there that are doing a good portion of the computing in the cloud right now. How do they differentiate from one another? And then how are they actually making their money? Where does the majority of their profits actually coming from?

Howard Holton  23:26

Oh, so that’s a really good question. But yeah, to be clear, IaaS, which is the kind of buying the hardware..

Simon Erickson  23:35

“Infrastructure as a service.”

Howard Holton  23:36

Right, yeah, Infrastructure as a Service. It has really become commoditized, as it should. And as we kind of all knew it was eventually going to be. It’s really been there for a while. The real value in these platforms comes from the platforms that they offer. The services that I subscribe to. And then philosophically, I think there’s a philosophical kind of difference between the big three.  And here’s your your five second Buyer’s Guide.  Are you a Microsoft Enterprise customer and very comfortable with the Microsoft ecosystem as a great place for you to be?  Are you a big software development shop? You’ve got a ton of software developers and you probably adopted DevOps relatively early on. Amazon is for you.  Are you a big proponent of open source? And open source is where you live? Google’s a great cloud.  Don’t make a decision based on on my five seconds. But but that kind of is a big differentiators between each of the providers, right? Microsoft is your is absolutely built on the back of all of the cloud services that Microsoft has been offering to its consumers for decades. And the way you interact with Microsoft through the procurement process matches the same way you interact with Microsoft as you would go through an ELA. There are additional ELA benefits for Azure that don’t exist in the other two lifeforms. It is how Microsoft has been built, and thus is the language that they speak. And so if you’re comfortable with that language that makes a big difference.  Amazon comes from kind of the hyper digital world, right? Meeting the needs of companies that start in a garage and build in a garage and then turn into something greater than that on a Monday morning. And so software developers around the world have become very comfortable with Amazon, AWS. That’s when you say, the cloud, that’s the first name that pops into their head.  And then Google is kind of the the wildcard outlier. Google is the least afraid to take a chance. The fastest to adopt open source. And if you roll on your own, Google’s a very comfortable place to be. It’s kind of the story of the three bears. Right, Mama Bear, Papa bear, and Baby bear. This one’s too hard. This one’s too soft. This one’s just right. And you kind of have to go through that, to find to find which one’s comfortable for you. And if at the end of it, you find these aren’t really comfortable. For me, there’s an infinite number of additional clouds, they’re not the only game in town.  And what we’re finding is, once you get comfortable with one, do not stop, don’t pause, don’t slow down at all. Pick another one. What was the second most comfortable, because diversification is really, really important. They all speak cloud, but they speak different dialects. So be prepared. But that diversification is important, because they’re not all to market. At the same time, they’re not all to market with the same capability at the same level at the same maturity at the same kind of ease of use or best fit. And if you become fluent in more than one cloud, more than one hyperscalar, you have the ability to make decisions about putting the right application in the right data, in the right place, for the right use case. And some of it may be as simple as it’s a very simple use case, it’s well understood by all the hyper scalars. But you have a very specific region requirement. And Azure has a data center closer to your customer than Amazon does. Or you may have a direct data sovereignty issue that’s met by one of them better than the other within a particular nation, because that’s also a huge, a huge issue and a huge piece of conversation, and why it really makes sense to be on top of kind of that multi cloud experience.

Simon Erickson  27:37

So know thyself. But find the solution that fits you best for your unique needs or your organization’s.

Howard Holton  27:43

And data sovereignty is a big piece of that. So, know thyself means not just know what you do, but know what you should be doing. What you could be doing or what you will be doing. And in many cases, like data sovereignty, what you must be doing. Because I may not be aware of it today, that doesn’t mean I that it doesn’t exist.

Simon Erickson  28:03

Can you chat a little bit more about the open source piece of this? This is where you open up the code for everybody to see. People can build things that are similar. I mean, we’re starting to see publicly traded companies that are open source, right? MongoDB, Elastic, Confluent. You know, others are out there, they tend to have a service or kind of enterprise plan. That’s not just the free version, where you can do it all on your own. But if you want a little bit of help with things or other features that you pay for. I don’t know how to ask the question, but are you seeing open source as opposed to like something that’s complete handholding at the enterprise level? Is it gaining a lot of adoption right now?

Howard Holton  28:39

So yeah, open source has a bunch of advantages and a bunch of disadvantages. There’s no kind of clear, one size fits all choice.  But open source is what it says on the box, right? The source code itself is open, you can freely copy it, you can freely modify it, and you can freely use it. It is what they say it is. It is free as in Oh, I know forget the tagline for open source. But it’s kind of free is as an opinion. You’re allowed to have one; everyone can have one and yours can differ from mine. None of it makes it right. It doesn’t necessarily make it well informed.  The advantage to open source. Let’s talk about security for a minute. Anyone can audit the code at any time for any reason. This idea that closed source is somehow more secure because it is closed source doesn’t really hold water. Any large open source project simply has so many eyes on the source code. It’s not necessarily secure. But large security holes are generally found very very, very quickly and patched by the community equally quickly.  The downside is there’s also a ton of chaos because of that. If I have 50,000 developers that report to no central authority whatsoever, there’s no veto team leader, that are submitting changes that then have to get voted or not voted to be incorporated in the core of the code, there’s a level of chaos there that can be very hard to follow, if you’re kind of new to open source.  That being said, the advantages are, I can try something by downloading code, incorporating it into an application, and publishing the application and see if it works. I don’t have to talk to a salesperson. I don’t have to go through procurement. I can simply do something because the open source library or application or framework already exists.  Now, where we get into trouble is I now want to take that and go into production. Or I want to scale that in a large way. A lot of open source works great out of the box. And when we decide to scale it, the fragility of the open source becomes apparent.  Kubernetes is a fantastic example of this. It is awesome out of the box. I run a little lab at home. My scale is six nodes, total end of statement. When you’re moving to 600 nodes, it’s extremely fragile. 60 nodes extremely fragile. And what we all found out was, it was really easy to run in the lab, it was really easy to run the dev environment. And then we put it in production and we and we’d let half a million users have at it is sort of the follower.  And that’s where some of the open source companies come with their commercial offering. Here’s our open source Kubernetes. But here’s our paid management tool. Here’s open source. Ansible is a really good example. Right for managing all of your nodes. Ansible is a wonderful tool, wonderful piece of software. Here’s our closed source tower for managing and orchestrating Ansible at scale. So that’s kind of one version of your publicly traded open source companies.  The second is, same exact code. Doesn’t matter, right? The commercial side will have support, I can pick up the phone and call someone. That’s so critical when it’s in production, and it goes down on a Friday night. And the people that have been doing what’s effective, what is effectively best effort, and it’s very good effort to keep it running. I don’t know where to go from here. Having someone that’s right there next to the code that’s right there next to the development team. It’s right there next to the decision makers. Pick up the phone and help work through it is a huge value.  And you’re really buying comfort before anything else. You’re buying insurance, right? The insurance that I’m not alone, that I’m not having to look through and try stuff I found on Google at three o’clock in the morning when my production is three in the afternoon when it’s much more stressful.  I love to see these companies go public. And it’s it’s kind of the ultimate in capitalism, because the cream is rising to the top. It’s all open source. Right? So they’re not waiting on price. They’re not getting big based on price. They’re getting big based on the package. They’re getting big based on the product itself.  It’s kind of the Michelin star rule, right? If you’re familiar with the Michelin program, sorted by the tire company of all things, right as part of a kind of a travel guide. But very, very French in that. They don’t care what the outside of the building looks like. They don’t care how good the parking is. All the things that framers points out. They don’t care about what the neighborhood is. Whether you can leave your $150,000 Mercedes in the parking lot or not. What they care about is what is the quality of the food on the plate. That’s how you get one Michelin star. The quality of the food on the plate. It can be a roadside cart that someone pushes down the road and is in a different place every day and you have to deal with exhaust fumes. But it is the greatest roast chicken you’ve ever had in your life that you are getting.

Simon Erickson  34:01

So open source. So step one, you know, move fast and break things. Step two, if you do break things have someone you can call on the phone. Let’s talk about the third step of that to maybe the protection the security piece of this. Open source cybersecurity, you mentioned Kubernetes and containers, and you have these platforms like Netflix that were the early adopters of this. You realize you can move fast, you can build things incredibly quickly by doing it this way.  But then of course, there’s also the cybersecurity protection, which is even more complicated. We’ve seen companies like CrowdStrike that have taken advantage of this secured endpoints for remote workforce for Kubernetes containers. Anything else like that? What’s your take on the cybersecurity industry right now? It seems like Microsoft has been spending a ton of money on addressing security protections. But it’s also a much, much harder world to defend out there. Right?

Howard Holton  34:49

It’s infinitely harder, and AI and quantum computing is going to just make it worse. The good news is quantum computing is very, very, very expensive. The bad news is is a lot of the large hacking groups are nation states. They’re funded by governments themselves and so budgetary concerns aren’t such a big deal. On the other hand, a lot of those they go after private organizations especially the the kind of SMB and small enterprises are not are not necessarily nation states. If your critical infrastructure doesn’t matter, your size, you’re probably going to be attacked.  But those are starting to be run like businesses to kind of look back right when I got started. It was very similar to the to the old movie Hackers. I don’t mean anything you did. I just mean the actors themselves. Where people trying to just experiment just kind of dig in right? If you watch the there were a couple of movies made about Kevin Mitnick. Mitnick was just experimenting, he found his way. And Shawn Polson kind of same thing, right? He found his way. And these were individuals that were hacking their way in and experimenting, some of them were malicious, no doubt.  Today, these things are run like businesses. They have board meetings, they have P&Ls. So so it is important, the barrier, the cost barrier for entry within security.  But outside of that, we have this thing now called Zero Trust, that I’m certain all of your listeners have have heard in passing. Much like digital transformation. Zero trust is not a product anymore than there was digital transformation. It is not a product, it is a philosophy.  And the philosophy is I can trust nothing. Thus I verify absolutely everything. I verify that that person at that place at that time with that application is all authorized true and correct. That what they are requesting is in line with what they have requested in the past. That the query that’s coming through that request is properly formed. And properly intentioned, very important on the intention part, and that the response back from the application matches the request.  I do that every time it’s requested, even if there’s no not necessarily a user, but it’s an automated system making that request, I still do not trust any piece of that, just because I’ve seen one or more pieces before.  The goal here is to is to effectively minimize the blast radius when something goes wrong. That’s really the goal with zero trust. It used to be I trusted everything that was in within the boundaries of my organization. I own this network, I trust this network. Well, that only works until one piece of that network is compromised. And if I can trust that network, all my data goes away through that compromised piece by saying I trust nothing, I don’t trust anything at all, I no longer trust the application itself. And so the goal is to inspect everything, the goal is to ask a lot of questions.  Now is that feasible? Is that real? No. But from an architectural standpoint, we then go okay, what is reasonable? What is feasible within the context of of zero trust? And how do I engineer this? So as I’m doing new things, as I’m doing additional things, as I’m updating things, I’m not failing to consider the ramifications of the blast radius. If something goes wrong, How bad will it be? And that’s really kind of the fundamental of of the security questions we need to be asking ourselves today. Especially with RSA, you know, right around the corner for us.

Simon Erickson  38:31

Minimize the blast radius. I love it. Keep the explosions to a minimum so there’s no collateral damage that comes from this.  It might be the perfect segue to one of the other topics I wanted to ask you about. This was something you mentioned on an previous livestream that when I heard you mentioned it, I literally almost spit out my coffee from laughing because it was so funny.  But you know, it was about the metaverse. You said that you wanted to see the metaverse die a fiery death. Howard it was a perfect quote that you used for it.  Oh, my goodness, what a project. Mark Zuckerberg is spending $30 billion a year on capital expenditures. I would say that most people are still not experiencing the Metaverse. Except for a time or two here or there at a conference. But you know, they’re still kind of the infinite promise of this becoming the Ready Player One, the Oasis that’s out there.  Tell me more about the fiery death? And then also let’s talk about how we could potentially even control the metaverse if that’s even possible at all. Other than the collateral damage, you know, the blast radius getting completely out of hand.

Howard Holton  39:29

So I think conceptually, the metaverse is is a fantastic idea. And it really like it was called the metaverse in the book Snow Crash. I think Neil Stevenson wrote it in 1989. It might have been 1987. It’s a book that got me started on the path that I that my career. It’s not particularly like innovative from a writing standpoint, the protagonists name is Hero Protagonist. But it’s a ton of fun and the metaverse exists and it exists in kind of the way Musk saw it. And he wrote it three decades before Ready Player One or two decades, two and a half decades before Ready Player One. But Ready Player One, another kind of virtual world that you exist in.  If we look at technology, where technology is going what the internet provides to us, I think that virtual world is the next evolution of the Internet. I think Zuckerberg was real smart to key in on that. I think he picked on exactly the right time and place. He bought Oculus, founded by John Carmack, the creator of Doom. And really set down a path of saying this, this is the next evolution of this kind of thing that started as static pages that became dynamic pages that is now the world we, we kind of seamlessly interact with. I don’t think so much about can I get, I need to buy this thing, let’s get the keys, let’s put on my shoes. As we get in the car, let’s drive to a store, what store carries it, let’s go look at five stores. Even if I’m going to go to the local Target, I check their online stock first.  So what’s the next version of that? Well, the next version of that is can I interact with that much like I interact with the real world? That’s effectively what the metaverse is trying to give to us. I have no problem with any of that. The problem that I have is the internet should not be controlled by one person. And if the metaverse is the next version of the internet, and that is controlled by the corporation, Meta, we will create a dystopian society of that I am I am totally sure. As I said, following that, quote, There is not a single movie you watch that has a virtual world in it, where the real world isn’t a complete dystopia.  And so that’s really my concern. We’ve already seen the danger of giving Facebook all of the data that it has on us as it sits much less, you know, in Instagram and the other apps that they own. I don’t want to give them the totality of the internet. Does that mean that technology is fraught? Absolutely not. But there needs to be an open, shared, distributed version of it. I’d love to see that the web three Dotto, right web three Dotto is is it just truly distributed version of the Internet where we all we being the people all controlling it is and is democratized? And that’s what has to happen with a virtual world that we interact in.

Simon Erickson  42:29

How about the digital advertising piece of this? A lot of the thesis that I see on the metaverse becoming a big deal, how is it going to be subsidized/funded?  And what’s the enterprise involvement in all of this? It tends to be that same progression, like you just mentioned. We now have the Facebook’s of the world, you know, the targeted advertisements of the world that appeal to who you personally are. What it knows about me is my demographic or my previous read behavior to put the perfect ad on the sidebar, or whatever site that I’m visiting. And the metaverse would be the next evolution of that, now I’ve got VR headset on there’s going to be a Coca Cola sign and whatever place that I am or anything like that.  Is this truly going to subsidize the metaverse just like advertising? Subsidized publishing of content on the internet? Or how do you see this progressing? Who’s going to pay for it? How do companies make money from this? Kind of an open ended question. But what do you think about it Howard?

Howard Holton  43:18

Well, that’s fundamentally the problem, right? The fundamental problem with web three Dotto in any of its forms is there isn’t anybody to pay for pay for it.  And I love being a bleeding heart liberal. And I love being kind of a fan of open source and a fan of openness and a fan of democratization and a fan of all of these things. But ultimately, somebody has to pay for it.  And none of us are in a position where we want to pay pay per click. Oh, I want to go visit you know, Amazon.com even if I don’t buy anything? Well, I’m sorry. But you have to pay to get into Amazon because Amazon no longer has the advertising that pays for it.  On the internet. If you are getting something for free, anything at all, you are the product. Not the thing you’re receiving for free. So google.com is not free. It’s just you’re the product, not the search. The search is the mechanism that keeps you engaged and you connected so they can collect data about you that they sell to advertisers. And whether we’ve actively made the decision that that’s okay or passively made the decision that we’ll deal with that we’ll settle for it, it doesn’t really change the fact that the reality is I’m the product, right?  So the metaverse is going to have to be be paid for and it’s going to have to be paid for using advertising. It’s going to have to be subsidized in many, many, many ways. And I doubt we’re all going to be okay with the added micro transaction costs that the advertising offsets. So, yes, so whatever the whatever that that virtual world is, it’s going to have to be subsidized in some way.  Let’s be honest, though, the Internet was not built by advertising the internet is supported now by advertising. The internet was built by porn. As was many, many, many, many, many other technologies. As much as I as I hate to say it, that’s the big moneymaker out there in kind of advancing technology. HD cameras, example, which is a horrible, horrible thing.  So but ultimately someone has to pay for it. And the consumer generally avoids being the one to pay for it. In which case, we will, again, be the product. And I say that fully aware that I don’t really want to pay another $100 a month to have access to an internet that is without advertising. Because I’m not sure that I’m not sure that my that those pieces of my data are really worth $100. I can easily click Delete on an email.  Now, I think there’s privacy issues that we need to be aware of right? The ability to predict someone’s health based on their buying patterns changing. I think that that supersedes the value of like Amazon.com using that to tell me, Hey, we noticed you were shopping for Levi’s, we have a better price on Levi’s. Like there’s some there’s some really innovative uses of that data that is very revealing and very problematic. And that’s the sort of stuff that needs to be legislated. Access to health care is something I think is really important for the entirety of the world. And, you know, we really need to limit how we make decisions for people’s health and wellness and well being based on their, you know, the data that’s available to them in the world wide web.

Simon Erickson  46:42

Absolutely. Howard, as we close this out, you know, I really love having you on the show. I mean, like you’ve just been right at the front line, like we said earlier, of innovation in technology. And it’s something that’s continually cutting edge, it’s always changing.  I want to bring in the hype cycle at this point. You know, the Gartner Hype Cycle kind of shows that there’s going to be a lot of expectations for new technology, sometimes they don’t live up to what they’re expected to, at least right out of the gate. There’s kind of this, this downfall of those expectations, and then they kind of sustainably grow later on over time.  The frame that I want to use that for the final question is, as someone who sees the new technologies that are getting it that are getting deployed and adopted out there, can you tell me either something that has too much expectations baked into it right now, where everyone thinks that all technology is awesome, this is going to be the next biggest thing in the world? And you kind of look at that with a raised eyebrow and saying I’m not really sure about that. Or the other side of the question is, is there something that’s not getting enough attention just yet, that you’re really excited about? You can take it either way, but you know, how you ride the ups and downs of the hype cycle of what’s going on out there?

Howard Holton  47:45

What I have to actually say that the answer is the same for both of them. It is, in fact, AI.  We jump on the bandwagon for AI almost instantly for any individual use. We ride the hype. We’ve spent a lot of time talking about language learning models here. And then we find out it’s not all it’s cracked up to be, and then run away. Right now there’s a ton of hype on LLM. And the thing that I’m not hearing talked about enough is the lack of trust. Solve the trust problem is underfunded. Don’t solve the trust problem. It’s just a hype cycle.  And I think we do that a lot. There’s a PhD student that I was working with in Tokyo, who was trying to figure out how to quantify trust. Think about it. Define trust, for me.

Simon Erickson  48:41

Accuracy score. Accuracy of predictions of what it’s doing what you expect,

Howard Holton  48:44

It is that. But that’s not how I define trust. Trust is my do I believe you do what you say you will do. But that’s a really that’s even a squishy statement. Right?  Trust is a human emotion that we built for other humans, that we have since applied to corporations. Which are as far from human as you could get until we have AI which is even further from human.  And yet trust is integral to the human experience. And something that we just we know it when we see it, which makes it nearly impossible to build to trust. And we’re now to the point where like, if I was to invest in something I would invest in an AI company that had a Chief Trust Officer or something along that lines.  But this is beyond ethics. Because I need to trust that the model is being built in a way that that isn’t free from bias. That’s ridiculous, but understands its bias. I need to be able to trust that the data that’s that it’s giving to me even it can give me some sort of confidence level and not not tell me as though it is always true. I need to be able to trust artificial intelligence and today I cannot. And I don’t mean I need to be able to trust that I’m gonna take over the world. I don’t actually care. I need to be able to trust that it’s not lying when it says two plus two is five. So that would be the big thing for me.

Simon Erickson  50:13

That was a fascinating discussion. Here again, Howard Holton is the Chief Technology Officer of GigaOm. We’ve covered quite a bit, from AI and then LLMs to hardware to trust. Howard, this is a real pleasure, thanks for being a part of the 7investing podcast this afternoon.

Howard Holton  50:26

This was great. I’d be happy to do it again. If anybody has any questions, you can reach me on LinkedIn.

Simon Erickson  50:32

Absolutely. Well, thanks again for tuning into this edition of our 7investing podcast. We’re here to empower you to invest in your future. We are 7investing. Have a great day!